Anthropic - Claude Code Docs Changes

agent-teams.md +387 −0 added

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# Orchestrate teams of Claude Code sessions

7> Coordinate multiple Claude Code instances working together as a team, with shared tasks, inter-agent messaging, and centralized management.

9<Warning>

10 Agent teams are experimental and disabled by default. Enable them by adding `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS` to your [settings.json](/en/settings) or environment. Agent teams have [known limitations](#limitations) around session resumption, task coordination, and shutdown behavior.

11</Warning>

13Agent teams let you coordinate multiple Claude Code instances working together. One session acts as the team lead, coordinating work, assigning tasks, and synthesizing results. Teammates work independently, each in its own context window, and communicate directly with each other.

15Unlike [subagents](/en/sub-agents), which run within a single session and can only report back to the main agent, you can also interact with individual teammates directly without going through the lead.

17This page covers:

19* [When to use agent teams](#when-to-use-agent-teams), including best use cases and how they compare with subagents

20* [Starting a team](#start-your-first-agent-team)

21* [Controlling teammates](#control-your-agent-team), including display modes, task assignment, and delegation

22* [Best practices for parallel work](#best-practices)

24## When to use agent teams

26Agent teams are most effective for tasks where parallel exploration adds real value. See [use case examples](#use-case-examples) for full scenarios. The strongest use cases are:

28* **Research and review**: multiple teammates can investigate different aspects of a problem simultaneously, then share and challenge each other's findings

29* **New modules or features**: teammates can each own a separate piece without stepping on each other

30* **Debugging with competing hypotheses**: teammates test different theories in parallel and converge on the answer faster

31* **Cross-layer coordination**: changes that span frontend, backend, and tests, each owned by a different teammate

33Agent teams add coordination overhead and use significantly more tokens than a single session. They work best when teammates can operate independently. For sequential tasks, same-file edits, or work with many dependencies, a single session or [subagents](/en/sub-agents) are more effective.

35### Compare with subagents

37Both agent teams and [subagents](/en/sub-agents) let you parallelize work, but they operate differently. Choose based on whether your workers need to communicate with each other:

39| | Subagents | Agent teams |

40| :---------------- | :----------------------------------------------- | :-------------------------------------------------- |

41| **Context** | Own context window; results return to the caller | Own context window; fully independent |

42| **Communication** | Report results back to the main agent only | Teammates message each other directly |

43| **Coordination** | Main agent manages all work | Shared task list with self-coordination |

44| **Best for** | Focused tasks where only the result matters | Complex work requiring discussion and collaboration |

45| **Token cost** | Lower: results summarized back to main context | Higher: each teammate is a separate Claude instance |

47Use subagents when you need quick, focused workers that report back. Use agent teams when teammates need to share findings, challenge each other, and coordinate on their own.

49## Enable agent teams

51Agent teams are disabled by default. Enable them by setting the `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS` environment variable to `1`, either in your shell environment or through [settings.json](/en/settings):

53```json settings.json theme={null}

54{

55 "env": {

56 "CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1"

57 }

58}

59```

61## Start your first agent team

63After enabling agent teams, tell Claude to create an agent team and describe the task and the team structure you want in natural language. Claude creates the team, spawns teammates, and coordinates work based on your prompt.

65This example works well because the three roles are independent and can explore the problem without waiting on each other:

67```

68I'm designing a CLI tool that helps developers track TODO comments across

69their codebase. Create an agent team to explore this from different angles: one

70teammate on UX, one on technical architecture, one playing devil's advocate.

71```

73From there, Claude creates a team with a [shared task list](/en/interactive-mode#task-list), spawns teammates for each perspective, has them explore the problem, synthesizes findings, and attempts to [clean up the team](#clean-up-the-team) when finished.

75The lead's terminal lists all teammates and what they're working on. Use Shift+Up/Down to select a teammate and message them directly.

77If you want each teammate in its own split pane, see [Choose a display mode](#choose-a-display-mode).

79## Control your agent team

81Tell the lead what you want in natural language. It handles team coordination, task assignment, and delegation based on your instructions.

83### Choose a display mode

85Agent teams support two display modes:

87* **In-process**: all teammates run inside your main terminal. Use Shift+Up/Down to select a teammate and type to message them directly. Works in any terminal, no extra setup required.

88* **Split panes**: each teammate gets its own pane. You can see everyone's output at once and click into a pane to interact directly. Requires tmux, or iTerm2.

90<Note>

91 `tmux` has known limitations on certain operating systems and traditionally works best on macOS. Using `tmux -CC` in iTerm2 is the suggested entrypoint into `tmux`.

92</Note>

94The default is `"auto"`, which uses split panes if you're already running inside a tmux session, and in-process otherwise. The `"tmux"` setting enables split-pane mode and auto-detects whether to use tmux or iTerm2 based on your terminal. To override, set `teammateMode` in your [settings.json](/en/settings):

96```json theme={null}

97{

98 "teammateMode": "in-process"

99}

100```

101

102To force in-process mode for a single session, pass it as a flag:

103

104```bash theme={null}

105claude --teammate-mode in-process

106```

107

108Split-pane mode requires either [tmux](https://github.com/tmux/tmux/wiki) or iTerm2 with the [`it2` CLI](https://github.com/mkusaka/it2). To install manually:

109

110* **tmux**: install through your system's package manager. See the [tmux wiki](https://github.com/tmux/tmux/wiki/Installing) for platform-specific instructions.

111* **iTerm2**: install the [`it2` CLI](https://github.com/mkusaka/it2), then enable the Python API in **iTerm2 → Settings → General → Magic → Enable Python API**.

112

113### Specify teammates and models

114

115Claude decides the number of teammates to spawn based on your task, or you can specify exactly what you want:

116

117```

118Create a team with 4 teammates to refactor these modules in parallel.

119Use Sonnet for each teammate.

120```

121

122### Require plan approval for teammates

123

124For complex or risky tasks, you can require teammates to plan before implementing. The teammate works in read-only plan mode until the lead approves their approach:

125

126```

127Spawn an architect teammate to refactor the authentication module.

128Require plan approval before they make any changes.

129```

130

131When a teammate finishes planning, it sends a plan approval request to the lead. The lead reviews the plan and either approves it or rejects it with feedback. If rejected, the teammate stays in plan mode, revises based on the feedback, and resubmits. Once approved, the teammate exits plan mode and begins implementation.

132

133The lead makes approval decisions autonomously. To influence the lead's judgment, give it criteria in your prompt, such as "only approve plans that include test coverage" or "reject plans that modify the database schema."

134

135### Use delegate mode

136

137Without delegate mode, the lead sometimes starts implementing tasks itself instead of waiting for teammates. Delegate mode prevents this by restricting the lead to coordination-only tools: spawning, messaging, shutting down teammates, and managing tasks.

138

139This is useful when you want the lead to focus entirely on orchestration, such as breaking down work, assigning tasks, and synthesizing results, without touching code directly.

140

141To enable it, start a team first, then press Shift+Tab to cycle into delegate mode.

142

143### Talk to teammates directly

144

145Each teammate is a full, independent Claude Code session. You can message any teammate directly to give additional instructions, ask follow-up questions, or redirect their approach.

146

147* **In-process mode**: use Shift+Up/Down to select a teammate, then type to send them a message. Press Enter to view a teammate's session, then Escape to interrupt their current turn. Press Ctrl+T to toggle the task list.

148* **Split-pane mode**: click into a teammate's pane to interact with their session directly. Each teammate has a full view of their own terminal.

149

150### Assign and claim tasks

151

152The shared task list coordinates work across the team. The lead creates tasks and teammates work through them. Tasks have three states: pending, in progress, and completed. Tasks can also depend on other tasks: a pending task with unresolved dependencies cannot be claimed until those dependencies are completed.

153

154The lead can assign tasks explicitly, or teammates can self-claim:

155

156* **Lead assigns**: tell the lead which task to give to which teammate

157* **Self-claim**: after finishing a task, a teammate picks up the next unassigned, unblocked task on its own

158

159Task claiming uses file locking to prevent race conditions when multiple teammates try to claim the same task simultaneously.

160

161### Shut down teammates

162

163To gracefully end a teammate's session:

164

165```

166Ask the researcher teammate to shut down

167```

168

169The lead sends a shutdown request. The teammate can approve, exiting gracefully, or reject with an explanation.

170

171### Clean up the team

172

173When you're done, ask the lead to clean up:

174

175```

176Clean up the team

177```

178

179This removes the shared team resources. When the lead runs cleanup, it checks for active teammates and fails if any are still running, so shut them down first.

180

181<Warning>

182 Always use the lead to clean up. Teammates should not run cleanup because their team context may not resolve correctly, potentially leaving resources in an inconsistent state.

183</Warning>

184

185### Enforce quality gates with hooks

186

187Use [hooks](/en/hooks) to enforce rules when teammates finish work or tasks complete:

188

189* [`TeammateIdle`](/en/hooks#teammateidle): runs when a teammate is about to go idle. Exit with code 2 to send feedback and keep the teammate working.

190* [`TaskCompleted`](/en/hooks#taskcompleted): runs when a task is being marked complete. Exit with code 2 to prevent completion and send feedback.

191

192## How agent teams work

193

194This section covers the architecture and mechanics behind agent teams. If you want to start using them, see [Control your agent team](#control-your-agent-team) above.

195

196### How Claude starts agent teams

197

198There are two ways agent teams get started:

199

200* **You request a team**: give Claude a task that benefits from parallel work and explicitly ask for an agent team. Claude creates one based on your instructions.

201* **Claude proposes a team**: if Claude determines your task would benefit from parallel work, it may suggest creating a team. You confirm before it proceeds.

202

203In both cases, you stay in control. Claude won't create a team without your approval.

204

205### Architecture

206

207An agent team consists of:

208

209| Component | Role |

210| :------------ | :----------------------------------------------------------------------------------------- |

211| **Team lead** | The main Claude Code session that creates the team, spawns teammates, and coordinates work |

212| **Teammates** | Separate Claude Code instances that each work on assigned tasks |

213| **Task list** | Shared list of work items that teammates claim and complete |

214| **Mailbox** | Messaging system for communication between agents |

215

216See [Choose a display mode](#choose-a-display-mode) for display configuration options. Teammate messages arrive at the lead automatically.

217

218The system manages task dependencies automatically. When a teammate completes a task that other tasks depend on, blocked tasks unblock without manual intervention.

219

220Teams and tasks are stored locally:

221

222* **Team config**: `~/.claude/teams/{team-name}/config.json`

223* **Task list**: `~/.claude/tasks/{team-name}/`

224

225The team config contains a `members` array with each teammate's name, agent ID, and agent type. Teammates can read this file to discover other team members.

226

227### Permissions

228

229Teammates start with the lead's permission settings. If the lead runs with `--dangerously-skip-permissions`, all teammates do too. After spawning, you can change individual teammate modes, but you can't set per-teammate modes at spawn time.

230

231### Context and communication

232

233Each teammate has its own context window. When spawned, a teammate loads the same project context as a regular session: CLAUDE.md, MCP servers, and skills. It also receives the spawn prompt from the lead. The lead's conversation history does not carry over.

234

235**How teammates share information:**

236

237* **Automatic message delivery**: when teammates send messages, they're delivered automatically to recipients. The lead doesn't need to poll for updates.

238* **Idle notifications**: when a teammate finishes and stops, they automatically notify the lead.

239* **Shared task list**: all agents can see task status and claim available work.

240

241**Teammate messaging:**

242

243* **message**: send a message to one specific teammate

244* **broadcast**: send to all teammates simultaneously. Use sparingly, as costs scale with team size.

245

246### Token usage

247

248Agent teams use significantly more tokens than a single session. Each teammate has its own context window, and token usage scales with the number of active teammates. For research, review, and new feature work, the extra tokens are usually worthwhile. For routine tasks, a single session is more cost-effective. See [agent team token costs](/en/costs#agent-team-token-costs) for usage guidance.

249

250## Use case examples

251

252These examples show how agent teams handle tasks where parallel exploration adds value.

253

254### Run a parallel code review

255

256A single reviewer tends to gravitate toward one type of issue at a time. Splitting review criteria into independent domains means security, performance, and test coverage all get thorough attention simultaneously. The prompt assigns each teammate a distinct lens so they don't overlap:

257

258```

259Create an agent team to review PR #142. Spawn three reviewers:

260- One focused on security implications

261- One checking performance impact

262- One validating test coverage

263Have them each review and report findings.

264```

265

266Each reviewer works from the same PR but applies a different filter. The lead synthesizes findings across all three after they finish.

267

268### Investigate with competing hypotheses

269

270When the root cause is unclear, a single agent tends to find one plausible explanation and stop looking. The prompt fights this by making teammates explicitly adversarial: each one's job is not only to investigate its own theory but to challenge the others'.

271

272```

273Users report the app exits after one message instead of staying connected.

274Spawn 5 agent teammates to investigate different hypotheses. Have them talk to

275each other to try to disprove each other's theories, like a scientific

276debate. Update the findings doc with whatever consensus emerges.

277```

278

279The debate structure is the key mechanism here. Sequential investigation suffers from anchoring: once one theory is explored, subsequent investigation is biased toward it.

280

281With multiple independent investigators actively trying to disprove each other, the theory that survives is much more likely to be the actual root cause.

282

283## Best practices

284

285### Give teammates enough context

286

287Teammates load project context automatically, including CLAUDE.md, MCP servers, and skills, but they don't inherit the lead's conversation history. See [Context and communication](#context-and-communication) for details. Include task-specific details in the spawn prompt:

288

289```

290Spawn a security reviewer teammate with the prompt: "Review the authentication module

291at src/auth/ for security vulnerabilities. Focus on token handling, session

292management, and input validation. The app uses JWT tokens stored in

293httpOnly cookies. Report any issues with severity ratings."

294```

295

296### Size tasks appropriately

297

298* **Too small**: coordination overhead exceeds the benefit

299* **Too large**: teammates work too long without check-ins, increasing risk of wasted effort

300* **Just right**: self-contained units that produce a clear deliverable, such as a function, a test file, or a review

301

302<Tip>

303 The lead breaks work into tasks and assigns them to teammates automatically. If it isn't creating enough tasks, ask it to split the work into smaller pieces. Having 5-6 tasks per teammate keeps everyone productive and lets the lead reassign work if someone gets stuck.

304</Tip>

305

306### Wait for teammates to finish

307

308Sometimes the lead starts implementing tasks itself instead of waiting for teammates. If you notice this:

309

310```

311Wait for your teammates to complete their tasks before proceeding

312```

313

314### Start with research and review

315

316If you're new to agent teams, start with tasks that have clear boundaries and don't require writing code: reviewing a PR, researching a library, or investigating a bug. These tasks show the value of parallel exploration without the coordination challenges that come with parallel implementation.

317

318### Avoid file conflicts

319

320Two teammates editing the same file leads to overwrites. Break the work so each teammate owns a different set of files.

321

322### Monitor and steer

323

324Check in on teammates' progress, redirect approaches that aren't working, and synthesize findings as they come in. Letting a team run unattended for too long increases the risk of wasted effort.

325

326## Troubleshooting

327

328### Teammates not appearing

329

330If teammates aren't appearing after you ask Claude to create a team:

331

332* In in-process mode, teammates may already be running but not visible. Press Shift+Down to cycle through active teammates.

333* Check that the task you gave Claude was complex enough to warrant a team. Claude decides whether to spawn teammates based on the task.

334* If you explicitly requested split panes, ensure tmux is installed and available in your PATH:

335 ```bash theme={null}

336 which tmux

337 ```

338* For iTerm2, verify the `it2` CLI is installed and the Python API is enabled in iTerm2 preferences.

339

340### Too many permission prompts

341

342Teammate permission requests bubble up to the lead, which can create friction. Pre-approve common operations in your [permission settings](/en/permissions) before spawning teammates to reduce interruptions.

343

344### Teammates stopping on errors

345

346Teammates may stop after encountering errors instead of recovering. Check their output using Shift+Up/Down in in-process mode or by clicking the pane in split mode, then either:

347

348* Give them additional instructions directly

349* Spawn a replacement teammate to continue the work

350

351### Lead shuts down before work is done

352

353The lead may decide the team is finished before all tasks are actually complete. If this happens, tell it to keep going. You can also tell the lead to wait for teammates to finish before proceeding if it starts doing work instead of delegating.

354

355### Orphaned tmux sessions

356

357If a tmux session persists after the team ends, it may not have been fully cleaned up. List sessions and kill the one created by the team:

358

359```bash theme={null}

360tmux ls

361tmux kill-session -t <session-name>

362```

363

364## Limitations

365

366Agent teams are experimental. Current limitations to be aware of:

367

368* **No session resumption with in-process teammates**: `/resume` and `/rewind` do not restore in-process teammates. After resuming a session, the lead may attempt to message teammates that no longer exist. If this happens, tell the lead to spawn new teammates.

369* **Task status can lag**: teammates sometimes fail to mark tasks as completed, which blocks dependent tasks. If a task appears stuck, check whether the work is actually done and update the task status manually or tell the lead to nudge the teammate.

370* **Shutdown can be slow**: teammates finish their current request or tool call before shutting down, which can take time.

371* **One team per session**: a lead can only manage one team at a time. Clean up the current team before starting a new one.

372* **No nested teams**: teammates cannot spawn their own teams or teammates. Only the lead can manage the team.

373* **Lead is fixed**: the session that creates the team is the lead for its lifetime. You can't promote a teammate to lead or transfer leadership.

374* **Permissions set at spawn**: all teammates start with the lead's permission mode. You can change individual teammate modes after spawning, but you can't set per-teammate modes at spawn time.

375* **Split panes require tmux or iTerm2**: the default in-process mode works in any terminal. Split-pane mode isn't supported in VS Code's integrated terminal, Windows Terminal, or Ghostty.

376

377<Tip>

378 **`CLAUDE.md` works normally**: teammates read `CLAUDE.md` files from their working directory. Use this to provide project-specific guidance to all teammates.

379</Tip>

380

381## Next steps

382

383Explore related approaches for parallel work and delegation:

384

385* **Lightweight delegation**: [subagents](/en/sub-agents) spawn helper agents for research or verification within your session, better for tasks that don't need inter-agent coordination

386* **Manual parallel sessions**: [Git worktrees](/en/common-workflows#run-parallel-claude-code-sessions-with-git-worktrees) let you run multiple Claude Code sessions yourself without automated team coordination

387* **Compare approaches**: see the [subagent vs agent team](/en/features-overview#compare-similar-features) comparison for a side-by-side breakdown

amazon-bedrock.md +19 −21

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code on Amazon Bedrock5# Claude Code on Amazon Bedrock

2 6

3> Learn about configuring Claude Code through Amazon Bedrock, including setup, IAM configuration, and troubleshooting.7> Learn about configuring Claude Code through Amazon Bedrock, including setup, IAM configuration, and troubleshooting.

143export DISABLE_PROMPT_CACHING=1147export DISABLE_PROMPT_CACHING=1

144```148```

145 149

146<Note>[Prompt caching](https://docs.claude.com/en/docs/build-with-claude/prompt-caching) may not be available in all regions.</Note>150<Note>[Prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching) may not be available in all regions.</Note>

~~147~~

148### 5. Output token configuration

~~149~~

150These are the recommended token settings for Claude Code with Amazon Bedrock:

~~151~~

152```bash theme={null}

153# Recommended output token settings for Bedrock

154export CLAUDE_CODE_MAX_OUTPUT_TOKENS=4096

155export MAX_THINKING_TOKENS=1024

156```

~~157~~

158**Why these values:**

~~159~~

160* **`CLAUDE_CODE_MAX_OUTPUT_TOKENS=4096`**: Bedrock's burndown throttling logic sets a minimum of 4096 tokens as the `max_token` penalty. Setting this lower won't reduce costs but may cut off long tool uses, causing the Claude Code agent loop to fail persistently. Claude Code typically uses less than 4096 output tokens without extended thinking, but may need this headroom for tasks involving significant file creation or Write tool usage.

~~161~~

162* **`MAX_THINKING_TOKENS=1024`**: This provides space for extended thinking without cutting off tool use responses, while still maintaining focused reasoning chains. This balance helps prevent trajectory changes that aren't always helpful for coding tasks specifically.

163 151

164## IAM configuration152## IAM configuration

165 153

209 We recommend creating a dedicated AWS account for Claude Code to simplify cost tracking and access control.197 We recommend creating a dedicated AWS account for Claude Code to simplify cost tracking and access control.

210</Note>198</Note>

211 199

200## AWS Guardrails

201

202[Amazon Bedrock Guardrails](https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html) let you implement content filtering for Claude Code. Create a Guardrail in the [Amazon Bedrock console](https://console.aws.amazon.com/bedrock/), publish a version, then add the Guardrail headers to your [settings file](/en/settings). Enable Cross-Region inference on your Guardrail if you're using cross-region inference profiles.

203

204Example configuration:

205

206```json theme={null}

207{

208 "env": {

209 "ANTHROPIC_CUSTOM_HEADERS": "X-Amzn-Bedrock-GuardrailIdentifier: your-guardrail-id\nX-Amzn-Bedrock-GuardrailVersion: 1"

210 }

211}

212```

213

212## Troubleshooting214## Troubleshooting

213 215

214If you encounter region issues:216If you encounter region issues:

229* [Bedrock pricing](https://aws.amazon.com/bedrock/pricing/)231* [Bedrock pricing](https://aws.amazon.com/bedrock/pricing/)

230* [Bedrock inference profiles](https://docs.aws.amazon.com/bedrock/latest/userguide/inference-profiles-support.html)232* [Bedrock inference profiles](https://docs.aws.amazon.com/bedrock/latest/userguide/inference-profiles-support.html)

231* [Claude Code on Amazon Bedrock: Quick Setup Guide](https://community.aws/content/2tXkZKrZzlrlu0KfH8gST5Dkppq/claude-code-on-amazon-bedrock-quick-setup-guide)- [Claude Code Monitoring Implementation (Bedrock)](https://github.com/aws-solutions-library-samples/guidance-for-claude-code-with-amazon-bedrock/blob/main/assets/docs/MONITORING.md)233* [Claude Code on Amazon Bedrock: Quick Setup Guide](https://community.aws/content/2tXkZKrZzlrlu0KfH8gST5Dkppq/claude-code-on-amazon-bedrock-quick-setup-guide)- [Claude Code Monitoring Implementation (Bedrock)](https://github.com/aws-solutions-library-samples/guidance-for-claude-code-with-amazon-bedrock/blob/main/assets/docs/MONITORING.md)

~~232~~

~~233~~

~~234~~

235> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

analytics.md +180 −47

Details

~~1# Analytics~~1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

2 4

~~3> View detailed usage insights and productivity metrics for your organization's Claude Code deployment.~~5# Track team usage with analytics

4 6

~~5Claude Code provides an analytics dashboard that helps organizations understand developer usage patterns, track productivity metrics, and optimize their Claude Code adoption.~~7> View Claude Code usage metrics, track adoption, and measure engineering velocity in the analytics dashboard.

9Claude Code provides analytics dashboards to help organizations understand developer usage patterns, track contribution metrics, and measure how Claude Code impacts engineering velocity. Access the dashboard for your plan:

12| ----------------------------- | -------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- | ----------------------------------------------------- |

16## Access analytics for Teams and Enterprise

18Navigate to [claude.ai/analytics/claude-code](https://claude.ai/analytics/claude-code). Admins and Owners can view the dashboard.

20The Teams and Enterprise dashboard includes:

22* **Usage metrics**: lines of code accepted, suggestion accept rate, daily active users and sessions

23* **Contribution metrics**: PRs and lines of code shipped with Claude Code assistance, with [GitHub integration](#enable-contribution-metrics)

24* **Leaderboard**: top contributors ranked by Claude Code usage

25* **Data export**: download contribution data as CSV for custom reporting

27### Enable contribution metrics

6 28

7<Note>29<Note>

~~8 Analytics are currently available only for organizations using Claude Code with the Claude API through the Claude Console.~~30 Contribution metrics are in public beta and available on Claude for Teams and Claude for Enterprise plans. These metrics only cover users within your claude.ai organization. Usage through the Claude Console API or third-party integrations is not included.

9</Note>31</Note>

10 32

~~11## Access analytics~~33Usage and adoption data is available for all Claude for Teams and Claude for Enterprise accounts. Contribution metrics require additional setup to connect your GitHub organization.

35You need the Owner role to configure analytics settings. A GitHub admin must install the GitHub app.

37<Warning>

38 Contribution metrics are not available for organizations with [Zero Data Retention](/en/data-usage#data-retention) enabled. The analytics dashboard will show usage metrics only.

39</Warning>

41<Steps>

42 <Step title="Install the GitHub app">

43 A GitHub admin installs the Claude GitHub app on your organization's GitHub account at [github.com/apps/claude](https://github.com/apps/claude).

44 </Step>

46 <Step title="Enable Claude Code analytics">

47 A Claude Owner navigates to [claude.ai/admin-settings/claude-code](https://claude.ai/admin-settings/claude-code) and enables the Claude Code analytics feature.

48 </Step>

50 <Step title="Enable GitHub analytics">

51 On the same page, enable the "GitHub analytics" toggle.

52 </Step>

54 <Step title="Authenticate with GitHub">

55 Complete the GitHub authentication flow and select which GitHub organizations to include in the analysis.

56 </Step>

57</Steps>

12 58

~~13Navigate to the analytics dashboard at [console.anthropic.com/claude-code](https://console.anthropic.com/claude-code).~~59Data typically appears within 24 hours after enabling, with daily updates. If no data appears, you may see one of these messages:

14 60

~~15### Required roles~~61* **"GitHub app required"**: install the GitHub app to view contribution metrics

62* **"Data processing in progress"**: check back in a few days and confirm the GitHub app is installed if data doesn't appear

16 63

~~17* **Primary Owner**~~64Contribution metrics support GitHub Cloud and GitHub Enterprise Server.

~~18* **Owner**~~65

~~19* **Billing**~~66### Review summary metrics

~~20* **Admin**~~

~~21* **Developer**~~

22 67

23<Note>68<Note>

~~24 Users with **User**, **Claude Code User** or **Membership Admin** roles cannot access analytics.~~69 These metrics are deliberately conservative and represent an underestimate of Claude Code's actual impact. Only lines and PRs where there is high confidence in Claude Code's involvement are counted.

25</Note>70</Note>

26 71

~~27## Available metrics~~72The dashboard displays these summary metrics at the top:

74* **PRs with CC**: total count of merged pull requests that contain at least one line of code written with Claude Code

75* **Lines of code with CC**: total lines of code across all merged PRs that were written with Claude Code assistance. Only "effective lines" are counted: lines with more than 3 characters after normalization, excluding empty lines and lines with only brackets or trivial punctuation.

76* **PRs with Claude Code (%)**: percentage of all merged PRs that contain Claude Code-assisted code

77* **Suggestion accept rate**: percentage of times users accept Claude Code's code editing suggestions, including Edit, Write, and NotebookEdit tool usage

78* **Lines of code accepted**: total lines of code written by Claude Code that users have accepted in their sessions. This excludes rejected suggestions and does not track subsequent deletions.

80### Explore the charts

82The dashboard includes several charts to visualize trends over time.

84#### Track adoption

86The Adoption chart shows daily usage trends:

88* **users**: daily active users

89* **sessions**: number of active Claude Code sessions per day

91#### Measure PRs per user

93This chart displays individual developer activity over time:

95* **PRs per user**: total number of PRs merged per day divided by daily active users

96* **users**: daily active users

98Use this to understand how individual productivity changes as Claude Code adoption increases.

100#### View pull requests breakdown

28 101

~~29### Lines of code accepted~~102The Pull requests chart shows a daily breakdown of merged PRs:

30 103

~~31Total lines of code written by Claude Code that users have accepted in their sessions.~~104* **PRs with CC**: pull requests containing Claude Code-assisted code

105* **PRs without CC**: pull requests without Claude Code-assisted code

32 106

~~33* Excludes rejected code suggestions~~107Toggle to **Lines of code** view to see the same breakdown by lines of code rather than PR count.

~~34* Doesn't track subsequent deletions~~

35 108

~~36### Suggestion accept rate~~109#### Find top contributors

37 110

~~38Percentage of times users accept code editing tool usage, including:~~111The Leaderboard shows the top 10 users ranked by contribution volume. Toggle between:

39 112

~~40* Edit~~113* **Pull requests**: shows PRs with Claude Code vs All PRs for each user

~~41* Write~~114* **Lines of code**: shows lines with Claude Code vs All lines for each user

~~42* NotebookEdit~~

43 115

~~44### Activity~~116Click **Export all users** to download complete contribution data for all users as a CSV file. The export includes all users, not just the top 10 displayed.

45 117

~~46**users**: Number of active users in a given day (number on left Y-axis)~~118### PR attribution

47 119

~~48**sessions**: Number of active sessions in a given day (number on right Y-axis)~~120When contribution metrics are enabled, Claude Code analyzes merged pull requests to determine which code was written with Claude Code assistance. This is done by matching Claude Code session activity against the code in each PR.

49 121

~~50### Spend~~122#### Tagging criteria

51 123

~~52**users**: Number of active users in a given day (number on left Y-axis)~~124PRs are tagged as "with Claude Code" if they contain at least one line of code written during a Claude Code session. The system uses conservative matching: only code where there is high confidence in Claude Code's involvement is counted as assisted.

53 125

~~54**spend**: Total dollars spent in a given day (number on right Y-axis)~~126#### Attribution process

55 127

~~56### Team insights~~128When a pull request is merged:

57 129

~~58**Members**: All users who have authenticated to Claude Code~~1301. Added lines are extracted from the PR diff

1312. Claude Code sessions that edited matching files within a time window are identified

1323. PR lines are matched against Claude Code output using multiple strategies

1334. Metrics are calculated for AI-assisted lines and total lines

59 134

~~60* API key users are displayed by **API key identifier**~~135Before comparison, lines are normalized: whitespace is trimmed, multiple spaces are collapsed, quotes are standardized, and text is converted to lowercase.

~~61* OAuth users are displayed by **email address**~~

62 136

~~63**Spend this month:** Per-user total spend for the current month.~~137Merged pull requests containing Claude Code-assisted lines are labeled as `claude-code-assisted` in GitHub.

64 138

~~65**Lines this month:** Per-user total of accepted code lines for the current month.~~139#### Time window

66 140

~~67## Using analytics effectively~~141Sessions from 21 days before to 2 days after the PR merge date are considered for attribution matching.

68 142

~~69### Monitor adoption~~143#### Excluded files

70 144

~~71Track team member status to identify:~~145Certain files are automatically excluded from analysis because they are auto-generated:

146

147* Lock files: package-lock.json, yarn.lock, Cargo.lock, and similar

148* Generated code: Protobuf outputs, build artifacts, minified files

149* Build directories: dist/, build/, node\_modules/, target/

150* Test fixtures: snapshots, cassettes, mock data

151* Lines over 1,000 characters, which are likely minified or generated

152

153#### Attribution notes

154

155Keep these additional details in mind when interpreting attribution data:

156

157* Code substantially rewritten by developers, with more than 20% difference, is not attributed to Claude Code

158* Sessions outside the 21-day window are not considered

159* The algorithm does not consider the PR source or destination branch when performing attribution

160

161### Get the most from analytics

162

163Use contribution metrics to demonstrate ROI, identify adoption patterns, and find team members who can help others get started.

164

165#### Monitor adoption

166

167Track the Adoption chart and user counts to identify:

72 168

73* Active users who can share best practices169* Active users who can share best practices

74* Overall adoption trends across your organization170* Overall adoption trends across your organization

171* Dips in usage that may indicate friction or issues

75 172

~~76### Measure productivity~~173#### Measure ROI

77 174

~~78Tool acceptance rates and code metrics help you:~~175Contribution metrics help answer "Is this tool worth the investment?" with data from your own codebase:

79 176

~~80* Understand developer satisfaction with Claude Code suggestions~~177* Track changes in PRs per user over time as adoption increases

~~81* Track code generation effectiveness~~178* Compare PRs and lines of code shipped with vs. without Claude Code

~~82* Identify opportunities for training or process improvements~~179* Use alongside [DORA metrics](https://dora.dev/), sprint velocity, or other engineering KPIs to understand changes from adopting Claude Code

83 180

~~84## Related resources~~181#### Identify power users

182

183The Leaderboard helps you find team members with high Claude Code adoption who can:

184

185* Share prompting techniques and workflows with the team

186* Provide feedback on what's working well

187* Help onboard new users

188

189#### Access data programmatically

190

191To query this data through GitHub, search for PRs labeled with `claude-code-assisted`.

85 192

~~86* [Monitoring usage with OpenTelemetry](/en/monitoring-usage) for custom metrics and alerting~~193## Access analytics for API customers

~~87* [Identity and access management](/en/iam) for role configuration~~

88 194

195API customers using the Claude Console can access analytics at [platform.claude.com/claude-code](https://platform.claude.com/claude-code). You need the UsageView permission to access the dashboard, which is granted to Developer, Billing, Admin, Owner, and Primary Owner roles.

89 196

197<Note>

198 Contribution metrics with GitHub integration are not currently available for API customers. The Console dashboard shows usage and spend metrics only.

199</Note>

200

201The Console dashboard displays:

202

203* **Lines of code accepted**: total lines of code written by Claude Code that users have accepted in their sessions. This excludes rejected suggestions and does not track subsequent deletions.

204* **Suggestion accept rate**: percentage of times users accept code editing tool usage, including Edit, Write, and NotebookEdit tools.

205* **Activity**: daily active users and sessions shown on a chart.

206* **Spend**: daily API costs in dollars alongside user count.

207

208### View team insights

209

210The team insights table shows per-user metrics:

211

212* **Members**: all users who have authenticated to Claude Code. API key users display by key identifier, OAuth users display by email address.

213* **Spend this month**: per-user total API costs for the current month.

214* **Lines this month**: per-user total of accepted code lines for the current month.

215

216<Note>

217 Spend figures in the Console dashboard are estimates for analytics purposes. For actual costs, refer to your billing page.

218</Note>

219

220## Related resources

90 221

~~91> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt~~222* [Monitoring with OpenTelemetry](/en/monitoring-usage): export real-time metrics and events to your observability stack

223* [Manage costs effectively](/en/costs): set spend limits and optimize token usage

224* [Permissions](/en/permissions): configure roles and permissions

authentication.md +104 −0 added

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# Authentication

7> Learn how to configure user authentication and credential management for Claude Code in your organization.

9## Authentication methods

11Setting up Claude Code requires access to Anthropic models. For teams, you can set up Claude Code access in one of these ways:

13* [Claude for Teams or Enterprise](#claude-for-teams-or-enterprise) (recommended)

14* [Claude Console](#claude-console-authentication)

15* [Amazon Bedrock](/en/amazon-bedrock)

16* [Google Vertex AI](/en/google-vertex-ai)

17* [Microsoft Foundry](/en/microsoft-foundry)

19### Claude for Teams or Enterprise

21[Claude for Teams](https://claude.com/pricing#team-&-enterprise) and [Claude for Enterprise](https://anthropic.com/contact-sales) provide the best experience for organizations using Claude Code. Team members get access to both Claude Code and Claude on the web with centralized billing and team management.

23* **Claude for Teams**: self-service plan with collaboration features, admin tools, and billing management. Best for smaller teams.

24* **Claude for Enterprise**: adds SSO, domain capture, role-based permissions, compliance API, and managed policy settings for organization-wide Claude Code configurations. Best for larger organizations with security and compliance requirements.

26<Steps>

27 <Step title="Subscribe">

28 Subscribe to [Claude for Teams](https://claude.com/pricing#team-&-enterprise) or contact sales for [Claude for Enterprise](https://anthropic.com/contact-sales).

29 </Step>

31 <Step title="Invite team members">

32 Invite team members from the admin dashboard.

33 </Step>

35 <Step title="Install and log in">

36 Team members install Claude Code and log in with their Claude.ai accounts.

37 </Step>

38</Steps>

40### Claude Console authentication

42For organizations that prefer API-based billing, you can set up access through the Claude Console.

44<Steps>

45 <Step title="Create or use a Console account">

46 Use your existing Claude Console account or create a new one.

47 </Step>

49 <Step title="Add users">

50 You can add users through either method:

52 * Bulk invite users from within the Console (Console -> Settings -> Members -> Invite)

53 * [Set up SSO](https://support.claude.com/en/articles/13132885-setting-up-single-sign-on-sso)

54 </Step>

56 <Step title="Assign roles">

57 When inviting users, assign one of:

59 * **Claude Code** role: users can only create Claude Code API keys

60 * **Developer** role: users can create any kind of API key

61 </Step>

63 <Step title="Users complete setup">

64 Each invited user needs to:

66 * Accept the Console invite

67 * [Check system requirements](/en/setup#system-requirements)

68 * [Install Claude Code](/en/setup#installation)

69 * Log in with Console account credentials

70 </Step>

71</Steps>

73### Cloud provider authentication

75For teams using Amazon Bedrock, Google Vertex AI, or Microsoft Azure:

77<Steps>

78 <Step title="Follow provider setup">

79 Follow the [Bedrock docs](/en/amazon-bedrock), [Vertex docs](/en/google-vertex-ai), or [Microsoft Foundry docs](/en/microsoft-foundry).

80 </Step>

82 <Step title="Distribute configuration">

83 Distribute the environment variables and instructions for generating cloud credentials to your users. Read more about how to [manage configuration here](/en/settings).

84 </Step>

86 <Step title="Install Claude Code">

87 Users can [install Claude Code](/en/setup#installation).

88 </Step>

89</Steps>

91## Credential management

93Claude Code securely manages your authentication credentials:

95* **Storage location**: on macOS, API keys, OAuth tokens, and other credentials are stored in the encrypted macOS Keychain.

96* **Supported authentication types**: Claude.ai credentials, Claude API credentials, Azure Auth, Bedrock Auth, and Vertex Auth.

97* **Custom credential scripts**: the [`apiKeyHelper`](/en/settings#available-settings) setting can be configured to run a shell script that returns an API key.

98* **Refresh intervals**: by default, `apiKeyHelper` is called after 5 minutes or on HTTP 401 response. Set `CLAUDE_CODE_API_KEY_HELPER_TTL_MS` environment variable for custom refresh intervals.

100## See also

101

102* [Permissions](/en/permissions): configure what Claude Code can access and do

103* [Settings](/en/settings): complete configuration reference

104* [Security](/en/security): security safeguards and best practices

best-practices.md +599 −0 added

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# Best Practices for Claude Code

7> Tips and patterns for getting the most out of Claude Code, from configuring your environment to scaling across parallel sessions.

9Claude Code is an agentic coding environment. Unlike a chatbot that answers questions and waits, Claude Code can read your files, run commands, make changes, and autonomously work through problems while you watch, redirect, or step away entirely.

11This changes how you work. Instead of writing code yourself and asking Claude to review it, you describe what you want and Claude figures out how to build it. Claude explores, plans, and implements.

13But this autonomy still comes with a learning curve. Claude works within certain constraints you need to understand.

15This guide covers patterns that have proven effective across Anthropic's internal teams and for engineers using Claude Code across various codebases, languages, and environments. For how the agentic loop works under the hood, see [How Claude Code works](/en/how-claude-code-works).

17***

19Most best practices are based on one constraint: Claude's context window fills up fast, and performance degrades as it fills.

21Claude's context window holds your entire conversation, including every message, every file Claude reads, and every command output. However, this can fill up fast. A single debugging session or codebase exploration might generate and consume tens of thousands of tokens.

23This matters since LLM performance degrades as context fills. When the context window is getting full, Claude may start "forgetting" earlier instructions or making more mistakes. The context window is the most important resource to manage. Track context usage continuously with a [custom status line](/en/statusline), and see [Reduce token usage](/en/costs#reduce-token-usage) for strategies on reducing token usage.

25***

27## Give Claude a way to verify its work

29<Tip>

30 Include tests, screenshots, or expected outputs so Claude can check itself. This is the single highest-leverage thing you can do.

31</Tip>

33Claude performs dramatically better when it can verify its own work, like run tests, compare screenshots, and validate outputs.

35Without clear success criteria, it might produce something that looks right but actually doesn't work. You become the only feedback loop, and every mistake requires your attention.

37| Strategy | Before | After |

38| ------------------------------------- | ------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

39| **Provide verification criteria** | *"implement a function that validates email addresses"* | *"write a validateEmail function. example test cases: [user@example.com](mailto:user@example.com) is true, invalid is false, [user@.com](mailto:user@.com) is false. run the tests after implementing"* |

40| **Verify UI changes visually** | *"make the dashboard look better"* | *"\[paste screenshot] implement this design. take a screenshot of the result and compare it to the original. list differences and fix them"* |

41| **Address root causes, not symptoms** | *"the build is failing"* | *"the build fails with this error: \[paste error]. fix it and verify the build succeeds. address the root cause, don't suppress the error"* |

43UI changes can be verified using the [Claude in Chrome extension](/en/chrome). It opens new tabs in your browser, tests the UI, and iterates until the code works.

45Your verification can also be a test suite, a linter, or a Bash command that checks output. Invest in making your verification rock-solid.

47***

49## Explore first, then plan, then code

51<Tip>

52 Separate research and planning from implementation to avoid solving the wrong problem.

53</Tip>

55Letting Claude jump straight to coding can produce code that solves the wrong problem. Use [Plan Mode](/en/common-workflows#use-plan-mode-for-safe-code-analysis) to separate exploration from execution.

57The recommended workflow has four phases:

59<Steps>

60 <Step title="Explore">

61 Enter Plan Mode. Claude reads files and answers questions without making changes.

63 ```txt claude (Plan Mode) theme={null}

64 read /src/auth and understand how we handle sessions and login.

65 also look at how we manage environment variables for secrets.

66 ```

67 </Step>

69 <Step title="Plan">

70 Ask Claude to create a detailed implementation plan.

72 ```txt claude (Plan Mode) theme={null}

73 I want to add Google OAuth. What files need to change?

74 What's the session flow? Create a plan.

75 ```

77 Press `Ctrl+G` to open the plan in your text editor for direct editing before Claude proceeds.

78 </Step>

80 <Step title="Implement">

81 Switch back to Normal Mode and let Claude code, verifying against its plan.

83 ```txt claude (Normal Mode) theme={null}

84 implement the OAuth flow from your plan. write tests for the

85 callback handler, run the test suite and fix any failures.

86 ```

87 </Step>

89 <Step title="Commit">

90 Ask Claude to commit with a descriptive message and create a PR.

92 ```txt claude (Normal Mode) theme={null}

93 commit with a descriptive message and open a PR

94 ```

95 </Step>

96</Steps>

98<Callout>

99 Plan Mode is useful, but also adds overhead.

100

101 For tasks where the scope is clear and the fix is small (like fixing a typo, adding a log line, or renaming a variable) ask Claude to do it directly.

102

103 Planning is most useful when you're uncertain about the approach, when the change modifies multiple files, or when you're unfamiliar with the code being modified. If you could describe the diff in one sentence, skip the plan.

104</Callout>

105

106***

107

108## Provide specific context in your prompts

109

110<Tip>

111 The more precise your instructions, the fewer corrections you'll need.

112</Tip>

113

114Claude can infer intent, but it can't read your mind. Reference specific files, mention constraints, and point to example patterns.

115

116| Strategy | Before | After |

117| ------------------------------------------------------------------------------------------------ | ---------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

118| **Scope the task.** Specify which file, what scenario, and testing preferences. | *"add tests for foo.py"* | *"write a test for foo.py covering the edge case where the user is logged out. avoid mocks."* |

119| **Point to sources.** Direct Claude to the source that can answer a question. | *"why does ExecutionFactory have such a weird api?"* | *"look through ExecutionFactory's git history and summarize how its api came to be"* |

120| **Reference existing patterns.** Point Claude to patterns in your codebase. | *"add a calendar widget"* | *"look at how existing widgets are implemented on the home page to understand the patterns. HotDogWidget.php is a good example. follow the pattern to implement a new calendar widget that lets the user select a month and paginate forwards/backwards to pick a year. build from scratch without libraries other than the ones already used in the codebase."* |

121| **Describe the symptom.** Provide the symptom, the likely location, and what "fixed" looks like. | *"fix the login bug"* | *"users report that login fails after session timeout. check the auth flow in src/auth/, especially token refresh. write a failing test that reproduces the issue, then fix it"* |

122

123Vague prompts can be useful when you're exploring and can afford to course-correct. A prompt like `"what would you improve in this file?"` can surface things you wouldn't have thought to ask about.

124

125### Provide rich content

126

127<Tip>

128 Use `@` to reference files, paste screenshots/images, or pipe data directly.

129</Tip>

130

131You can provide rich data to Claude in several ways:

132

133* **Reference files with `@`** instead of describing where code lives. Claude reads the file before responding.

134* **Paste images directly**. Copy/paste or drag and drop images into the prompt.

135* **Give URLs** for documentation and API references. Use `/permissions` to allowlist frequently-used domains.

136* **Pipe in data** by running `cat error.log | claude` to send file contents directly.

137* **Let Claude fetch what it needs**. Tell Claude to pull context itself using Bash commands, MCP tools, or by reading files.

138

139***

140

141## Configure your environment

142

143A few setup steps make Claude Code significantly more effective across all your sessions. For a full overview of extension features and when to use each one, see [Extend Claude Code](/en/features-overview).

144

145### Write an effective CLAUDE.md

146

147<Tip>

148 Run `/init` to generate a starter CLAUDE.md file based on your current project structure, then refine over time.

149</Tip>

150

151CLAUDE.md is a special file that Claude reads at the start of every conversation. Include Bash commands, code style, and workflow rules. This gives Claude persistent context **it can't infer from code alone**.

152

153The `/init` command analyzes your codebase to detect build systems, test frameworks, and code patterns, giving you a solid foundation to refine.

154

155There's no required format for CLAUDE.md files, but keep it short and human-readable. For example:

156

157```markdown CLAUDE.md theme={null}

158# Code style

159- Use ES modules (import/export) syntax, not CommonJS (require)

160- Destructure imports when possible (eg. import { foo } from 'bar')

161

162# Workflow

163- Be sure to typecheck when you're done making a series of code changes

164- Prefer running single tests, and not the whole test suite, for performance

165```

166

167CLAUDE.md is loaded every session, so only include things that apply broadly. For domain knowledge or workflows that are only relevant sometimes, use [skills](/en/skills) instead. Claude loads them on demand without bloating every conversation.

168

169Keep it concise. For each line, ask: *"Would removing this cause Claude to make mistakes?"* If not, cut it. Bloated CLAUDE.md files cause Claude to ignore your actual instructions!

170

171| ✅ Include | ❌ Exclude |

172| ---------------------------------------------------- | -------------------------------------------------- |

173| Bash commands Claude can't guess | Anything Claude can figure out by reading code |

174| Code style rules that differ from defaults | Standard language conventions Claude already knows |

175| Testing instructions and preferred test runners | Detailed API documentation (link to docs instead) |

176| Repository etiquette (branch naming, PR conventions) | Information that changes frequently |

177| Architectural decisions specific to your project | Long explanations or tutorials |

178| Developer environment quirks (required env vars) | File-by-file descriptions of the codebase |

179| Common gotchas or non-obvious behaviors | Self-evident practices like "write clean code" |

180

181If Claude keeps doing something you don't want despite having a rule against it, the file is probably too long and the rule is getting lost. If Claude asks you questions that are answered in CLAUDE.md, the phrasing might be ambiguous. Treat CLAUDE.md like code: review it when things go wrong, prune it regularly, and test changes by observing whether Claude's behavior actually shifts.

182

183You can tune instructions by adding emphasis (e.g., "IMPORTANT" or "YOU MUST") to improve adherence. Check CLAUDE.md into git so your team can contribute. The file compounds in value over time.

184

185CLAUDE.md files can import additional files using `@path/to/import` syntax:

186

187```markdown CLAUDE.md theme={null}

188See @README.md for project overview and @package.json for available npm commands.

189

190# Additional Instructions

191- Git workflow: @docs/git-instructions.md

192- Personal overrides: @~/.claude/my-project-instructions.md

193```

194

195You can place CLAUDE.md files in several locations:

196

197* **Home folder (`~/.claude/CLAUDE.md`)**: Applies to all Claude sessions

198* **Project root (`./CLAUDE.md`)**: Check into git to share with your team, or name it `CLAUDE.local.md` and `.gitignore` it

199* **Parent directories**: Useful for monorepos where both `root/CLAUDE.md` and `root/foo/CLAUDE.md` are pulled in automatically

200* **Child directories**: Claude pulls in child CLAUDE.md files on demand when working with files in those directories

201

202### Configure permissions

203

204<Tip>

205 Use `/permissions` to allowlist safe commands or `/sandbox` for OS-level isolation. This reduces interruptions while keeping you in control.

206</Tip>

207

208By default, Claude Code requests permission for actions that might modify your system: file writes, Bash commands, MCP tools, etc. This is safe but tedious. After the tenth approval you're not really reviewing anymore, you're just clicking through. There are two ways to reduce these interruptions:

209

210* **Permission allowlists**: Permit specific tools you know are safe (like `npm run lint` or `git commit`)

211* **Sandboxing**: Enable OS-level isolation that restricts filesystem and network access, allowing Claude to work more freely within defined boundaries

212

213Alternatively, use `--dangerously-skip-permissions` to bypass all permission checks for contained workflows like fixing lint errors or generating boilerplate.

214

215<Warning>

216 Letting Claude run arbitrary commands can result in data loss, system corruption, or data exfiltration via prompt injection. Only use `--dangerously-skip-permissions` in a sandbox without internet access.

217</Warning>

218

219Read more about [configuring permissions](/en/settings) and [enabling sandboxing](/en/sandboxing#sandboxing).

220

221### Use CLI tools

222

223<Tip>

224 Tell Claude Code to use CLI tools like `gh`, `aws`, `gcloud`, and `sentry-cli` when interacting with external services.

225</Tip>

226

227CLI tools are the most context-efficient way to interact with external services. If you use GitHub, install the `gh` CLI. Claude knows how to use it for creating issues, opening pull requests, and reading comments. Without `gh`, Claude can still use the GitHub API, but unauthenticated requests often hit rate limits.

228

229Claude is also effective at learning CLI tools it doesn't already know. Try prompts like `Use 'foo-cli-tool --help' to learn about foo tool, then use it to solve A, B, C.`

230

231### Connect MCP servers

232

233<Tip>

234 Run `claude mcp add` to connect external tools like Notion, Figma, or your database.

235</Tip>

236

237With [MCP servers](/en/mcp), you can ask Claude to implement features from issue trackers, query databases, analyze monitoring data, integrate designs from Figma, and automate workflows.

238

239### Set up hooks

240

241<Tip>

242 Use hooks for actions that must happen every time with zero exceptions.

243</Tip>

244

245[Hooks](/en/hooks-guide) run scripts automatically at specific points in Claude's workflow. Unlike CLAUDE.md instructions which are advisory, hooks are deterministic and guarantee the action happens.

246

247Claude can write hooks for you. Try prompts like *"Write a hook that runs eslint after every file edit"* or *"Write a hook that blocks writes to the migrations folder."* Run `/hooks` for interactive configuration, or edit `.claude/settings.json` directly.

248

249### Create skills

250

251<Tip>

252 Create `SKILL.md` files in `.claude/skills/` to give Claude domain knowledge and reusable workflows.

253</Tip>

254

255[Skills](/en/skills) extend Claude's knowledge with information specific to your project, team, or domain. Claude applies them automatically when relevant, or you can invoke them directly with `/skill-name`.

256

257Create a skill by adding a directory with a `SKILL.md` to `.claude/skills/`:

258

259```markdown .claude/skills/api-conventions/SKILL.md theme={null}

260---

261name: api-conventions

262description: REST API design conventions for our services

263---

264# API Conventions

265- Use kebab-case for URL paths

266- Use camelCase for JSON properties

267- Always include pagination for list endpoints

268- Version APIs in the URL path (/v1/, /v2/)

269```

270

271Skills can also define repeatable workflows you invoke directly:

272

273```markdown .claude/skills/fix-issue/SKILL.md theme={null}

274---

275name: fix-issue

276description: Fix a GitHub issue

277disable-model-invocation: true

278---

279Analyze and fix the GitHub issue: $ARGUMENTS.

280

2811. Use `gh issue view` to get the issue details

2822. Understand the problem described in the issue

2833. Search the codebase for relevant files

2844. Implement the necessary changes to fix the issue

2855. Write and run tests to verify the fix

2866. Ensure code passes linting and type checking

2877. Create a descriptive commit message

2888. Push and create a PR

289```

290

291Run `/fix-issue 1234` to invoke it. Use `disable-model-invocation: true` for workflows with side effects that you want to trigger manually.

292

293### Create custom subagents

294

295<Tip>

296 Define specialized assistants in `.claude/agents/` that Claude can delegate to for isolated tasks.

297</Tip>

298

299[Subagents](/en/sub-agents) run in their own context with their own set of allowed tools. They're useful for tasks that read many files or need specialized focus without cluttering your main conversation.

300

301```markdown .claude/agents/security-reviewer.md theme={null}

302---

303name: security-reviewer

304description: Reviews code for security vulnerabilities

305tools: Read, Grep, Glob, Bash

306model: opus

307---

308You are a senior security engineer. Review code for:

309- Injection vulnerabilities (SQL, XSS, command injection)

310- Authentication and authorization flaws

311- Secrets or credentials in code

312- Insecure data handling

313

314Provide specific line references and suggested fixes.

315```

316

317Tell Claude to use subagents explicitly: *"Use a subagent to review this code for security issues."*

318

319### Install plugins

320

321<Tip>

322 Run `/plugin` to browse the marketplace. Plugins add skills, tools, and integrations without configuration.

323</Tip>

324

325[Plugins](/en/plugins) bundle skills, hooks, subagents, and MCP servers into a single installable unit from the community and Anthropic. If you work with a typed language, install a [code intelligence plugin](/en/discover-plugins#code-intelligence) to give Claude precise symbol navigation and automatic error detection after edits.

326

327For guidance on choosing between skills, subagents, hooks, and MCP, see [Extend Claude Code](/en/features-overview#match-features-to-your-goal).

328

329***

330

331## Communicate effectively

332

333The way you communicate with Claude Code significantly impacts the quality of results.

334

335### Ask codebase questions

336

337<Tip>

338 Ask Claude questions you'd ask a senior engineer.

339</Tip>

340

341When onboarding to a new codebase, use Claude Code for learning and exploration. You can ask Claude the same sorts of questions you would ask another engineer:

342

343* How does logging work?

344* How do I make a new API endpoint?

345* What does `async move { ... }` do on line 134 of `foo.rs`?

346* What edge cases does `CustomerOnboardingFlowImpl` handle?

347* Why does this code call `foo()` instead of `bar()` on line 333?

348

349Using Claude Code this way is an effective onboarding workflow, improving ramp-up time and reducing load on other engineers. No special prompting required: ask questions directly.

350

351### Let Claude interview you

352

353<Tip>

354 For larger features, have Claude interview you first. Start with a minimal prompt and ask Claude to interview you using the `AskUserQuestion` tool.

355</Tip>

356

357Claude asks about things you might not have considered yet, including technical implementation, UI/UX, edge cases, and tradeoffs.

358

359```

360I want to build [brief description]. Interview me in detail using the AskUserQuestion tool.

361

362Ask about technical implementation, UI/UX, edge cases, concerns, and tradeoffs. Don't ask obvious questions, dig into the hard parts I might not have considered.

363

364Keep interviewing until we've covered everything, then write a complete spec to SPEC.md.

365```

366

367Once the spec is complete, start a fresh session to execute it. The new session has clean context focused entirely on implementation, and you have a written spec to reference.

368

369***

370

371## Manage your session

372

373Conversations are persistent and reversible. Use this to your advantage!

374

375### Course-correct early and often

376

377<Tip>

378 Correct Claude as soon as you notice it going off track.

379</Tip>

380

381The best results come from tight feedback loops. Though Claude occasionally solves problems perfectly on the first attempt, correcting it quickly generally produces better solutions faster.

382

383* **`Esc`**: Stop Claude mid-action with the `Esc` key. Context is preserved, so you can redirect.

384* **`Esc + Esc` or `/rewind`**: Press `Esc` twice or run `/rewind` to open the rewind menu and restore previous conversation and code state, or summarize from a selected message.

385* **`"Undo that"`**: Have Claude revert its changes.

386* **`/clear`**: Reset context between unrelated tasks. Long sessions with irrelevant context can reduce performance.

387

388If you've corrected Claude more than twice on the same issue in one session, the context is cluttered with failed approaches. Run `/clear` and start fresh with a more specific prompt that incorporates what you learned. A clean session with a better prompt almost always outperforms a long session with accumulated corrections.

389

390### Manage context aggressively

391

392<Tip>

393 Run `/clear` between unrelated tasks to reset context.

394</Tip>

395

396Claude Code automatically compacts conversation history when you approach context limits, which preserves important code and decisions while freeing space.

397

398During long sessions, Claude's context window can fill with irrelevant conversation, file contents, and commands. This can reduce performance and sometimes distract Claude.

399

400* Use `/clear` frequently between tasks to reset the context window entirely

401* When auto compaction triggers, Claude summarizes what matters most, including code patterns, file states, and key decisions

402* For more control, run `/compact <instructions>`, like `/compact Focus on the API changes`

403* To compact only part of the conversation, use `Esc + Esc` or `/rewind`, select a message checkpoint, and choose **Summarize from here**. This condenses messages from that point forward while keeping earlier context intact.

404* Customize compaction behavior in CLAUDE.md with instructions like `"When compacting, always preserve the full list of modified files and any test commands"` to ensure critical context survives summarization

405

406### Use subagents for investigation

407

408<Tip>

409 Delegate research with `"use subagents to investigate X"`. They explore in a separate context, keeping your main conversation clean for implementation.

410</Tip>

411

412Since context is your fundamental constraint, subagents are one of the most powerful tools available. When Claude researches a codebase it reads lots of files, all of which consume your context. Subagents run in separate context windows and report back summaries:

413

414```

415Use subagents to investigate how our authentication system handles token

416refresh, and whether we have any existing OAuth utilities I should reuse.

417```

418

419The subagent explores the codebase, reads relevant files, and reports back with findings, all without cluttering your main conversation.

420

421You can also use subagents for verification after Claude implements something:

422

423```

424use a subagent to review this code for edge cases

425```

426

427### Rewind with checkpoints

428

429<Tip>

430 Every action Claude makes creates a checkpoint. You can restore conversation, code, or both to any previous checkpoint.

431</Tip>

432

433Claude automatically checkpoints before changes. Double-tap `Escape` or run `/rewind` to open the rewind menu. You can restore conversation only, restore code only, restore both, or summarize from a selected message. See [Checkpointing](/en/checkpointing) for details.

434

435Instead of carefully planning every move, you can tell Claude to try something risky. If it doesn't work, rewind and try a different approach. Checkpoints persist across sessions, so you can close your terminal and still rewind later.

436

437<Warning>

438 Checkpoints only track changes made *by Claude*, not external processes. This isn't a replacement for git.

439</Warning>

440

441### Resume conversations

442

443<Tip>

444 Run `claude --continue` to pick up where you left off, or `--resume` to choose from recent sessions.

445</Tip>

446

447Claude Code saves conversations locally. When a task spans multiple sessions (you start a feature, get interrupted, come back the next day) you don't have to re-explain the context:

448

449```bash theme={null}

450claude --continue # Resume the most recent conversation

451claude --resume # Select from recent conversations

452```

453

454Use `/rename` to give sessions descriptive names (`"oauth-migration"`, `"debugging-memory-leak"`) so you can find them later. Treat sessions like branches. Different workstreams can have separate, persistent contexts.

455

456***

457

458## Automate and scale

459

460Once you're effective with one Claude, multiply your output with parallel sessions, headless mode, and fan-out patterns.

461

462Everything so far assumes one human, one Claude, and one conversation. But Claude Code scales horizontally. The techniques in this section show how you can get more done.

463

464### Run headless mode

465

466<Tip>

467 Use `claude -p "prompt"` in CI, pre-commit hooks, or scripts. Add `--output-format stream-json` for streaming JSON output.

468</Tip>

469

470With `claude -p "your prompt"`, you can run Claude headlessly, without an interactive session. Headless mode is how you integrate Claude into CI pipelines, pre-commit hooks, or any automated workflow. The output formats (plain text, JSON, streaming JSON) let you parse results programmatically.

471

472```bash theme={null}

473# One-off queries

474claude -p "Explain what this project does"

475

476# Structured output for scripts

477claude -p "List all API endpoints" --output-format json

478

479# Streaming for real-time processing

480claude -p "Analyze this log file" --output-format stream-json

481```

482

483### Run multiple Claude sessions

484

485<Tip>

486 Run multiple Claude sessions in parallel to speed up development, run isolated experiments, or start complex workflows.

487</Tip>

488

489There are three main ways to run parallel sessions:

490

491* [Claude Desktop](/en/desktop): Manage multiple local sessions visually. Each session gets its own isolated worktree.

492* [Claude Code on the web](/en/claude-code-on-the-web): Run on Anthropic's secure cloud infrastructure in isolated VMs.

493* [Agent teams](/en/agent-teams): Automated coordination of multiple sessions with shared tasks, messaging, and a team lead.

494

495Beyond parallelizing work, multiple sessions enable quality-focused workflows. A fresh context improves code review since Claude won't be biased toward code it just wrote.

496

497For example, use a Writer/Reviewer pattern:

498

499| Session A (Writer) | Session B (Reviewer) |

500| ----------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |

501| `Implement a rate limiter for our API endpoints` | |

502| | `Review the rate limiter implementation in @src/middleware/rateLimiter.ts. Look for edge cases, race conditions, and consistency with our existing middleware patterns.` |

503| `Here's the review feedback: [Session B output]. Address these issues.` | |

504

505You can do something similar with tests: have one Claude write tests, then another write code to pass them.

506

507### Fan out across files

508

509<Tip>

510 Loop through tasks calling `claude -p` for each. Use `--allowedTools` to scope permissions for batch operations.

511</Tip>

512

513For large migrations or analyses, you can distribute work across many parallel Claude invocations:

514

515<Steps>

516 <Step title="Generate a task list">

517 Have Claude list all files that need migrating (e.g., `list all 2,000 Python files that need migrating`)

518 </Step>

519

520 <Step title="Write a script to loop through the list">

521 ```bash theme={null}

522 for file in $(cat files.txt); do

523 claude -p "Migrate $file from React to Vue. Return OK or FAIL." \

524 --allowedTools "Edit,Bash(git commit *)"

525 done

526 ```

527 </Step>

528

529 <Step title="Test on a few files, then run at scale">

530 Refine your prompt based on what goes wrong with the first 2-3 files, then run on the full set. The `--allowedTools` flag restricts what Claude can do, which matters when you're running unattended.

531 </Step>

532</Steps>

533

534You can also integrate Claude into existing data/processing pipelines:

535

536```bash theme={null}

537claude -p "<your prompt>" --output-format json | your_command

538```

539

540Use `--verbose` for debugging during development, and turn it off in production.

541

542### Safe Autonomous Mode

543

544Use `claude --dangerously-skip-permissions` to bypass all permission checks and let Claude work uninterrupted. This works well for workflows like fixing lint errors or generating boilerplate code.

545

546<Warning>

547 Letting Claude run arbitrary commands is risky and can result in data loss, system corruption, or data exfiltration (e.g., via prompt injection attacks). To minimize these risks, use `--dangerously-skip-permissions` in a container without internet access.

548

549 With sandboxing enabled (`/sandbox`), you get similar autonomy with better security. Sandbox defines upfront boundaries rather than bypassing all checks.

550</Warning>

551

552***

553

554## Avoid common failure patterns

555

556These are common mistakes. Recognizing them early saves time:

557

558* **The kitchen sink session.** You start with one task, then ask Claude something unrelated, then go back to the first task. Context is full of irrelevant information.

559 > **Fix**: `/clear` between unrelated tasks.

560* **Correcting over and over.** Claude does something wrong, you correct it, it's still wrong, you correct again. Context is polluted with failed approaches.

561 > **Fix**: After two failed corrections, `/clear` and write a better initial prompt incorporating what you learned.

562* **The over-specified CLAUDE.md.** If your CLAUDE.md is too long, Claude ignores half of it because important rules get lost in the noise.

563 > **Fix**: Ruthlessly prune. If Claude already does something correctly without the instruction, delete it or convert it to a hook.

564* **The trust-then-verify gap.** Claude produces a plausible-looking implementation that doesn't handle edge cases.

565 > **Fix**: Always provide verification (tests, scripts, screenshots). If you can't verify it, don't ship it.

566* **The infinite exploration.** You ask Claude to "investigate" something without scoping it. Claude reads hundreds of files, filling the context.

567 > **Fix**: Scope investigations narrowly or use subagents so the exploration doesn't consume your main context.

568

569***

570

571## Develop your intuition

572

573The patterns in this guide aren't set in stone. They're starting points that work well in general, but might not be optimal for every situation.

574

575Sometimes you *should* let context accumulate because you're deep in one complex problem and the history is valuable. Sometimes you should skip planning and let Claude figure it out because the task is exploratory. Sometimes a vague prompt is exactly right because you want to see how Claude interprets the problem before constraining it.

576

577Pay attention to what works. When Claude produces great output, notice what you did: the prompt structure, the context you provided, the mode you were in. When Claude struggles, ask why. Was the context too noisy? The prompt too vague? The task too big for one pass?

578

579Over time, you'll develop intuition that no guide can capture. You'll know when to be specific and when to be open-ended, when to plan and when to explore, when to clear context and when to let it accumulate.

580

581## Related resources

582

583<CardGroup cols={2}>

584 <Card title="How Claude Code works" icon="gear" href="/en/how-claude-code-works">

585 Understand the agentic loop, tools, and context management

586 </Card>

587

588 <Card title="Extend Claude Code" icon="puzzle-piece" href="/en/features-overview">

589 Choose between skills, hooks, MCP, subagents, and plugins

590 </Card>

591

592 <Card title="Common workflows" icon="list-check" href="/en/common-workflows">

593 Step-by-step recipes for debugging, testing, PRs, and more

594 </Card>

595

596 <Card title="CLAUDE.md" icon="file-lines" href="/en/memory">

597 Store project conventions and persistent context

598 </Card>

599</CardGroup>

checkpointing.md +34 −14

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Checkpointing5# Checkpointing

2 6

~~3> Automatically track and rewind Claude's edits to quickly recover from unwanted changes.~~7> Track, rewind, and summarize Claude's edits and conversation to manage session state.

4 8

5Claude Code automatically tracks Claude's file edits as you work, allowing you to quickly undo changes and rewind to previous states if anything gets off track.9Claude Code automatically tracks Claude's file edits as you work, allowing you to quickly undo changes and rewind to previous states if anything gets off track.

6 10

16* Checkpoints persist across sessions, so you can access them in resumed conversations20* Checkpoints persist across sessions, so you can access them in resumed conversations

17* Automatically cleaned up along with sessions after 30 days (configurable)21* Automatically cleaned up along with sessions after 30 days (configurable)

18 22

~~19### Rewinding changes~~23### Rewind and summarize

25Press `Esc` twice (`Esc` + `Esc`) or use the `/rewind` command to open the rewind menu. A scrollable list shows each of your prompts from the session. Select the point you want to act on, then choose an action:

27* **Restore code and conversation**: revert both code and conversation to that point

28* **Restore conversation**: rewind to that message while keeping current code

29* **Restore code**: revert file changes while keeping the conversation

30* **Summarize from here**: compress the conversation from this point forward into a summary, freeing context window space

31* **Never mind**: return to the message list without making changes

33After restoring the conversation or summarizing, the original prompt from the selected message is restored into the input field so you can re-send or edit it.

35#### Restore vs. summarize

20 36

~~21Press `Esc` twice (`Esc` + `Esc`) or use the `/rewind` command to open up the rewind menu. You can choose to restore:~~37The three restore options revert state: they undo code changes, conversation history, or both. "Summarize from here" works differently:

22 38

~~23* **Conversation only**: Rewind to a user message while keeping code changes~~39* Messages before the selected message stay intact

~~24* **Code only**: Revert file changes while keeping the conversation~~40* The selected message and all subsequent messages get replaced with a compact AI-generated summary

~~25* **Both code and conversation**: Restore both to a prior point in the session~~41* No files on disk are changed

42* The original messages are preserved in the session transcript, so Claude can reference the details if needed

44This is similar to `/compact`, but targeted: instead of summarizing the entire conversation, you keep early context in full detail and only compress the parts that are using up space. You can type optional instructions to guide what the summary focuses on.

46<Note>

47 Summarize keeps you in the same session and compresses context. If you want to branch off and try a different approach while preserving the original session intact, use [fork](/en/how-claude-code-works#resume-or-fork-sessions) instead (`claude --continue --fork-session`).

48</Note>

26 49

27## Common use cases50## Common use cases

28 51

29Checkpoints are particularly useful when:52Checkpoints are particularly useful when:

30 53

~~31* **Exploring alternatives**: Try different implementation approaches without losing your starting point~~54* **Exploring alternatives**: try different implementation approaches without losing your starting point

~~32* **Recovering from mistakes**: Quickly undo changes that introduced bugs or broke functionality~~55* **Recovering from mistakes**: quickly undo changes that introduced bugs or broke functionality

~~33* **Iterating on features**: Experiment with variations knowing you can revert to working states~~56* **Iterating on features**: experiment with variations knowing you can revert to working states

57* **Freeing context space**: summarize a verbose debugging session from the midpoint forward, keeping your initial instructions intact

34 58

35## Limitations59## Limitations

36 60

61## See also85## See also

62 86

63* [Interactive mode](/en/interactive-mode) - Keyboard shortcuts and session controls87* [Interactive mode](/en/interactive-mode) - Keyboard shortcuts and session controls

~~64* [Slash commands](/en/slash-commands) - Accessing checkpoints using `/rewind`~~88* [Built-in commands](/en/interactive-mode#built-in-commands) - Accessing checkpoints using `/rewind`

65* [CLI reference](/en/cli-reference) - Command-line options89* [CLI reference](/en/cli-reference) - Command-line options

~~69> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt~~

chrome.md +101 −89

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Use Claude Code with Chrome (beta)5# Use Claude Code with Chrome (beta)

2 6

~~3> Connect Claude Code to your browser to test web apps, debug with console logs, and automate browser tasks.~~7> Connect Claude Code to your Chrome browser to test web apps, debug with console logs, automate form filling, and extract data from web pages.

4 8

~~5<Note>~~9Claude Code integrates with the Claude in Chrome browser extension to give you browser automation capabilities from the CLI or the [VS Code extension](/en/vs-code#automate-browser-tasks-with-chrome). Build your code, then test and debug in the browser without switching contexts.

6 Chrome integration is in beta and currently works with Google Chrome only. It is not yet supported on Brave, Arc, or other Chromium-based browsers. WSL (Windows Subsystem for Linux) is also not supported.

~~7</Note>~~

8 10

9Claude Code integrates with the Claude in Chrome browser extension to give you browser automation capabilities directly from your terminal. Build in your terminal, then test and debug in your browser without switching contexts.11Claude opens new tabs for browser tasks and shares your browser's login state, so it can access any site you're already signed into. Browser actions run in a visible Chrome window in real time. When Claude encounters a login page or CAPTCHA, it pauses and asks you to handle it manually.

10 12

~~11## What the integration enables~~13<Note>

14 Chrome integration is in beta and currently works with Google Chrome and Microsoft Edge. It is not yet supported on Brave, Arc, or other Chromium-based browsers. WSL (Windows Subsystem for Linux) is also not supported.

15</Note>

12 16

13With Chrome connected, you can chain browser actions with terminal commands in a single workflow. For example: scrape documentation from a website, analyze it, generate code based on what you learned, and commit the result.17## Capabilities

14 18

~~15Key capabilities include:~~19With Chrome connected, you can chain browser actions with coding tasks in a single workflow:

16 20

~~17* **Live debugging**: Claude reads console errors and DOM state directly, then fixes the code that caused them~~21* **Live debugging**: read console errors and DOM state directly, then fix the code that caused them

~~18* **Design verification**: Build a UI from a Figma mock, then have Claude open it in the browser and verify it matches~~22* **Design verification**: build a UI from a Figma mock, then open it in the browser to verify it matches

~~19* **Web app testing**: Test form validation, check for visual regressions, or verify user flows work correctly~~23* **Web app testing**: test form validation, check for visual regressions, or verify user flows

~~20* **Authenticated web apps**: Interact with Google Docs, Gmail, Notion, or any app you're logged into without needing API connectors~~24* **Authenticated web apps**: interact with Google Docs, Gmail, Notion, or any app you're logged into without API connectors

~~21* **Data extraction**: Pull structured information from web pages and save it locally~~25* **Data extraction**: pull structured information from web pages and save it locally

~~22* **Task automation**: Automate repetitive browser tasks like data entry, form filling, or multi-site workflows~~26* **Task automation**: automate repetitive browser tasks like data entry, form filling, or multi-site workflows

~~23* **Session recording**: Record browser interactions as GIFs to document or share what happened~~27* **Session recording**: record browser interactions as GIFs to document or share what happened

24 28

25## Prerequisites29## Prerequisites

26 30

27Before using Claude Code with Chrome, you need:31Before using Claude Code with Chrome, you need:

28 32

~~29* [Google Chrome](https://www.google.com/chrome/) browser~~33* [Google Chrome](https://www.google.com/chrome/) or [Microsoft Edge](https://www.microsoft.com/edge) browser

~~30* [Claude in Chrome extension](https://chromewebstore.google.com/detail/claude/fcoeoabgfenejglbffodgkkbkcdhcgfn) version 1.0.36 or higher~~34* [Claude in Chrome extension](https://chromewebstore.google.com/detail/claude/fcoeoabgfenejglbffodgkkbkcdhcgfn) version 1.0.36 or higher, available in the Chrome Web Store for both browsers

~~31* [Claude Code CLI](/en/quickstart#step-1:-install-claude-code) version 2.0.73 or higher~~35* [Claude Code](/en/quickstart#step-1-install-claude-code) version 2.0.73 or higher

~~32* A paid Claude plan (Pro, Team, or Enterprise)~~36* A direct Anthropic plan (Pro, Max, Team, or Enterprise)

~~34## How the integration works~~

36Claude Code communicates with Chrome through the Claude in Chrome browser extension. The extension uses Chrome's [Native Messaging API](https://developer.chrome.com/docs/extensions/develop/concepts/native-messaging) to receive commands from Claude Code and execute them in your browser. This architecture lets Claude Code control browser tabs, read page content, and perform actions while you continue working in your terminal.

38When Claude encounters a login page, CAPTCHA, or other blocker, it pauses and asks you to handle it. You can provide credentials for Claude to enter, or log in manually in the browser. Once you're past the blocker, tell Claude to continue and it picks up where it left off.

40Claude opens new tabs for browser tasks rather than taking over existing ones. However, it shares your browser's login state, so if you're already signed into a site in Chrome, Claude can access it without re-authenticating.

41 37

42<Note>38<Note>

43 The Chrome integration requires a visible browser window. When Claude performs browser actions, you'll see Chrome open and navigate in real time. There's no headless mode since the integration relies on your actual browser session with its login state.39 Chrome integration is not available through third-party providers like Amazon Bedrock, Google Cloud Vertex AI, or Microsoft Foundry. If you access Claude exclusively through a third-party provider, you need a separate claude.ai account to use this feature.

44</Note>40</Note>

45 41

~~46## Set up the integration~~42## Get started in the CLI

47 43

48<Steps>44<Steps>

~~49 <Step title="Update Claude Code">~~45 <Step title="Launch Claude Code with Chrome">

50 Chrome integration requires a recent version of Claude Code. If you installed using the [native installer](/en/quickstart#step-1:-install-claude-code), updates happen automatically. Otherwise, run:46 Start Claude Code with the `--chrome` flag:

51 47

52 ```bash theme={null}48 ```bash theme={null}

~~53 claude update~~49 claude --chrome

54 ```50 ```

52 You can also enable Chrome from within an existing session by running `/chrome`.

55 </Step>53 </Step>

56 54

~~57 <Step title="Start Claude Code with Chrome enabled">~~55 <Step title="Ask Claude to use the browser">

~~58 Launch Claude Code with the `--chrome` flag:~~56 This example navigates to a page, interacts with it, and reports what it finds, all from your terminal or editor:

59 57

~~60 ```bash theme={null}~~58 ```text theme={null}

~~61 claude --chrome~~59 Go to code.claude.com/docs, click on the search box,

60 type "hooks", and tell me what results appear

62 ```61 ```

63 </Step>62 </Step>

~~65 <Step title="Verify the connection">~~

~~66 Run `/chrome` to check the connection status and manage settings. If the extension isn't detected, you'll see a warning with a link to install it.~~

~~67 </Step>~~

68</Steps>63</Steps>

69 64

~~70You can also enable Chrome integration from within an existing session using the `/chrome` slash command.~~65Run `/chrome` at any time to check the connection status, manage permissions, or reconnect the extension.

71 66

~~72## Try it out~~67For VS Code, see [browser automation in VS Code](/en/vs-code#automate-browser-tasks-with-chrome).

73 68

~~74Once connected, type this into Claude to see the integration in action:~~69### Enable Chrome by default

75 70

~~76```~~71To avoid passing `--chrome` each session, run `/chrome` and select "Enabled by default".

~~77Go to code.claude.com/docs, click on the search box,~~

~~78type "hooks", and tell me what results appear~~

~~79```~~

80 72

~~81Claude opens the page, clicks into the search field, types the query, and reports the autocomplete results. This shows navigation, clicking, and typing in a single workflow.~~73In the [VS Code extension](/en/vs-code#automate-browser-tasks-with-chrome), Chrome is available whenever the Chrome extension is installed. No additional flag is needed.

82 74

~~83## Example workflows~~75<Note>

76 Enabling Chrome by default in the CLI increases context usage since browser tools are always loaded. If you notice increased context consumption, disable this setting and use `--chrome` only when needed.

77</Note>

79### Manage site permissions

81Site-level permissions are inherited from the Chrome extension. Manage permissions in the Chrome extension settings to control which sites Claude can browse, click, and type on.

84 82

85Claude can navigate pages, click and type, fill forms, scroll, read console logs and network requests, manage tabs, resize windows, and record GIFs. Run `/mcp` and click into `claude-in-chrome` to see the full list of available tools.83## Example workflows

86 84

~~87The following examples show common patterns for browser automation.~~85These examples show common ways to combine browser actions with coding tasks. Run `/mcp` and select `claude-in-chrome` to see the full list of available browser tools.

88 86

89### Test a local web application87### Test a local web application

90 88

91When developing a web app, ask Claude to verify your changes work correctly:89When developing a web app, ask Claude to verify your changes work correctly:

92 90

~~93```~~91```text theme={null}

94I just updated the login form validation. Can you open localhost:3000,92I just updated the login form validation. Can you open localhost:3000,

95try submitting the form with invalid data, and check if the error93try submitting the form with invalid data, and check if the error

96messages appear correctly?94messages appear correctly?

100 98

101### Debug with console logs99### Debug with console logs

102 100

103If your app has issues, Claude can read console output to help diagnose problems:101Claude can read console output to help diagnose problems. Tell Claude what patterns to look for rather than asking for all console output, since logs can be verbose:

104 102

105```103```text theme={null}

106Open the dashboard page and check the console for any errors when104Open the dashboard page and check the console for any errors when

107the page loads.105the page loads.

108```106```

113 111

114Speed up repetitive data entry tasks:112Speed up repetitive data entry tasks:

115 113

116```114```text theme={null}

117I have a spreadsheet of customer contacts in contacts.csv. For each row,115I have a spreadsheet of customer contacts in contacts.csv. For each row,

118go to our CRM at crm.example.com, click "Add Contact", and fill in the116go to the CRM at crm.example.com, click "Add Contact", and fill in the

119name, email, and phone fields.117name, email, and phone fields.

120```118```

121 119

125 123

126Use Claude to write directly in your documents without API setup:124Use Claude to write directly in your documents without API setup:

127 125

128```126```text theme={null}

129Draft a project update based on our recent commits and add it to my127Draft a project update based on the recent commits and add it to my

130Google Doc at docs.google.com/document/d/abc123128Google Doc at docs.google.com/document/d/abc123

131```129```

132 130

136 134

137Pull structured information from websites:135Pull structured information from websites:

138 136

139```137```text theme={null}

140Go to the product listings page and extract the name, price, and138Go to the product listings page and extract the name, price, and

141availability for each item. Save the results as a CSV file.139availability for each item. Save the results as a CSV file.

142```140```

147 145

148Coordinate tasks across multiple websites:146Coordinate tasks across multiple websites:

149 147

150```148```text theme={null}

151Check my calendar for meetings tomorrow, then for each meeting with149Check my calendar for meetings tomorrow, then for each meeting with

152an external attendee, look up their company on LinkedIn and add a150an external attendee, look up their company website and add a note

153note about what they do.151about what they do.

154```152```

155 153

156Claude works across tabs to gather information and complete the workflow.154Claude works across tabs to gather information and complete the workflow.

159 157

160Create shareable recordings of browser interactions:158Create shareable recordings of browser interactions:

161 159

162```160```text theme={null}

163Record a GIF showing how to complete the checkout flow, from adding161Record a GIF showing how to complete the checkout flow, from adding

164an item to the cart through to the confirmation page.162an item to the cart through to the confirmation page.

165```163```

166 164

167Claude records the interaction sequence and saves it as a GIF file.165Claude records the interaction sequence and saves it as a GIF file.

168 166

169## Best practices

~~170~~

171When using browser automation, keep these guidelines in mind:

~~172~~

173* **Modal dialogs can interrupt the flow**: JavaScript alerts, confirms, and prompts block browser events and prevent Claude from receiving commands. If a dialog appears, dismiss it manually and tell Claude to continue.

174* **Use fresh tabs**: Claude creates new tabs for each session. If a tab becomes unresponsive, ask Claude to create a new one.

175* **Filter console output**: Console logs can be verbose. When debugging, tell Claude what patterns to look for rather than asking for all console output.

~~176~~

177## Troubleshooting167## Troubleshooting

178 168

179### Extension not detected169### Extension not detected

180 170

181If Claude Code shows "Chrome extension not detected":171If Claude Code shows "Chrome extension not detected":

182 172

1831. Verify the Chrome extension (version 1.0.36 or higher) is installed1731. Verify the Chrome extension is installed and enabled in `chrome://extensions`

1842. Verify Claude Code is version 2.0.73 or higher by running `claude --version`1742. Verify Claude Code is up to date by running `claude --version`

1853. Check that Chrome is running1753. Check that Chrome is running

1864. Run `/chrome` and select "Reconnect extension" to re-establish the connection1764. Run `/chrome` and select "Reconnect extension" to re-establish the connection

1875. If the issue persists, restart both Claude Code and Chrome1775. If the issue persists, restart both Claude Code and Chrome

188 178

179The first time you enable Chrome integration, Claude Code installs a native messaging host configuration file. Chrome reads this file on startup, so if the extension isn't detected on your first attempt, restart Chrome to pick up the new configuration.

180

181If the connection still fails, verify the host configuration file exists at:

182

183For Chrome:

184

185* **macOS**: `~/Library/Application Support/Google/Chrome/NativeMessagingHosts/com.anthropic.claude_code_browser_extension.json`

186* **Linux**: `~/.config/google-chrome/NativeMessagingHosts/com.anthropic.claude_code_browser_extension.json`

187* **Windows**: check `HKCU\Software\Google\Chrome\NativeMessagingHosts\` in the Windows Registry

188

189For Edge:

190

191* **macOS**: `~/Library/Application Support/Microsoft Edge/NativeMessagingHosts/com.anthropic.claude_code_browser_extension.json`

192* **Linux**: `~/.config/microsoft-edge/NativeMessagingHosts/com.anthropic.claude_code_browser_extension.json`

193* **Windows**: check `HKCU\Software\Microsoft\Edge\NativeMessagingHosts\` in the Windows Registry

194

189### Browser not responding195### Browser not responding

190 196

191If Claude's browser commands stop working:197If Claude's browser commands stop working:

192 198

1931. Check if a modal dialog (alert, confirm, prompt) is blocking the page1991. Check if a modal dialog (alert, confirm, prompt) is blocking the page. JavaScript dialogs block browser events and prevent Claude from receiving commands. Dismiss the dialog manually, then tell Claude to continue.

1942. Ask Claude to create a new tab and try again2002. Ask Claude to create a new tab and try again

1953. Restart the Chrome extension by disabling and re-enabling it2013. Restart the Chrome extension by disabling and re-enabling it in `chrome://extensions`

196 202

197### First-time setup203### Connection drops during long sessions

198 204

199The first time you use the integration, Claude Code installs a native messaging host that allows communication between the CLI and Chrome. If you encounter permission errors, you may need to restart Chrome for the installation to take effect.205The Chrome extension's service worker can go idle during extended sessions, which breaks the connection. If browser tools stop working after a period of inactivity, run `/chrome` and select "Reconnect extension".

200 206

201## Enable by default207### Windows-specific issues

202 208

203Chrome integration requires the `--chrome` flag each time you start Claude Code. To enable it by default, run `/chrome` and select "Enabled by default".209On Windows, you may encounter:

204 210

205<Note>211* **Named pipe conflicts (EADDRINUSE)**: if another process is using the same named pipe, restart Claude Code. Close any other Claude Code sessions that might be using Chrome.

206 Enabling Chrome by default increases context usage since browser tools are always loaded. If you notice increased context consumption, disable this setting and use `--chrome` only when needed.212* **Native messaging host errors**: if the native messaging host crashes on startup, try reinstalling Claude Code to regenerate the host configuration.

207</Note>

208 213

209Site-level permissions are inherited from the Chrome extension. Manage permissions in the Chrome extension settings to control which sites Claude can browse, click, and type on. Run `/chrome` to see current permission settings.214### Common error messages

210 215

211## See also216These are the most frequently encountered errors and how to resolve them:

212 217

213* [CLI reference](/en/cli-reference) - Command-line flags including `--chrome`218| Error | Cause | Fix |

214* [Common workflows](/en/common-workflows) - More ways to use Claude Code219| ------------------------------------ | ------------------------------------------------ | --------------------------------------------------------------- |

215* [Getting started with Claude for Chrome](https://support.anthropic.com/en/articles/12012173-getting-started-with-claude-for-chrome) - Full documentation for the Chrome extension, including shortcuts, scheduling, and permissions220| "Browser extension is not connected" | Native messaging host cannot reach the extension | Restart Chrome and Claude Code, then run `/chrome` to reconnect |

221| "Extension not detected" | Chrome extension is not installed or is disabled | Install or enable the extension in `chrome://extensions` |

222| "No tab available" | Claude tried to act before a tab was ready | Ask Claude to create a new tab and retry |

223| "Receiving end does not exist" | Extension service worker went idle | Run `/chrome` and select "Reconnect extension" |

216 224

225## See also

217 226

~~218~~ 227* [Use Claude Code in VS Code](/en/vs-code#automate-browser-tasks-with-chrome): browser automation in the VS Code extension

219> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt228* [CLI reference](/en/cli-reference): command-line flags including `--chrome`

229* [Common workflows](/en/common-workflows): more ways to use Claude Code

230* [Data and privacy](/en/data-usage): how Claude Code handles your data

231* [Getting started with Claude in Chrome](https://support.claude.com/en/articles/12012173-getting-started-with-claude-in-chrome): full documentation for the Chrome extension, including shortcuts, scheduling, and permissions

claude-code-on-the-web.md +157 −34

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code on the web5# Claude Code on the web

2 6

3> Run Claude Code tasks asynchronously on secure cloud infrastructure7> Run Claude Code tasks asynchronously on secure cloud infrastructure

16* **Repositories not on your local machine**: Work on code you don't have checked out locally20* **Repositories not on your local machine**: Work on code you don't have checked out locally

17* **Backend changes**: Where Claude Code can write tests and then write code to pass those tests21* **Backend changes**: Where Claude Code can write tests and then write code to pass those tests

18 22

~~19Claude Code is also available on the Claude iOS app. This is perfect for:~~23Claude Code is also available on the Claude iOS app for kicking off tasks on the go and monitoring work in progress.

~~21* **On the go**: Kick off tasks while commuting or away from laptop~~

~~22* **Monitoring**: Watch the trajectory and steer the agent's work~~

23 24

~~24Developers can also move Claude Code sessions from the Claude app to their terminal to continue tasks locally.~~25You can move between local and remote development: [send tasks from your terminal to run on the web](#from-terminal-to-web) with the `&` prefix, or [teleport web sessions back to your terminal](#from-web-to-terminal) to continue locally.

25 26

26## Who can use Claude Code on the web?27## Who can use Claude Code on the web?

27 28

29 30

30* **Pro users**31* **Pro users**

31* **Max users**32* **Max users**

~~32* **Team premium seat users**~~33* **Team users**

~~33* **Enterprise premium seat users**~~34* **Enterprise users** with premium seats or Chat + Claude Code seats

34 35

35## Getting started36## Getting started

36 37

393. Install the Claude GitHub app in your repositories403. Install the Claude GitHub app in your repositories

404. Select your default environment414. Select your default environment

415. Submit your coding task425. Submit your coding task

~~426. Review changes and create a pull request in GitHub~~436. Review changes in diff view, iterate with comments, then create a pull request

43 44

44## How it works45## How it works

45 46

525. **Completion**: You're notified when finished and can create a PR with the changes535. **Completion**: You're notified when finished and can create a PR with the changes

536. **Results**: Changes are pushed to a branch, ready for pull request creation546. **Results**: Changes are pushed to a branch, ready for pull request creation

54 55

56## Review changes with diff view

58Diff view lets you see exactly what Claude changed before creating a pull request. Instead of clicking "Create PR" to review changes in GitHub, view the diff directly in the app and iterate with Claude until the changes are ready.

60When Claude makes changes to files, a diff stats indicator appears showing the number of lines added and removed (for example, `+12 -1`). Select this indicator to open the diff viewer, which displays a file list on the left and the changes for each file on the right.

62From the diff view, you can:

64* Review changes file by file

65* Comment on specific changes to request modifications

66* Continue iterating with Claude based on what you see

68This lets you refine changes through multiple rounds of feedback without creating draft PRs or switching to GitHub.

55## Moving tasks between web and terminal70## Moving tasks between web and terminal

56 71

72You can start tasks on the web and continue them in your terminal, or send tasks from your terminal to run on the web. Web sessions persist even if you close your laptop, and you can monitor them from anywhere including the Claude iOS app.

74<Note>

75 Session handoff is one-way: you can pull web sessions into your terminal, but you can't push an existing terminal session to the web. The [`&` prefix](#from-terminal-to-web) creates a *new* web session with your current conversation context.

76</Note>

78### From terminal to web

80Start a message with `&` inside Claude Code to send a task to run on the web:

82```

83& Fix the authentication bug in src/auth/login.ts

84```

86This creates a new web session on claude.ai with your current conversation context. The task runs in the cloud while you continue working locally. Use `/tasks` to check progress, or open the session on claude.ai or the Claude iOS app to interact directly. From there you can steer Claude, provide feedback, or answer questions just like any other conversation.

88You can also start a web session directly from the command line:

90```bash theme={null}

91claude --remote "Fix the authentication bug in src/auth/login.ts"

92```

94#### Tips for background tasks

96**Plan locally, execute remotely**: For complex tasks, start Claude in plan mode to collaborate on the approach before sending work to the web:

98```bash theme={null}

99claude --permission-mode plan

100```

101

102In plan mode, Claude can only read files and explore the codebase. Once you're satisfied with the plan, send it to the web for autonomous execution:

103

104```

105& Execute the migration plan we discussed

106```

107

108This pattern gives you control over the strategy while letting Claude execute autonomously in the cloud.

109

110**Run tasks in parallel**: Each `&` command creates its own web session that runs independently. You can kick off multiple tasks and they'll all run simultaneously in separate sessions:

111

112```

113& Fix the flaky test in auth.spec.ts

114& Update the API documentation

115& Refactor the logger to use structured output

116```

117

118Monitor all sessions with `/tasks`. When a session completes, you can create a PR from the web interface or [teleport](#from-web-to-terminal) the session to your terminal to continue working.

119

57### From web to terminal120### From web to terminal

58 121

~~59After starting a task on the web:~~122There are several ways to pull a web session into your terminal:

123

124* **Using `/teleport`**: From within Claude Code, run `/teleport` (or `/tp`) to see an interactive picker of your web sessions. If you have uncommitted changes, you'll be prompted to stash them first.

125* **Using `--teleport`**: From the command line, run `claude --teleport` for an interactive session picker, or `claude --teleport <session-id>` to resume a specific session directly.

126* **From `/tasks`**: Run `/tasks` to see your background sessions, then press `t` to teleport into one

127* **From the web interface**: Click "Open in CLI" to copy a command you can paste into your terminal

128

129When you teleport a session, Claude verifies you're in the correct repository, fetches and checks out the branch from the remote session, and loads the full conversation history into your terminal.

130

131#### Requirements for teleporting

132

133Teleport checks these requirements before resuming a session. If any requirement isn't met, you'll see an error or be prompted to resolve the issue.

134

135| Requirement | Details |

136| ------------------ | ---------------------------------------------------------------------------------------------------------------------- |

137| Clean git state | Your working directory must have no uncommitted changes. Teleport prompts you to stash changes if needed. |

138| Correct repository | You must run `--teleport` from a checkout of the same repository, not a fork. |

139| Branch available | The branch from the web session must have been pushed to the remote. Teleport automatically fetches and checks it out. |

140| Same account | You must be authenticated to the same Claude.ai account used in the web session. |

141

142### Sharing sessions

143

144To share a session, toggle its visibility according to the account

145types below. After that, share the session link as-is. Recipients who open your

146shared session will see the latest state of the session upon load, but the

147recipient's page will not update in real time.

148

149#### Sharing from an Enterprise or Teams account

60 150

~~611. Click the "Open in CLI" button~~151For Enterprise and Teams accounts, the two visibility options are **Private**

~~622. Paste and run the command in your terminal in a checkout of the repo~~152and **Team**. Team visibility makes the session visible to other members of your

~~633. Any existing local changes will be stashed, and the remote session will be loaded~~153Claude.ai organization. Repository access verification is enabled by default,

~~644. Continue working locally~~154based on the GitHub account connected to the recipient's account. Your account's

155display name is visible to all recipients with access. [Claude in Slack](/en/slack)

156sessions are automatically shared with Team visibility.

157

158#### Sharing from a Max or Pro account

159

160For Max and Pro accounts, the two visibility options are **Private**

161and **Public**. Public visibility makes the session visible to any user logged

162into claude.ai.

163

164Check your session for sensitive content before sharing. Sessions may contain

165code and credentials from private GitHub repositories. Repository access

166verification is not enabled by default.

167

168Enable repository access verification and/or withhold your name from your shared

169sessions by going to Settings > Claude Code > Sharing settings.

65 170

66## Cloud environment171## Cloud environment

67 172

127 232

128**To update an existing environment:** Select the current environment, to the right of the environment name, and select the settings button. This will open a dialog where you can update the environment name, network access, and environment variables.233**To update an existing environment:** Select the current environment, to the right of the environment name, and select the settings button. This will open a dialog where you can update the environment name, network access, and environment variables.

129 234

235**To select your default environment from the terminal:** If you have multiple environments configured, run `/remote-env` to choose which one to use when starting web sessions from your terminal with `&` or `--remote`. With a single environment, this command shows your current configuration.

236

130<Note>237<Note>

131 Environment variables must be specified as key-value pairs, in [`.env` format](https://www.dotenv.org/). For example:238 Environment variables must be specified as key-value pairs, in [`.env` format](https://www.dotenv.org/). For example:

132 239

138 245

139### Dependency management246### Dependency management

140 247

141Configure automatic dependency installation using [SessionStart hooks](/en/hooks#sessionstart). This can be configured in your repository's `.claude/settings.json` file:248Custom environment images and snapshots are not yet supported. As a workaround, you can use [SessionStart hooks](/en/hooks#sessionstart) to install packages when a session starts. This approach has [known limitations](#dependency-management-limitations).

249

250To configure automatic dependency installation, add a SessionStart hook to your repository's `.claude/settings.json` file:

142 251

143```json theme={null}252```json theme={null}

144{253{

160 269

161Create the corresponding script at `scripts/install_pkgs.sh`:270Create the corresponding script at `scripts/install_pkgs.sh`:

162 271

163```bash theme={null}

164#!/bin/bash

165npm install

166pip install -r requirements.txt

167exit 0

168```

~~169~~

170Make it executable: `chmod +x scripts/install_pkgs.sh`

~~171~~

172#### Local vs remote execution

~~173~~

174By default, all hooks execute both locally and in remote (web) environments. To run a hook only in one environment, check the `CLAUDE_CODE_REMOTE` environment variable in your hook script.

~~175~~

176```bash theme={null}272```bash theme={null}

177#!/bin/bash273#!/bin/bash

178 274

179# Example: Only run in remote environments275# Only run in remote environments

180if [ "$CLAUDE_CODE_REMOTE" != "true" ]; then276if [ "$CLAUDE_CODE_REMOTE" != "true" ]; then

181 exit 0277 exit 0

182fi278fi

183 279

184npm install280npm install

185pip install -r requirements.txt281pip install -r requirements.txt

282exit 0

186```283```

187 284

188#### Persisting environment variables285Make it executable: `chmod +x scripts/install_pkgs.sh`

286

287#### Persist environment variables

288

289SessionStart hooks can persist environment variables for subsequent Bash commands by writing to the file specified in the `CLAUDE_ENV_FILE` environment variable. For details, see [SessionStart hooks](/en/hooks#sessionstart) in the hooks reference.

189 290

190SessionStart hooks can persist environment variables for subsequent bash commands by writing to the file specified in the `CLAUDE_ENV_FILE` environment variable. For details, see [SessionStart hooks](/en/hooks#sessionstart) in the hooks reference.291#### Dependency management limitations

292

293* **Hooks fire for all sessions**: SessionStart hooks run in both local and remote environments. There is no hook configuration to scope a hook to remote sessions only. To skip local execution, check the `CLAUDE_CODE_REMOTE` environment variable in your script as shown above.

294* **Requires network access**: Install commands need network access to reach package registries. If your environment is configured with "No internet" access, these hooks will fail. Use "Limited" (the default) or "Full" network access. The [default allowlist](#default-allowed-domains) includes common registries like npm, PyPI, RubyGems, and crates.io.

295* **Proxy compatibility**: All outbound traffic in remote environments passes through a [security proxy](#security-proxy). Some package managers do not work correctly with this proxy. Bun is a known example.

296* **Runs on every session start**: Hooks run each time a session starts or resumes, adding startup latency. Keep install scripts fast by checking whether dependencies are already present before reinstalling.

191 297

192## Network access and security298## Network access and security

193 299

223 329

224* api.anthropic.com330* api.anthropic.com

225* statsig.anthropic.com331* statsig.anthropic.com

332* platform.claude.com

333* code.claude.com

226* claude.ai334* claude.ai

227 335

228#### Version Control336#### Version Control

230* github.com338* github.com

231* [www.github.com](http://www.github.com)339* [www.github.com](http://www.github.com)

232* api.github.com340* api.github.com

341* npm.pkg.github.com

233* raw\.githubusercontent.com342* raw\.githubusercontent.com

343* pkg-npm.githubusercontent.com

234* objects.githubusercontent.com344* objects.githubusercontent.com

235* codeload.github.com345* codeload.github.com

236* avatars.githubusercontent.com346* avatars.githubusercontent.com

252* [www.docker.com](http://www.docker.com)362* [www.docker.com](http://www.docker.com)

253* production.cloudflare.docker.com363* production.cloudflare.docker.com

254* download.docker.com364* download.docker.com

365* gcr.io

255* \*.gcr.io366* \*.gcr.io

256* ghcr.io367* ghcr.io

257* mcr.microsoft.com368* mcr.microsoft.com

258* \*.data.mcr.microsoft.com369* \*.data.mcr.microsoft.com

370* public.ecr.aws

259 371

260#### Cloud Platforms372#### Cloud Platforms

261 373

276* dot.net388* dot.net

277* visualstudio.com389* visualstudio.com

278* dev.azure.com390* dev.azure.com

391* \*.amazonaws.com

392* \*.api.aws

279* oracle.com393* oracle.com

280* [www.oracle.com](http://www.oracle.com)394* [www.oracle.com](http://www.oracle.com)

281* java.com395* java.com

325 439

326* crates.io440* crates.io

327* [www.crates.io](http://www.crates.io)441* [www.crates.io](http://www.crates.io)

442* index.crates.io

328* static.crates.io443* static.crates.io

329* rustup.rs444* rustup.rs

330* static.rust-lang.org445* static.rust-lang.org

350* gradle.org465* gradle.org

351* [www.gradle.org](http://www.gradle.org)466* [www.gradle.org](http://www.gradle.org)

352* services.gradle.org467* services.gradle.org

468* plugins.gradle.org

469* kotlin.org

470* [www.kotlin.org](http://www.kotlin.org)

353* spring.io471* spring.io

354* repo.spring.io472* repo.spring.io

355 473

423* statsig.com541* statsig.com

424* [www.statsig.com](http://www.statsig.com)542* [www.statsig.com](http://www.statsig.com)

425* api.statsig.com543* api.statsig.com

544* sentry.io

426* \*.sentry.io545* \*.sentry.io

546* http-intake.logs.datadoghq.com

547* \*.datadoghq.com

548* \*.datadoghq.eu

427 549

428#### Content Delivery & Mirrors550#### Content Delivery & Mirrors

429 551

552* sourceforge.net

430* \*.sourceforge.net553* \*.sourceforge.net

431* packagecloud.io554* packagecloud.io

432* \*.packagecloud.io555* \*.packagecloud.io

438* json.schemastore.org561* json.schemastore.org

439* [www.schemastore.org](http://www.schemastore.org)562* [www.schemastore.org](http://www.schemastore.org)

440 563

564#### Model Context Protocol

565

566* \*.modelcontextprotocol.io

567

441<Note>568<Note>

442 Domains marked with `*` indicate wildcard subdomain matching. For example, `*.gcr.io` allows access to any subdomain of `gcr.io`.569 Domains marked with `*` indicate wildcard subdomain matching. For example, `*.gcr.io` allows access to any subdomain of `gcr.io`.

443</Note>570</Note>

482* [Settings reference](/en/settings)609* [Settings reference](/en/settings)

483* [Security](/en/security)610* [Security](/en/security)

484* [Data usage](/en/data-usage)611* [Data usage](/en/data-usage)

~~485~~

~~486~~

~~487~~

488> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

cli-reference.md +44 −32

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# CLI reference5# CLI reference

2 6

3> Complete reference for Claude Code command-line interface, including commands and flags.7> Complete reference for Claude Code command-line interface, including commands and flags.

5## CLI commands9## CLI commands

6 10

~~8| :------------------------------ | :--------------------------------------------- | :------------------------------------------------ |~~12| :------------------------------ | :----------------------------------------------------- | :------------------------------------------------ |

21Customize Claude Code's behavior with these command-line flags:25Customize Claude Code's behavior with these command-line flags:

22 26

24| :------------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | :------------------------------------------------------------------------------------------------- |28| :------------------------------------- | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | :------------------------------------------------------------------------------------------------- |

27| `--agents` | Define custom [subagents](/en/sub-agents) dynamically via JSON (see below for format) | `claude --agents '{"reviewer":{"description":"Reviews code","prompt":"You are a code reviewer"}}'` |31| `--agents` | Define custom [subagents](/en/sub-agents) dynamically via JSON (see below for format) | `claude --agents '{"reviewer":{"description":"Reviews code","prompt":"You are a code reviewer"}}'` |

28| `--allowedTools` | Tools that execute without prompting for permission. To restrict which tools are available, use `--tools` instead | `"Bash(git log:*)" "Bash(git diff:*)" "Read"` |32| `--allow-dangerously-skip-permissions` | Enable permission bypassing as an option without immediately activating it. Allows composing with `--permission-mode` (use with caution) | `claude --permission-mode plan --allow-dangerously-skip-permissions` |

33| `--allowedTools` | Tools that execute without prompting for permission. See [permission rule syntax](/en/settings#permission-rule-syntax) for pattern matching. To restrict which tools are available, use `--tools` instead | `"Bash(git log *)" "Bash(git diff *)" "Read"` |

29| `--append-system-prompt` | Append custom text to the end of the default system prompt (works in both interactive and print modes) | `claude --append-system-prompt "Always use TypeScript"` |34| `--append-system-prompt` | Append custom text to the end of the default system prompt (works in both interactive and print modes) | `claude --append-system-prompt "Always use TypeScript"` |

35| `--append-system-prompt-file` | Load additional system prompt text from a file and append to the default prompt (print mode only) | `claude -p --append-system-prompt-file ./extra-rules.txt "query"` |

38| `--fork-session` | When resuming, create a new session ID instead of reusing the original (use with `--resume` or `--continue`) | `claude --resume abc123 --fork-session` |44| `--fork-session` | When resuming, create a new session ID instead of reusing the original (use with `--resume` or `--continue`) | `claude --resume abc123 --fork-session` |

45| `--from-pr` | Resume sessions linked to a specific GitHub PR. Accepts a PR number or URL. Sessions are automatically linked when created via `gh pr create` | `claude --from-pr 123` |

39| `--ide` | Automatically connect to IDE on startup if exactly one valid IDE is available | `claude --ide` |46| `--ide` | Automatically connect to IDE on startup if exactly one valid IDE is available | `claude --ide` |

47| `--init` | Run initialization hooks and start interactive mode | `claude --init` |

48| `--init-only` | Run initialization hooks and exit (no interactive session) | `claude --init-only` |

40| `--include-partial-messages` | Include partial streaming events in output (requires `--print` and `--output-format=stream-json`) | `claude -p --output-format stream-json --include-partial-messages "query"` |49| `--include-partial-messages` | Include partial streaming events in output (requires `--print` and `--output-format=stream-json`) | `claude -p --output-format stream-json --include-partial-messages "query"` |

42| `--json-schema` | Get validated JSON output matching a JSON Schema after agent completes its workflow (print mode only, see [Agent SDK Structured Outputs](https://docs.claude.com/en/docs/agent-sdk/structured-outputs)) | `claude -p --json-schema '{"type":"object","properties":{...}}' "query"` |51| `--json-schema` | Get validated JSON output matching a JSON Schema after agent completes its workflow (print mode only, see [structured outputs](https://platform.claude.com/docs/en/agent-sdk/structured-outputs)) | `claude -p --json-schema '{"type":"object","properties":{...}}' "query"` |

53| `--max-budget-usd` | Maximum dollar amount to spend on API calls before stopping (print mode only) | `claude -p --max-budget-usd 5.00 "query"` |

54| `--max-turns` | Limit the number of agentic turns (print mode only). Exits with an error when the limit is reached. No limit by default | `claude -p --max-turns 3 "query"` |

45| `--model` | Sets the model for the current session with an alias for the latest model (`sonnet` or `opus`) or a model's full name | `claude --model claude-sonnet-4-5-20250929` |56| `--model` | Sets the model for the current session with an alias for the latest model (`sonnet` or `opus`) or a model's full name | `claude --model claude-sonnet-4-5-20250929` |

58| `--no-session-persistence` | Disable session persistence so sessions are not saved to disk and cannot be resumed (print mode only) | `claude -p --no-session-persistence "query"` |

51| `--print`, `-p` | Print response without interactive mode (see [SDK documentation](https://docs.claude.com/en/docs/agent-sdk) for programmatic usage details) | `claude -p "query"` |63| `--print`, `-p` | Print response without interactive mode (see [Agent SDK documentation](https://platform.claude.com/docs/en/agent-sdk/overview) for programmatic usage details) | `claude -p "query"` |

64| `--remote` | Create a new [web session](/en/claude-code-on-the-web) on claude.ai with the provided task description | `claude --remote "Fix the login bug"` |

73| `--teammate-mode` | Set how [agent team](/en/agent-teams) teammates display: `auto` (default), `in-process`, or `tmux`. See [set up agent teams](/en/agent-teams#set-up-agent-teams) | `claude --teammate-mode in-process` |

74| `--tools` | Restrict which built-in tools Claude can use (works in both interactive and print modes). Use `""` to disable all, `"default"` for all, or tool names like `"Bash,Edit,Read"` | `claude --tools "Bash,Edit,Read"` |

62 77

70The `--agents` flag accepts a JSON object that defines one or more custom subagents. Each subagent requires a unique name (as the key) and a definition object with the following fields:85The `--agents` flag accepts a JSON object that defines one or more custom subagents. Each subagent requires a unique name (as the key) and a definition object with the following fields:

71 86

~~73| :------------ | :------- | :--------------------------------------------------------------------------------------------------------------------- |~~88| :---------------- | :------- | :----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

74| `description` | Yes | Natural language description of when the subagent should be invoked |89| `description` | Yes | Natural language description of when the subagent should be invoked |

75| `prompt` | Yes | The system prompt that guides the subagent's behavior |90| `prompt` | Yes | The system prompt that guides the subagent's behavior |

~~76| `tools` | No | Array of specific tools the subagent can use (for example, `["Read", "Edit", "Bash"]`). If omitted, inherits all tools |~~91| `tools` | No | Array of specific tools the subagent can use, for example `["Read", "Edit", "Bash"]`. If omitted, inherits all tools. Supports [`Task(agent_type)`](/en/sub-agents#restrict-which-subagents-can-be-spawned) syntax |

~~77| `model` | No | Model alias to use: `sonnet`, `opus`, or `haiku`. If omitted, uses the default subagent model |~~92| `disallowedTools` | No | Array of tool names to explicitly deny for this subagent |

93| `model` | No | Model alias to use: `sonnet`, `opus`, `haiku`, or `inherit`. If omitted, defaults to `inherit` |

94| `skills` | No | Array of [skill](/en/skills) names to preload into the subagent's context |

95| `mcpServers` | No | Array of [MCP servers](/en/mcp) for this subagent. Each entry is a server name string or a `{name: config}` object |

96| `maxTurns` | No | Maximum number of agentic turns before the subagent stops |

78 97

79Example:98Example:

80 99

97 116

98### System prompt flags117### System prompt flags

99 118

100Claude Code provides three flags for customizing the system prompt, each serving a different purpose:119Claude Code provides four flags for customizing the system prompt, each serving a different purpose:

101 120

103| :----------------------- | :--------------------------------- | :------------------ | :------------------------------------------------------------------- |122| :---------------------------- | :------------------------------------------ | :------------------ | :------------------------------------------------------------------- |

107 127

108**When to use each:**128**When to use each:**

109 129

122 claude --append-system-prompt "Always use TypeScript and include JSDoc comments"142 claude --append-system-prompt "Always use TypeScript and include JSDoc comments"

123 ```143 ```

124 144

125<Note>145* **`--append-system-prompt-file`**: Use when you want to append instructions from a file while keeping Claude Code's defaults. Useful for version-controlled additions.

126 `--system-prompt` and `--system-prompt-file` are mutually exclusive. You cannot use both flags simultaneously.146 ```bash theme={null}

127</Note>147 claude -p --append-system-prompt-file ./prompts/style-rules.txt "Review this PR"

148 ```

128 149

129<Tip>150`--system-prompt` and `--system-prompt-file` are mutually exclusive. The append flags can be used together with either replacement flag.

130 For most use cases, `--append-system-prompt` is recommended as it preserves Claude Code's built-in capabilities while adding your custom requirements. Use `--system-prompt` or `--system-prompt-file` only when you need complete control over the system prompt.

131</Tip>

132 151

133For detailed information about print mode (`-p`) including output formats,152For most use cases, `--append-system-prompt` or `--append-system-prompt-file` is recommended as they preserve Claude Code's built-in capabilities while adding your custom requirements. Use `--system-prompt` or `--system-prompt-file` only when you need complete control over the system prompt.

134streaming, verbose logging, and programmatic usage, see the

135[SDK documentation](https://docs.claude.com/en/docs/agent-sdk).

136 153

137## See also154## See also

138 155

139* [Chrome extension](/en/chrome) - Browser automation and web testing156* [Chrome extension](/en/chrome) - Browser automation and web testing

140* [Interactive mode](/en/interactive-mode) - Shortcuts, input modes, and interactive features157* [Interactive mode](/en/interactive-mode) - Shortcuts, input modes, and interactive features

141* [Slash commands](/en/slash-commands) - Interactive session commands

142* [Quickstart guide](/en/quickstart) - Getting started with Claude Code158* [Quickstart guide](/en/quickstart) - Getting started with Claude Code

143* [Common workflows](/en/common-workflows) - Advanced workflows and patterns159* [Common workflows](/en/common-workflows) - Advanced workflows and patterns

144* [Settings](/en/settings) - Configuration options160* [Settings](/en/settings) - Configuration options

145* [SDK documentation](https://docs.claude.com/en/docs/agent-sdk) - Programmatic usage and integrations161* [Agent SDK documentation](https://platform.claude.com/docs/en/agent-sdk/overview) - Programmatic usage and integrations

~~146~~

~~147~~

~~148~~

149> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

common-workflows.md +62 −171

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Common workflows5# Common workflows

2 6

~~3> Learn about common workflows with Claude Code.~~7> Step-by-step guides for exploring codebases, fixing bugs, refactoring, testing, and other everyday tasks with Claude Code.

4 8

~~5Each task in this document includes clear instructions, example commands, and best practices to help you get the most from Claude Code.~~9This page covers practical workflows for everyday development: exploring unfamiliar code, debugging, refactoring, writing tests, creating PRs, and managing sessions. Each section includes example prompts you can adapt to your own projects. For higher-level patterns and tips, see [Best practices](/en/best-practices).

6 10

7## Understand new codebases11## Understand new codebases

8 12

81 85

82 * Be specific about what you're looking for86 * Be specific about what you're looking for

83 * Use domain language from the project87 * Use domain language from the project

88 * Install a [code intelligence plugin](/en/discover-plugins#code-intelligence) for your language to give Claude precise "go to definition" and "find references" navigation

84</Tip>89</Tip>

85 90

86***91***

221 226

222## Use Plan Mode for safe code analysis227## Use Plan Mode for safe code analysis

223 228

224Plan Mode instructs Claude to create a plan by analyzing the codebase with read-only operations, perfect for exploring codebases, planning complex changes, or reviewing code safely.229Plan Mode instructs Claude to create a plan by analyzing the codebase with read-only operations, perfect for exploring codebases, planning complex changes, or reviewing code safely. In Plan Mode, Claude uses [`AskUserQuestion`](/en/settings#tools-available-to-claude) to gather requirements and clarify your goals before proposing a plan.

225 230

226### When to use Plan Mode231### When to use Plan Mode

227 232

235 240

236You can switch into Plan Mode during a session using **Shift+Tab** to cycle through permission modes.241You can switch into Plan Mode during a session using **Shift+Tab** to cycle through permission modes.

237 242

238If you are in Normal Mode, **Shift+Tab** first switches into Auto-Accept Mode, indicated by `⏵⏵ accept edits on` at the bottom of the terminal. A subsequent **Shift+Tab** will switch into Plan Mode, indicated by `⏸ plan mode on`.243If you are in Normal Mode, **Shift+Tab** first switches into Auto-Accept Mode, indicated by `⏵⏵ accept edits on` at the bottom of the terminal. A subsequent **Shift+Tab** will switch into Plan Mode, indicated by `⏸ plan mode on`. When an [agent team](/en/agent-teams) is active, the cycle also includes Delegate Mode.

239 244

240**Start a new session in Plan Mode**245**Start a new session in Plan Mode**

241 246

270> How should we handle database migration?275> How should we handle database migration?

271```276```

272 277

278<Tip>Press `Ctrl+G` to open the plan in your default text editor, where you can edit it directly before Claude proceeds.</Tip>

279

273### Configure Plan Mode as default280### Configure Plan Mode as default

274 281

275```json theme={null}282```json theme={null}

323 330

324## Create pull requests331## Create pull requests

325 332

326Suppose you need to create a well-documented pull request for your changes.333You can create pull requests by asking Claude directly ("create a pr for my changes") or by using the `/commit-push-pr` skill, which commits, pushes, and opens a PR in one step.

334

335```

336> /commit-push-pr

337```

338

339If you have a Slack MCP server configured and specify channels in your CLAUDE.md (for example, "post PR URLs to #team-prs"), the skill automatically posts the PR URL to those channels.

340

341For more control over the process, guide Claude through it step-by-step or [create your own skill](/en/skills):

327 342

328<Steps>343<Steps>

329 <Step title="Summarize your changes">344 <Step title="Summarize your changes">

332 ```347 ```

333 </Step>348 </Step>

334 349

335 <Step title="Generate a pull request with Claude">350 <Step title="Generate a pull request">

336 ```351 ```

337 > create a pr352 > create a pr

338 ```353 ```

343 > enhance the PR description with more context about the security improvements358 > enhance the PR description with more context about the security improvements

344 ```359 ```

345 </Step>360 </Step>

~~346~~

347 <Step title="Add testing details">

348 ```

349 > add information about how these changes were tested

350 ```

351 </Step>

352</Steps>361</Steps>

353 362

354<Tip>363When you create a PR using `gh pr create`, the session is automatically linked to that PR. You can resume it later with `claude --from-pr <number>`.

355 Tips:

356 364

357 * Ask Claude directly to make a PR for you365<Tip>

358 * Review Claude's generated PR before submitting366 Review Claude's generated PR before submitting and ask Claude to highlight potential risks or considerations.

359 * Ask Claude to highlight potential risks or considerations

360</Tip>367</Tip>

361 368

362## Handle documentation369## Handle documentation

454 * Include screenshots of errors, UI designs, or diagrams for better context461 * Include screenshots of errors, UI designs, or diagrams for better context

455 * You can work with multiple images in a conversation462 * You can work with multiple images in a conversation

456 * Image analysis works with diagrams, screenshots, mockups, and more463 * Image analysis works with diagrams, screenshots, mockups, and more

464 * When Claude references images (for example, `[Image #1]`), `Cmd+Click` (Mac) or `Ctrl+Click` (Windows/Linux) the link to open the image in your default viewer

457</Tip>465</Tip>

458 466

459***467***

501 509

502## Use extended thinking (thinking mode)510## Use extended thinking (thinking mode)

503 511

504[Extended thinking](https://docs.claude.com/en/docs/build-with-claude/extended-thinking) reserves a portion of the total output token budget for Claude to reason through complex problems step-by-step. This reasoning is visible in verbose mode, which you can toggle on with `Ctrl+O`.512[Extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) is enabled by default, giving Claude space to reason through complex problems step-by-step before responding. This reasoning is visible in verbose mode, which you can toggle on with `Ctrl+O`.

513

514Additionally, Opus 4.6 introduces adaptive reasoning: instead of a fixed thinking token budget, the model dynamically allocates thinking based on your [effort level](/en/model-config#adjust-effort-level) setting. Extended thinking and adaptive reasoning work together to give you control over how deeply Claude reasons before responding.

505 515

506Extended thinking is particularly valuable for complex architectural decisions, challenging bugs, multi-step implementation planning, and evaluating tradeoffs between different approaches. It provides more space for exploring multiple solutions, analyzing edge cases, and self-correcting mistakes.516Extended thinking is particularly valuable for complex architectural decisions, challenging bugs, multi-step implementation planning, and evaluating tradeoffs between different approaches.

507 517

508<Note>518<Note>

509 Sonnet 4.5 and Opus 4.5 have thinking enabled by default. All other models have thinking disabled by default. Use `/model` to view or switch your current model.519 Phrases like "think", "think hard", "ultrathink", and "think more" are interpreted as regular prompt instructions and don't allocate thinking tokens.

510</Note>520</Note>

511 521

512You can configure thinking mode for Claude Code in two ways:522### Configure thinking mode

~~513~~

514| Scope | How to enable | Details |

515| --------------------------------- | ------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------ |

516| **Global default** | Use `/config` to toggle thinking mode on | Sets your default across all projects.<br />Saved as `alwaysThinkingEnabled` in `~/.claude/settings.json` |

517| **Environment variable override** | Set [`MAX_THINKING_TOKENS`](/en/settings#environment-variables) environment variable | When set, applies a custom token budget to all requests, overriding your thinking mode configuration. Example: `export MAX_THINKING_TOKENS=1024` |

~~518~~

519### Per-request thinking with `ultrathink`

~~520~~

521You can include `ultrathink` as a keyword in your message to enable thinking for a single request:

522 523

523```524Thinking is enabled by default, but you can adjust or disable it.

524> ultrathink: design a caching layer for our API

525```

~~526~~

527Note that `ultrathink` both allocates the thinking budget AND semantically signals to Claude to reason more thoroughly, which may result in deeper thinking than necessary for your task.

528 525

529The `ultrathink` keyword only works when `MAX_THINKING_TOKENS` is not set. When `MAX_THINKING_TOKENS` is configured, it takes priority and controls the thinking budget for all requests.526| Scope | How to configure | Details |

~~530~~ 527| ---------------------- | ------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------- |

531Other phrases like "think", "think hard", and "think more" are interpreted as regular prompt instructions and don't allocate thinking tokens.528| **Effort level** | Adjust in `/model` or set [`CLAUDE_CODE_EFFORT_LEVEL`](/en/settings#environment-variables) | Control thinking depth for Opus 4.6: low, medium, high (default). See [Adjust effort level](/en/model-config#adjust-effort-level) |

529| **Toggle shortcut** | Press `Option+T` (macOS) or `Alt+T` (Windows/Linux) | Toggle thinking on/off for the current session (all models). May require [terminal configuration](/en/terminal-config) to enable Option key shortcuts |

530| **Global default** | Use `/config` to toggle thinking mode | Sets your default across all projects (all models).<br />Saved as `alwaysThinkingEnabled` in `~/.claude/settings.json` |

531| **Limit token budget** | Set [`MAX_THINKING_TOKENS`](/en/settings#environment-variables) environment variable | Limit the thinking budget to a specific number of tokens (ignored on Opus 4.6 unless set to 0). Example: `export MAX_THINKING_TOKENS=10000` |

532 532

533To view Claude's thinking process, press `Ctrl+O` to toggle verbose mode and see the internal reasoning displayed as gray italic text.533To view Claude's thinking process, press `Ctrl+O` to toggle verbose mode and see the internal reasoning displayed as gray italic text.

534 534

535See the [token budget section below](#how-extended-thinking-token-budgets-work) for detailed budget information and cost implications.535### How extended thinking works

~~536~~

537### How extended thinking token budgets work

~~538~~

539Extended thinking uses a **token budget** that controls how much internal reasoning Claude can perform before responding.

540 536

541A larger thinking token budget provides:537Extended thinking controls how much internal reasoning Claude performs before responding. More thinking provides more space to explore solutions, analyze edge cases, and self-correct mistakes.

542 538

543* More space to explore multiple solution approaches step-by-step539**With Opus 4.6**, thinking uses adaptive reasoning: the model dynamically allocates thinking tokens based on the [effort level](/en/model-config#adjust-effort-level) you select (low, medium, high). This is the recommended way to tune the tradeoff between speed and reasoning depth.

544* Room to analyze edge cases and evaluate tradeoffs thoroughly

545* Ability to revise reasoning and self-correct mistakes

546 540

547Token budgets for thinking mode:541**With other models**, thinking uses a fixed budget of up to 31,999 tokens from your output budget. You can limit this with the [`MAX_THINKING_TOKENS`](/en/settings#environment-variables) environment variable, or disable thinking entirely via `/config` or the `Option+T`/`Alt+T` toggle.

548 542

549* When thinking is **enabled** (via `/config` or `ultrathink`), Claude can use up to **31,999 tokens** from your output budget for internal reasoning543`MAX_THINKING_TOKENS` is ignored when using Opus 4.6, since adaptive reasoning controls thinking depth instead. The one exception: setting `MAX_THINKING_TOKENS=0` still disables thinking entirely on any model.

550* When thinking is **disabled**, Claude uses **0 tokens** for thinking

~~551~~

552**Custom token budgets:**

~~553~~

554* You can set a custom thinking token budget using the [`MAX_THINKING_TOKENS` environment variable](/en/settings#environment-variables)

555* This takes highest priority and overrides the default 31,999 token budget

556* See the [extended thinking documentation](https://docs.claude.com/en/docs/build-with-claude/extended-thinking) for valid token ranges

557 544

558<Warning>545<Warning>

559 You're charged for all thinking tokens used, even though Claude 4 models show summarized thinking546 You're charged for all thinking tokens used, even though Claude 4 models show summarized thinking

567 554

568* `claude --continue` continues the most recent conversation in the current directory555* `claude --continue` continues the most recent conversation in the current directory

569* `claude --resume` opens a conversation picker or resumes by name556* `claude --resume` opens a conversation picker or resumes by name

557* `claude --from-pr 123` resumes sessions linked to a specific pull request

570 558

571From inside an active session, use `/resume` to switch to a different conversation.559From inside an active session, use `/resume` to switch to a different conversation.

572 560

635 Tips:623 Tips:

636 624

637 * **Name sessions early**: Use `/rename` when starting work on a distinct task—it's much easier to find "payment-integration" than "explain this function" later625 * **Name sessions early**: Use `/rename` when starting work on a distinct task—it's much easier to find "payment-integration" than "explain this function" later

638 * Use `--continue` for quick access to your most recent conversation626 * Use `--continue` for quick access to your most recent conversation in the current directory

639 * Use `--resume session-name` when you know which session you need627 * Use `--resume session-name` when you know which session you need

640 * Use `--resume` (without a name) when you need to browse and select628 * Use `--resume` (without a name) when you need to browse and select

641 * For scripts, use `claude --continue --print "prompt"` to resume in non-interactive mode629 * For scripts, use `claude --continue --print "prompt"` to resume in non-interactive mode

719 * Other languages: Following your project's standard setup process707 * Other languages: Following your project's standard setup process

720</Tip>708</Tip>

721 709

710For automated coordination of parallel sessions with shared tasks and messaging, see [agent teams](/en/agent-teams).

711

722***712***

723 713

724## Use Claude as a unix-style utility714## Use Claude as a unix-style utility

806 796

807***797***

808 798

809## Create custom slash commands

~~810~~

811Claude Code supports custom slash commands that you can create to quickly execute specific prompts or tasks.

~~812~~

813For more details, see the [Slash commands](/en/slash-commands) reference page.

~~814~~

815### Create project-specific commands

~~816~~

817Suppose you want to create reusable slash commands for your project that all team members can use.

~~818~~

819<Steps>

820 <Step title="Create a commands directory in your project">

821 ```bash theme={null}

822 mkdir -p .claude/commands

823 ```

824 </Step>

~~825~~

826 <Step title="Create a Markdown file for each command">

827 ```bash theme={null}

828 echo "Analyze the performance of this code and suggest three specific optimizations:" > .claude/commands/optimize.md

829 ```

830 </Step>

~~831~~

832 <Step title="Use your custom command in Claude Code">

833 ```

834 > /optimize

835 ```

836 </Step>

837</Steps>

~~838~~

839<Tip>

840 Tips:

~~841~~

842 * Command names are derived from the filename (for example, `optimize.md` becomes `/optimize`)

843 * You can organize commands in subdirectories (for example, `.claude/commands/frontend/component.md` creates `/component` with "(project:frontend)" shown in the description)

844 * Project commands are available to everyone who clones the repository

845 * The Markdown file content becomes the prompt sent to Claude when the command is invoked

846</Tip>

~~847~~

848### Add command arguments with \$ARGUMENTS

~~849~~

850Suppose you want to create flexible slash commands that can accept additional input from users.

~~851~~

852<Steps>

853 <Step title="Create a command file with the $ARGUMENTS placeholder">

854 ```bash theme={null}

855 echo 'Find and fix issue #$ARGUMENTS. Follow these steps: 1.

856 Understand the issue described in the ticket 2. Locate the relevant code in

857 our codebase 3. Implement a solution that addresses the root cause 4. Add

858 appropriate tests 5. Prepare a concise PR description' >

859 .claude/commands/fix-issue.md

860 ```

861 </Step>

~~862~~

863 <Step title="Use the command with an issue number">

864 In your Claude session, use the command with arguments.

~~865~~

866 ```

867 > /fix-issue 123

868 ```

~~869~~

870 This replaces \$ARGUMENTS with "123" in the prompt.

871 </Step>

872</Steps>

~~873~~

874<Tip>

875 Tips:

~~876~~

877 * The \$ARGUMENTS placeholder is replaced with any text that follows the command

878 * You can position \$ARGUMENTS anywhere in your command template

879 * Other useful applications: generating test cases for specific functions, creating documentation for components, reviewing code in particular files, or translating content to specified languages

880</Tip>

~~881~~

882### Create personal slash commands

~~883~~

884Suppose you want to create personal slash commands that work across all your projects.

~~885~~

886<Steps>

887 <Step title="Create a commands directory in your home folder">

888 ```bash theme={null}

889 mkdir -p ~/.claude/commands

890 ```

891 </Step>

~~892~~

893 <Step title="Create a Markdown file for each command">

894 ```bash theme={null}

895 echo "Review this code for security vulnerabilities, focusing on:" >

896 ~/.claude/commands/security-review.md

897 ```

898 </Step>

~~899~~

900 <Step title="Use your personal custom command">

901 ```

902 > /security-review

903 ```

904 </Step>

905</Steps>

~~906~~

907<Tip>

908 Tips:

~~909~~

910 * Personal commands show "(user)" in their description when listed with `/help`

911 * Personal commands are only available to you and not shared with your team

912 * Personal commands work across all your projects

913 * You can use these for consistent workflows across different codebases

914</Tip>

~~915~~

916***

~~917~~

918## Ask Claude about its capabilities799## Ask Claude about its capabilities

919 800

920Claude has built-in access to its documentation and can answer questions about its own features and limitations.801Claude has built-in access to its documentation and can answer questions about its own features and limitations.

930```811```

931 812

932```813```

933> what slash commands are available?814> what skills are available?

934```815```

935 816

936```817```

961 842

962## Next steps843## Next steps

963 844

964<Card title="Claude Code reference implementation" icon="code" href="https://github.com/anthropics/claude-code/tree/main/.devcontainer">845<CardGroup cols={2}>

965 Clone our development container reference implementation.846 <Card title="Best practices" icon="lightbulb" href="/en/best-practices">

966</Card>847 Patterns for getting the most out of Claude Code

848 </Card>

967 849

850 <Card title="How Claude Code works" icon="gear" href="/en/how-claude-code-works">

851 Understand the agentic loop and context management

852 </Card>

968 853

854 <Card title="Extend Claude Code" icon="puzzle-piece" href="/en/features-overview">

855 Add skills, hooks, MCP, subagents, and plugins

856 </Card>

969 857

970> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt858 <Card title="Reference implementation" icon="code" href="https://github.com/anthropics/claude-code/tree/main/.devcontainer">

859 Clone our development container reference implementation

860 </Card>

861</CardGroup>

costs.md +125 −58

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Manage costs effectively5# Manage costs effectively

2 6

~~3> Learn how to track and optimize token usage and costs when using Claude Code.~~7> Track token usage, set team spend limits, and reduce Claude Code costs with context management, model selection, extended thinking settings, and preprocessing hooks.

4 8

~~5Claude Code consumes tokens for each interaction. The average cost is \$6 per developer per day, with daily costs remaining below \$12 for 90% of users.~~9Claude Code consumes tokens for each interaction. Costs vary based on codebase size, query complexity, and conversation length. The average cost is \$6 per developer per day, with daily costs remaining below \$12 for 90% of users.

6 10

7For team usage, Claude Code charges by API token consumption. On average, Claude Code costs \~\$100-200/developer per month with Sonnet 4.5 though there is large variance depending on how many instances users are running and whether they're using it in automation.11For team usage, Claude Code charges by API token consumption. On average, Claude Code costs \~\$100-200/developer per month with Sonnet 4.5 though there is large variance depending on how many instances users are running and whether they're using it in automation.

8 12

13This page covers how to [track your costs](#track-your-costs), [manage costs for teams](#managing-costs-for-teams), and [reduce token usage](#reduce-token-usage).

9## Track your costs15## Track your costs

10 16

11### Using the `/cost` command17### Using the `/cost` command

12 18

13<Note>19<Note>

~~14 The `/cost` command is not intended for Claude Max and Pro subscribers.~~20 The `/cost` command shows API token usage and is intended for API users. Claude Max and Pro subscribers have usage included in their subscription, so `/cost` data isn't relevant for billing purposes. Subscribers can use `/stats` to view usage patterns.

15</Note>21</Note>

16 22

17The `/cost` command provides detailed token usage statistics for your current session:23The `/cost` command provides detailed token usage statistics for your current session:

23Total code changes: 0 lines added, 0 lines removed29Total code changes: 0 lines added, 0 lines removed

24```30```

25 31

~~26### Additional tracking options~~32## Managing costs for teams

27 33

28Check [historical usage](https://support.claude.com/en/articles/9534590-cost-and-usage-reporting-in-console) in the Claude Console (requires Admin or Billing role) and set [workspace spend limits](https://support.claude.com/en/articles/9796807-creating-and-managing-workspaces) for the Claude Code workspace (requires Admin role).34When using Claude API, you can [set workspace spend limits](https://platform.claude.com/docs/en/build-with-claude/workspaces#workspace-limits) on the total Claude Code workspace spend. Admins can [view cost and usage reporting](https://platform.claude.com/docs/en/build-with-claude/workspaces#usage-and-cost-tracking) in the Console.

29 35

30<Note>36<Note>

31 When you first authenticate Claude Code with your Claude Console account, a workspace called "Claude Code" is automatically created for you. This workspace provides centralized cost tracking and management for all Claude Code usage in your organization. You cannot create API keys for this workspace - it is exclusively for Claude Code authentication and usage.37 When you first authenticate Claude Code with your Claude Console account, a workspace called "Claude Code" is automatically created for you. This workspace provides centralized cost tracking and management for all Claude Code usage in your organization. You cannot create API keys for this workspace; it is exclusively for Claude Code authentication and usage.

32</Note>38</Note>

33 39

~~34## Managing costs for teams~~40On Bedrock, Vertex, and Foundry, Claude Code does not send metrics from your cloud. To get cost metrics, several large enterprises reported using [LiteLLM](/en/llm-gateway#litellm-configuration), which is an open-source tool that helps companies [track spend by key](https://docs.litellm.ai/docs/proxy/virtual_keys#tracking-spend). This project is unaffiliated with Anthropic and has not been audited for security.

36When using Claude API, you can limit the total Claude Code workspace spend. To configure, [follow these instructions](https://support.claude.com/en/articles/9796807-creating-and-managing-workspaces). Admins can view cost and usage reporting by [following these instructions](https://support.claude.com/en/articles/9534590-cost-and-usage-reporting-in-console).

38On Bedrock and Vertex, Claude Code does not send metrics from your cloud. In order to get cost metrics, several large enterprises reported using [LiteLLM](/en/third-party-integrations#litellm), which is an open-source tool that helps companies [track spend by key](https://docs.litellm.ai/docs/proxy/virtual_keys#tracking-spend). This project is unaffiliated with Anthropic and we have not audited its security.

39 41

40### Rate limit recommendations42### Rate limit recommendations

41 43

52 54

53For example, if you have 200 users, you might request 20k TPM for each user, or 4 million total TPM (200\*20,000 = 4 million).55For example, if you have 200 users, you might request 20k TPM for each user, or 4 million total TPM (200\*20,000 = 4 million).

54 56

55The TPM per user decreases as team size grows because we expect fewer users to use Claude Code concurrently in larger organizations. These rate limits apply at the organization level, not per individual user, which means individual users can temporarily consume more than their calculated share when others aren't actively using the service.57The TPM per user decreases as team size grows because fewer users tend to use Claude Code concurrently in larger organizations. These rate limits apply at the organization level, not per individual user, which means individual users can temporarily consume more than their calculated share when others aren't actively using the service.

56 58

57<Note>59<Note>

58 If you anticipate scenarios with unusually high concurrent usage (such as live training sessions with large groups), you may need higher TPM allocations per user.60 If you anticipate scenarios with unusually high concurrent usage (such as live training sessions with large groups), you may need higher TPM allocations per user.

59</Note>61</Note>

60 62

63### Agent team token costs

65[Agent teams](/en/agent-teams) spawn multiple Claude Code instances, each with its own context window. Token usage scales with the number of active teammates and how long each one runs.

67To keep agent team costs manageable:

69* Use Sonnet for teammates. It balances capability and cost for coordination tasks.

70* Keep teams small. Each teammate runs its own context window, so token usage is roughly proportional to team size.

71* Keep spawn prompts focused. Teammates load CLAUDE.md, MCP servers, and skills automatically, but everything in the spawn prompt adds to their context from the start.

72* Clean up teams when work is done. Active teammates continue consuming tokens even if idle.

73* Agent teams are disabled by default. Set `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1` in your [settings.json](/en/settings) or environment to enable them. See [enable agent teams](/en/agent-teams#enable-agent-teams).

61## Reduce token usage75## Reduce token usage

62 76

~~63* **Compact conversations:**~~77Token costs scale with context size: the more context Claude processes, the more tokens you use. Claude Code automatically optimizes costs through prompt caching (which reduces costs for repeated content like system prompts) and auto-compaction (which summarizes conversation history when approaching context limits).

64 78

~~65 * Claude uses auto-compact by default when context exceeds 95% capacity~~79The following strategies help you keep context small and reduce per-message costs.

~~66 * Toggle auto-compact: Run `/config` and navigate to "Auto-compact enabled"~~

~~67 * Use `/compact` manually when context gets large~~

~~68 * Add custom instructions: `/compact Focus on code samples and API usage`~~

~~69 * Customize compaction by adding to CLAUDE.md:~~

70 80

~~71 ```markdown theme={null}~~81### Manage context proactively

~~72 # Summary instructions~~

73 82

~~74 When you are using compact, please focus on test output and code changes~~83Use `/cost` to check your current token usage, or [configure your status line](/en/statusline#context-window-usage) to display it continuously.

~~75 ```~~

76 84

~~77* **Write specific queries:** Avoid vague requests that trigger unnecessary scanning~~85* **Clear between tasks**: Use `/clear` to start fresh when switching to unrelated work. Stale context wastes tokens on every subsequent message. Use `/rename` before clearing so you can easily find the session later, then `/resume` to return to it.

86* **Add custom compaction instructions**: `/compact Focus on code samples and API usage` tells Claude what to preserve during summarization.

78 87

~~79* **Break down complex tasks:** Split large tasks into focused interactions~~88You can also customize compaction behavior in your CLAUDE.md:

80 89

~~81* **Clear history between tasks:** Use `/clear` to reset context~~90```markdown theme={null}

91# Compact instructions

82 92

~~83Costs can vary significantly based on:~~93When you are using compact, please focus on test output and code changes

94```

84 95

~~85* Size of codebase being analyzed~~96### Choose the right model

~~86* Complexity of queries~~

~~87* Number of files being searched or modified~~

~~88* Length of conversation history~~

~~89* Frequency of compacting conversations~~

90 97

~~91## Background token usage~~98Sonnet handles most coding tasks well and costs less than Opus. Reserve Opus for complex architectural decisions or multi-step reasoning. Use `/model` to switch models mid-session, or set a default in `/config`. For simple subagent tasks, specify `model: haiku` in your [subagent configuration](/en/sub-agents#choose-a-model).

92 99

~~93Claude Code uses tokens for some background functionality even when idle:~~100### Reduce MCP server overhead

94 101

~~95* **Conversation summarization**: Background jobs that summarize previous conversations for the `claude --resume` feature~~102Each MCP server adds tool definitions to your context, even when idle. Run `/context` to see what's consuming space.

~~96* **Command processing**: Some commands like `/cost` may generate requests to check status~~

97 103

~~98These background processes consume a small amount of tokens (typically under \$0.04 per session) even without active interaction.~~104* **Prefer CLI tools when available**: Tools like `gh`, `aws`, `gcloud`, and `sentry-cli` are more context-efficient than MCP servers because they don't add persistent tool definitions. Claude can run CLI commands directly without the overhead.

105* **Disable unused servers**: Run `/mcp` to see configured servers and disable any you're not actively using.

106* **Tool search is automatic**: When MCP tool descriptions exceed 10% of your context window, Claude Code automatically defers them and loads tools on-demand via [tool search](/en/mcp#scale-with-mcp-tool-search). Since deferred tools only enter context when actually used, a lower threshold means fewer idle tool definitions consuming space. Set a lower threshold with `ENABLE_TOOL_SEARCH=auto:<N>` (for example, `auto:5` triggers when tools exceed 5% of your context window).

99 107

100## Tracking version changes and updates108### Install code intelligence plugins for typed languages

101 109

102### Current version information110[Code intelligence plugins](/en/discover-plugins#code-intelligence) give Claude precise symbol navigation instead of text-based search, reducing unnecessary file reads when exploring unfamiliar code. A single "go to definition" call replaces what might otherwise be a grep followed by reading multiple candidate files. Installed language servers also report type errors automatically after edits, so Claude catches mistakes without running a compiler.

103 111

104To check your current Claude Code version and installation details:112### Offload processing to hooks and skills

105 113

106```bash theme={null}114Custom [hooks](/en/hooks) can preprocess data before Claude sees it. Instead of Claude reading a 10,000-line log file to find errors, a hook can grep for `ERROR` and return only matching lines, reducing context from tens of thousands of tokens to hundreds.

107claude doctor

108```

109 115

110This command shows your version, installation type, and system information.116A [skill](/en/skills) can give Claude domain knowledge so it doesn't have to explore. For example, a "codebase-overview" skill could describe your project's architecture, key directories, and naming conventions. When Claude invokes the skill, it gets this context immediately instead of spending tokens reading multiple files to understand the structure.

111 117

112### Understanding changes in Claude Code behavior118For example, this PreToolUse hook filters test output to show only failures:

113 119

114Claude Code regularly receives updates that may change how features work, including cost reporting:120<Tabs>

121 <Tab title="settings.json">

122 Add this to your [settings.json](/en/settings#settings-files) to run the hook before every Bash command:

115 123

116* **Version tracking**: Use `claude doctor` to see your current version124 ```json theme={null}

117* **Behavior changes**: Features like `/cost` may display information differently across versions125 {

118* **Documentation access**: Claude always has access to the latest documentation, which can help explain current feature behavior126 "hooks": {

127 "PreToolUse": [

128 {

129 "matcher": "Bash",

130 "hooks": [

131 {

132 "type": "command",

133 "command": "~/.claude/hooks/filter-test-output.sh"

134 }

135 ]

136 }

137 ]

138 }

139 }

140 ```

141 </Tab>

142

143 <Tab title="filter-test-output.sh">

144 The hook calls this script, which checks if the command is a test runner and modifies it to show only failures:

145

146 ```bash theme={null}

147 #!/bin/bash

148 input=$(cat)

149 cmd=$(echo "$input" | jq -r '.tool_input.command')

150

151 # If running tests, filter to show only failures

152 if [[ "$cmd" =~ ^(npm test|pytest|go test) ]]; then

153 filtered_cmd="$cmd 2>&1 | grep -A 5 -E '(FAIL|ERROR|error:)' | head -100"

154 echo "{\"hookSpecificOutput\":{\"hookEventName\":\"PreToolUse\",\"permissionDecision\":\"allow\",\"updatedInput\":{\"command\":\"$filtered_cmd\"}}}"

155 else

156 echo "{}"

157 fi

158 ```

159 </Tab>

160</Tabs>

119 161

120### When cost reporting changes162### Move instructions from CLAUDE.md to skills

121 163

122If you notice changes in how costs are displayed (such as the `/cost` command showing different information):164Your [CLAUDE.md](/en/memory) file is loaded into context at session start. If it contains detailed instructions for specific workflows (like PR reviews or database migrations), those tokens are present even when you're doing unrelated work. [Skills](/en/skills) load on-demand only when invoked, so moving specialized instructions into skills keeps your base context smaller. Aim to keep CLAUDE.md under \~500 lines by including only essentials.

123 165

1241. **Verify your version**: Run `claude doctor` to confirm your current version166### Adjust extended thinking

1252. **Consult documentation**: Ask Claude directly about current feature behavior, as it has access to up-to-date documentation

1263. **Contact support**: For specific billing questions, contact Anthropic support through your Console account

127 167

128<Note>168Extended thinking is enabled by default with a budget of 31,999 tokens because it significantly improves performance on complex planning and reasoning tasks. However, thinking tokens are billed as output tokens, so for simpler tasks where deep reasoning isn't needed, you can reduce costs by lowering the [effort level](/en/model-config#adjust-effort-level) in `/model` for Opus 4.6, disabling thinking in `/config`, or lowering the budget (for example, `MAX_THINKING_TOKENS=8000`).

129 For team deployments, we recommend starting with a small pilot group to169

130 establish usage patterns before wider rollout.170### Delegate verbose operations to subagents

131</Note>171

172Running tests, fetching documentation, or processing log files can consume significant context. Delegate these to [subagents](/en/sub-agents#isolate-high-volume-operations) so the verbose output stays in the subagent's context while only a summary returns to your main conversation.

173

174### Manage agent team costs

175

176Agent teams use approximately 7x more tokens than standard sessions when teammates run in plan mode, because each teammate maintains its own context window and runs as a separate Claude instance. Keep team tasks small and self-contained to limit per-teammate token usage. See [agent teams](/en/agent-teams) for details.

177

178### Write specific prompts

179

180Vague requests like "improve this codebase" trigger broad scanning. Specific requests like "add input validation to the login function in auth.ts" let Claude work efficiently with minimal file reads.

181

182### Work efficiently on complex tasks

183

184For longer or more complex work, these habits help avoid wasted tokens from going down the wrong path:

132 185

186* **Use plan mode for complex tasks**: Press Shift+Tab to enter [plan mode](/en/common-workflows#use-plan-mode-for-safe-code-analysis) before implementation. Claude explores the codebase and proposes an approach for your approval, preventing expensive re-work when the initial direction is wrong.

187* **Course-correct early**: If Claude starts heading the wrong direction, press Escape to stop immediately. Use `/rewind` or double-tap Escape to restore conversation and code to a previous checkpoint.

188* **Give verification targets**: Include test cases, paste screenshots, or define expected output in your prompt. When Claude can verify its own work, it catches issues before you need to request fixes.

189* **Test incrementally**: Write one file, test it, then continue. This catches issues early when they're cheap to fix.

190

191## Background token usage

192

193Claude Code uses tokens for some background functionality even when idle:

194

195* **Conversation summarization**: Background jobs that summarize previous conversations for the `claude --resume` feature

196* **Command processing**: Some commands like `/cost` may generate requests to check status

197

198These background processes consume a small amount of tokens (typically under \$0.04 per session) even without active interaction.

133 199

200## Understanding changes in Claude Code behavior

134 201

135> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt202Claude Code regularly receives updates that may change how features work, including cost reporting. Run `claude --version` to check your current version. For specific billing questions, contact Anthropic support through your [Console account](https://platform.claude.com/login). For team deployments, start with a small pilot group to establish usage patterns before wider rollout.

data-usage.md +24 −30

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Data usage5# Data usage

2 6

3> Learn about Anthropic's data usage policies for Claude7> Learn about Anthropic's data usage policies for Claude

7### Data training policy11### Data training policy

8 12

9**Consumer users (Free, Pro, and Max plans)**:13**Consumer users (Free, Pro, and Max plans)**:

~~10Starting August 28, 2025, we're giving you the choice to allow your data to be used to improve future Claude models.~~14We give you the choice to allow your data to be used to improve future Claude models. We will train new models using data from Free, Pro, and Max accounts when this setting is on (including when you use Claude Code from these accounts).

~~12We will train new models using data from Free, Pro, and Max accounts when this setting is on (including when you use Claude Code from these accounts).~~

~~14* If you're a current user, you can select your preference now and your selection will immediately go into effect.~~

~~15 This setting will only apply to new or resumed chats and coding sessions on Claude. Previous chats with no additional activity will not be used for model training.~~

~~16* You have until October 8, 2025 to make your selection.~~

~~17 If you're a new user, you can pick your setting for model training during the signup process.~~

~~18 You can change your selection at any time in your Privacy Settings.~~

19 15

20**Commercial users**: (Team and Enterprise plans, API, 3rd-party platforms, and Claude Gov) maintain existing policies: Anthropic does not train generative models using code or prompts sent to Claude Code under commercial terms, unless the customer has chosen to provide their data to us for model improvement (for example, the [Developer Partner Program](https://support.claude.com/en/articles/11174108-about-the-development-partner-program)).16**Commercial users**: (Team and Enterprise plans, API, 3rd-party platforms, and Claude Gov) maintain existing policies: Anthropic does not train generative models using code or prompts sent to Claude Code under commercial terms, unless the customer has chosen to provide their data to us for model improvement (for example, the [Developer Partner Program](https://support.claude.com/en/articles/11174108-about-the-development-partner-program)).

21 17

31 27

32When you see the "How is Claude doing this session?" prompt in Claude Code, responding to this survey (including selecting "Dismiss"), only your numeric rating (1, 2, 3, or dismiss) is recorded. We do not collect or store any conversation transcripts, inputs, outputs, or other session data as part of this survey. Unlike thumbs up/down feedback or `/bug` reports, this session quality survey is a simple product satisfaction metric. Your responses to this survey do not impact your data training preferences and cannot be used to train our AI models.28When you see the "How is Claude doing this session?" prompt in Claude Code, responding to this survey (including selecting "Dismiss"), only your numeric rating (1, 2, 3, or dismiss) is recorded. We do not collect or store any conversation transcripts, inputs, outputs, or other session data as part of this survey. Unlike thumbs up/down feedback or `/bug` reports, this session quality survey is a simple product satisfaction metric. Your responses to this survey do not impact your data training preferences and cannot be used to train our AI models.

33 29

30To disable these surveys, set `CLAUDE_CODE_DISABLE_FEEDBACK_SURVEY=1`. The survey is also automatically disabled when using third-party providers (Bedrock, Vertex, Foundry) or when telemetry is disabled.

34### Data retention32### Data retention

35 33

36Anthropic retains Claude Code data based on your account type and preferences.34Anthropic retains Claude Code data based on your account type and preferences.

51 49

52For full details, please review our [Commercial Terms of Service](https://www.anthropic.com/legal/commercial-terms) (for Team, Enterprise, and API users) or [Consumer Terms](https://www.anthropic.com/legal/consumer-terms) (for Free, Pro, and Max users) and [Privacy Policy](https://www.anthropic.com/legal/privacy).50For full details, please review our [Commercial Terms of Service](https://www.anthropic.com/legal/commercial-terms) (for Team, Enterprise, and API users) or [Consumer Terms](https://www.anthropic.com/legal/consumer-terms) (for Free, Pro, and Max users) and [Privacy Policy](https://www.anthropic.com/legal/privacy).

53 51

~~54## Data flow and dependencies~~52## Data access

54For all first party users, you can learn more about what data is logged for [local Claude Code](#local-claude-code-data-flow-and-dependencies) and [remote Claude Code](#cloud-execution-data-flow-and-dependencies). Note for remote Claude Code, Claude accesses the repository where you initiate your Claude Code session. Claude does not access repositories that you have connected but have not started a session in.

56## Local Claude Code: Data flow and dependencies

55 57

56The diagram below shows how Claude Code connects to external services during installation and normal operation. Solid lines indicate required connections, while dashed lines represent optional or user-initiated data flows.58The diagram below shows how Claude Code connects to external services during installation and normal operation. Solid lines indicate required connections, while dashed lines represent optional or user-initiated data flows.

57 59

61 63

62Claude Code is built on Anthropic's APIs. For details regarding our API's security controls, including our API logging procedures, please refer to compliance artifacts offered in the [Anthropic Trust Center](https://trust.anthropic.com).64Claude Code is built on Anthropic's APIs. For details regarding our API's security controls, including our API logging procedures, please refer to compliance artifacts offered in the [Anthropic Trust Center](https://trust.anthropic.com).

63 65

~~64### Cloud execution~~66### Cloud execution: Data flow and dependencies

~~66<Note>~~

~~67 The above data flow diagram and description applies to Claude Code CLI running locally on your machine. For cloud-based sessions using Claude Code on the web, see the section below.~~

~~68</Note>~~

69 67

70When using [Claude Code on the web](/en/claude-code-on-the-web), sessions run in Anthropic-managed virtual machines instead of locally. In cloud environments:68When using [Claude Code on the web](/en/claude-code-on-the-web), sessions run in Anthropic-managed virtual machines instead of locally. In cloud environments:

71 69

~~72* **Code storage**: Your repository is cloned to an isolated VM and automatically deleted after session completion~~70* **Code and data storage:** Your repository is cloned to an isolated VM. Code and session data are subject to the retention and usage policies for your account type (see Data retention section above)

~~73* **Credentials**: GitHub authentication is handled through a secure proxy; your GitHub credentials never enter the sandbox~~71* **Credentials:** GitHub authentication is handled through a secure proxy; your GitHub credentials never enter the sandbox

~~74* **Network traffic**: All outbound traffic goes through a security proxy for audit logging and abuse prevention~~72* **Network traffic:** All outbound traffic goes through a security proxy for audit logging and abuse prevention

~~75* **Data retention**: Code and session data are subject to the retention and usage policies for your account type~~73* **Session data:** Prompts, code changes, and outputs follow the same data policies as local Claude Code usage

~~76* **Session data**: Prompts, code changes, and outputs follow the same data policies as local Claude Code usage~~

77 74

78For security details about cloud execution, see [Security](/en/security#cloud-execution-security).75For security details about cloud execution, see [Security](/en/security#cloud-execution-security).

79 76

87 84

88## Default behaviors by API provider85## Default behaviors by API provider

89 86

90By default, we disable all non-essential traffic (including error reporting, telemetry, and bug reporting functionality) when using Bedrock or Vertex. You can also opt out of all of these at once by setting the `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` environment variable. Here are the full default behaviors:87By default, we disable all non-essential traffic (including error reporting, telemetry, bug reporting functionality, and session quality surveys) when using Bedrock, Vertex, or Foundry. You can also opt out of all of these at once by setting the `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` environment variable. Here are the full default behaviors:

91 88

93| ------------------------------- | -------------------------------------------------------- | ----------------------------------------------------- | ------------------------------------------------------ |90| ------------------------------- | -------------------------------------------------------------------- | ----------------------------------------------------- | ------------------------------------------------------ | ------------------------------------------------------ |

94| **Session quality surveys** | Default on.<br />`CLAUDE_CODE_DISABLE_FEEDBACK_SURVEY=1` to disable. | Default off.<br />`CLAUDE_CODE_USE_VERTEX` must be 1. | Default off.<br />`CLAUDE_CODE_USE_BEDROCK` must be 1. | Default off.<br />`CLAUDE_CODE_USE_FOUNDRY` must be 1. |

97 95

98All environment variables can be checked into `settings.json` ([read more](/en/settings)).96All environment variables can be checked into `settings.json` ([read more](/en/settings)).

~~100~~

~~101~~

102> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

desktop.md +259 −59

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code on desktop5# Claude Code on desktop

2 6

3> Run Claude Code tasks locally or on secure cloud infrastructure with the Claude desktop app7> Run Claude Code tasks locally or on secure cloud infrastructure with the Claude desktop app

4 8

5<img src="https://mintcdn.com/claude-code/zEGbGSbinVtT3BLw/images/desktop-interface.png?fit=max&auto=format&n=zEGbGSbinVtT3BLw&q=85&s=c4e9dc9737b437d36ab253b75a1cc595" alt="Claude Code on desktop interface" data-og-width="4132" width="4132" data-og-height="2620" height="2620" data-path="images/desktop-interface.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/zEGbGSbinVtT3BLw/images/desktop-interface.png?w=280&fit=max&auto=format&n=zEGbGSbinVtT3BLw&q=85&s=b1a8421a544c3e8c78679fa1a7b56190 280w, https://mintcdn.com/claude-code/zEGbGSbinVtT3BLw/images/desktop-interface.png?w=560&fit=max&auto=format&n=zEGbGSbinVtT3BLw&q=85&s=79cf4ea0923098cc429198678ea50903 560w, https://mintcdn.com/claude-code/zEGbGSbinVtT3BLw/images/desktop-interface.png?w=840&fit=max&auto=format&n=zEGbGSbinVtT3BLw&q=85&s=14bcbcd569d179770fe656686ffbf6bf 840w, https://mintcdn.com/claude-code/zEGbGSbinVtT3BLw/images/desktop-interface.png?w=1100&fit=max&auto=format&n=zEGbGSbinVtT3BLw&q=85&s=b873274db1e9ff8585ba545032aa24a5 1100w, https://mintcdn.com/claude-code/zEGbGSbinVtT3BLw/images/desktop-interface.png?w=1650&fit=max&auto=format&n=zEGbGSbinVtT3BLw&q=85&s=25553dced783c3a8c2a1134a53295f7e 1650w, https://mintcdn.com/claude-code/zEGbGSbinVtT3BLw/images/desktop-interface.png?w=2500&fit=max&auto=format&n=zEGbGSbinVtT3BLw&q=85&s=9ad49e6468c2f87b1895093deeea7bb2 2500w" />9<Note>

10 Claude Code on desktop is currently in preview.

11</Note>

13Claude Code is an AI coding assistant that works directly with your codebase. Unlike Claude.ai chat, it can read your project files, edit code, run terminal commands, and understand how different parts of your code connect. You watch changes happen in real time.

15You can use Claude Code through the terminal ([CLI](/en/quickstart)) or through the desktop app described here. Both provide the same core capabilities. The desktop app adds a graphical interface and visual session management.

17<CardGroup cols={2}>

18 <Card title="New to Claude Code?" icon="rocket" href="#installation-and-setup">

19 Start here to install and make your first edit

20 </Card>

22 <Card title="Coming from the CLI?" icon="terminal" href="#how-desktop-relates-to-cli">

23 See what's shared and what's different

24 </Card>

25</CardGroup>

27The desktop app has three tabs:

29* **Chat**: A conversational interface for general questions and tasks (like Claude.ai)

30* **Cowork**: An autonomous agent that works on tasks in the background

31* **Code**: An AI coding assistant that reads and edits your project files directly

33This documentation covers the **Code** tab. For the chat interface, see the [Claude Desktop support articles](https://support.claude.com/en/collections/16163169-claude-desktop).

35## Installation and setup

37<Steps>

38 <Step title="Download the app">

39 Download Claude for your platform. You'll need an Anthropic account ([sign up at claude.ai](https://claude.ai) if you don't have one).

41 <CardGroup cols={2}>

42 <Card title="macOS" icon="apple" href="https://claude.ai/api/desktop/darwin/universal/dmg/latest/redirect?utm_source=claude_code&utm_medium=docs">

43 Universal build for Intel and Apple Silicon

44 </Card>

46 <Card title="Windows" icon="windows" href="https://claude.ai/api/desktop/win32/x64/exe/latest/redirect?utm_source=claude_code&utm_medium=docs">

47 For x64 processors

48 </Card>

49 </CardGroup>

51 For Windows ARM64, [download here](https://claude.ai/api/desktop/win32/arm64/exe/latest/redirect?utm_source=claude_code\&utm_medium=docs). Local sessions are not available on ARM64 devices, so use remote sessions instead.

53 Linux is not currently supported.

54 </Step>

56 <Step title="Open the app and sign in">

57 Launch Claude from your Applications folder (macOS) or Start menu (Windows). Sign in with your Anthropic account.

58 </Step>

60 <Step title="Select the Code tab">

61 Click the **Code** tab in the top left. If clicking Code prompts you to sign in online, complete the sign-in and restart the app.

62 </Step>

63</Steps>

65## Getting started

67If you already use the CLI, you can skip to [How Desktop relates to CLI](#how-desktop-relates-to-cli) for a quick overview of differences.

69<Steps>

70 <Step title="Choose a folder and environment">

71 Select **Local** to run Claude on your machine using your files directly. This is the best choice for getting started. Click **Select folder** and choose your project directory.

73 You can also run [remote sessions](/en/claude-code-on-the-web) that continue in the cloud even if you close the app.

74 </Step>

6 75

~~7## Claude Code on desktop (Preview)~~76 <Step title="Start a session">

77 Type what you want Claude to do:

8 78

~~9The Claude desktop app provides a native interface for running multiple Claude Code sessions on your local machine and seamless integration with Claude Code on the web.~~79 * "Find a TODO comment and fix it"

80 * "Add tests for the main function"

81 * "Create a CLAUDE.md with instructions for this codebase"

10 82

~~11## Features~~83 A **session** is a conversation with Claude about your code. Each session tracks its own context and changes, so you can work on multiple tasks without them interfering with each other.

84 </Step>

12 85

~~13Claude Code on desktop provides:~~86 <Step title="Review and accept changes">

87 By default, Code is in **Ask** mode, where Claude proposes changes and waits for your approval before applying them. You'll see:

14 88

~~15* **Parallel local sessions with `git` worktrees**: Run multiple Claude Code sessions simultaneously in the same repository, each with its own isolated `git` worktree~~89 1. **A diff view** showing exactly what will change in each file

~~16* **Include files listed in your `.gitignore` in your worktrees**: Automatically copy files in your `.gitignore`, like `.env`, to new worktrees using `.worktreeinclude`~~90 2. **Accept/Reject buttons** to approve or decline each change

~~17* **Launch Claude Code on the web**: Kick off secure cloud sessions directly from the desktop app~~91 3. **Real-time updates** as Claude works through your request

18 92

~~19## Installation~~93 If you reject a change, Claude will ask how you'd like to proceed differently. Your files aren't modified until you accept.

94 </Step>

95</Steps>

20 96

~~21Download and install the Claude Desktop app from [claude.ai/download](https://claude.ai/download)~~97The sections below cover commands, permission modes, parallel sessions, and ways to extend Claude Code with custom workflows and integrations.

99## What you can do

100

101Claude Code can edit files, run terminal commands, and understand how your code connects. Try prompts like:

102

103* `Fix the bug in the login function`

104* `Run the tests and fix any failures`

105* `How does the authentication flow work?`

106

107You can rename, resume, and archive sessions through the sidebar.

108

109### Choose a permission mode

110

111Control how Claude works using the mode selector next to the send button:

112

113* **Ask** (recommended for new users): Claude asks for your approval before each file edit or command. You see a diff view and can accept or reject each change.

114* **Code**: Claude auto-accepts file edits but still asks before running terminal commands. Use this when you trust file changes and want faster iteration.

115* **Plan**: Claude creates a detailed plan for your approval before making any changes. Good for complex tasks where you want to review the approach first.

116* **Act**: Claude runs without permission checks, automatically executing file edits and terminal commands. Only use this mode in trusted environments.

117

118<Warning title="Act mode">

119 Act runs in `bypassPermissions` mode, which disables all permission checks and should only be used in isolated environments like containers or VMs where Claude Code cannot cause damage. This mode is disabled by default. For personal accounts, enable it in [Claude Code personal settings](https://claude.ai/settings/claude-code). For Team and Enterprise plans, admins must enable it in [Claude Code admin settings](https://claude.ai/admin-settings/claude-code). Act mode does not persist across sessions.

120</Warning>

121

122To stop Claude mid-task, click the stop button.

123

124Remote sessions only support **Code** and **Plan** modes because they continue running in the background without requiring your active participation. See [permission modes](/en/permissions#permission-modes) for details on how these work internally.

125

126### Work in parallel with sessions

127

128Click **+ New session** in the sidebar to work on multiple tasks in parallel. For Git repositories, each session gets its own isolated copy of your project using worktrees, so changes in one session don't affect another until you commit them. Worktrees are stored in `~/.claude-worktrees/` by default.

22 129

23<Note>130<Note>

~~24 Local sessions are not available on Windows arm64 architectures.~~131 Session isolation requires [Git](https://git-scm.com/downloads). Without Git, sessions in the same directory edit the same files, so changes in one session are immediately visible in others.

25</Note>132</Note>

26 133

~~27## Using Git worktrees~~134To include files listed in your `.gitignore` (like `.env`) in new worktrees, create a `.worktreeinclude` file in your project root listing the file patterns to copy.

135

136To manage a session, click its dropdown in the sidebar to rename it, archive it, or check context usage. When context fills up, Claude automatically summarizes the conversation. You can also ask Claude to compact if you want to free up space earlier.

137

138### Run long-running tasks remotely

139

140For large refactors, test suites, migrations, or other long-running tasks, select **Remote** instead of **Local** when starting a session. Remote sessions run on Anthropic's cloud infrastructure and continue even if you close the app or shut down your computer. Check back anytime to see progress or steer Claude in a different direction.

141

142Remote sessions support **Code** and **Plan** modes. See [Claude Code on the web](/en/claude-code-on-the-web) for details on configuring remote environments.

143

144### Review changes with diff view

145

146After Claude makes changes to your code, the diff view lets you review modifications file by file before creating a pull request.

147

148When Claude changes files, a diff stats indicator appears showing the number of lines added and removed (for example, `+12 -1`). Click this indicator to open the diff viewer, which displays a file list on the left and the changes for each file on the right.

149

150To comment on specific lines, click any line in the diff to open a comment box. Type your feedback and press **Enter** to send. In the full diff view, press **Enter** to accept each comment, then **Cmd+Enter** to send them all. Claude reads your comments and makes the requested changes, which appear as a new diff you can review.

151

152## Extend Claude Code

28 153

29Claude Code on desktop enables running multiple Claude Code sessions in the same repository using Git worktrees. Each session gets its own isolated worktree, allowing Claude to work on different tasks without conflicts. The default location for worktrees is `~/.claude-worktrees` but this can be configured in your settings on the Claude desktop app.154You can extend Claude Code with custom commands, automated workflows, and external integrations.

155

156### Connect external tools

157

158For local sessions, click the **...** button before starting and select **Connectors** to add integrations like Google Calendar, Slack, GitHub, Linear, Notion, and more. Connectors must be configured before the session starts and are only available for local sessions. Once connected, Claude can read your calendar, send messages, create issues, and interact with your tools directly. You can ask Claude what connectors are configured in your session.

159

160Connectors are [MCP (Model Context Protocol) servers](/en/mcp) with built-in setup. You can also [create custom connectors](https://support.claude.com/en/articles/11175166-getting-started-with-custom-connectors-using-remote-mcp) or add MCP servers manually via [configuration files](/en/mcp#configure-mcp-servers).

161

162### Create custom skills

163

164[Skills](/en/skills) are reusable prompts that extend Claude's capabilities. For example, you could create a `review` skill that runs your standard code review checklist, or a `deploy` skill that walks through your deployment steps. Skills are defined as markdown files in `.claude/skills/` and can include instructions, context, and even call other tools. Ask Claude what skills are available or to run a specific skill. Claude can also help you create a skill if you describe what you want, or see [skills](/en/skills) to learn how to write them yourself.

165

166### Automate workflows with hooks

167

168[Hooks](/en/hooks) run shell commands automatically in response to Claude Code events. For example, you could run a linter after every file edit, auto-format code, or send notifications when tasks complete. Hooks are configured in your [settings files](/en/settings). See [hooks](/en/hooks) for available events and configuration examples.

169

170## Environment configuration

171

172When starting a session, you choose between **Local** (runs on your machine) or **Remote** (runs on Anthropic's cloud).

173

174**Local sessions** inherit environment variables from your shell. If you need additional variables, set them in your shell profile (`~/.zshrc`, `~/.bashrc`) and restart the desktop app. See [environment variables](/en/settings#environment-variables) for the full list of supported variables.

175

176[Extended thinking](/en/common-workflows#use-extended-thinking-thinking-mode) is enabled by default, which improves performance on complex reasoning tasks but uses additional tokens. The thinking process runs in the background but isn't displayed in the Desktop interface. To disable it or adjust the budget, set `MAX_THINKING_TOKENS` in your shell profile (use `0` to disable).

177

178**Remote sessions** run on Anthropic's cloud infrastructure and continue even if you close the app. Usage counts toward your subscription plan limits with no separate compute charges. See [Claude Code on the web](/en/claude-code-on-the-web) for details on configuring remote environments.

179

180## How Desktop relates to CLI

181

182If you already use the Claude Code CLI, Desktop runs the same underlying engine with a graphical interface. You can run both simultaneously on the same machine, even on the same project. Each maintains separate session history, but they share configuration and project memory (CLAUDE.md files).

183

184### CLI flag equivalents

185

186If you're used to CLI flags, the table below shows the Desktop equivalent for each. Some flags have no Desktop equivalent because they're designed for scripting or automation.

187

188| CLI | Desktop equivalent |

189| ------------------------------------- | ---------------------------------------------- |

190| `--model sonnet` | **...** menu > Model (before starting session) |

191| `--resume`, `--continue` | Click a session in the sidebar |

192| `--allowedTools`, `--disallowedTools` | Not available in Desktop |

193| `--dangerously-skip-permissions` | Not available in Desktop |

194| `--print` | Not available (Desktop is interactive) |

195

196### Shared configuration

197

198Desktop and CLI read the same configuration files, so your setup carries over:

199

200* **[CLAUDE.md](/en/memory)** and **CLAUDE.local.md** files in your project are used by both

201* **[MCP servers](/en/mcp)** configured in `~/.claude.json` or `.mcp.json` work in both

202* **[Hooks](/en/hooks)** and **[skills](/en/skills)** defined in settings apply to both

203* **[Settings](/en/settings)** in `~/.claude.json` and `~/.claude/settings.json` are shared

204* **Models** (Sonnet, Opus, Haiku) are available in both (Desktop requires selecting before starting a session)

30 205

31<Note>206<Note>

~~32 If you start a local session in a folder that does not have Git initialized, the desktop app will not create a new worktree.~~207 MCP servers configured for the **Claude Desktop chat app** (in `claude_desktop_config.json`) are separate from Claude Code. To use MCP servers in Claude Code, configure them in `~/.claude.json` or your project's `.mcp.json` file. See [MCP configuration](/en/mcp#configure-mcp-servers) for details.

33</Note>208</Note>

34 209

~~35### Copying files ignored with `.gitignore`~~210### What's different

36 211

37When Claude Code creates a worktree, files ignored via `.gitignore` aren't automatically available. Including a `.worktreeinclude` file solves this by specifying which ignored files should be copied to new worktrees.212**Desktop adds:**

38 213

~~39Create a `.worktreeinclude` file in your repository root:~~214* Graphical interface with visual session management

215* Built-in connectors for common integrations

216* Automatic session isolation for Git repositories (each session gets its own worktree)

40 217

~~41```~~218**CLI adds:**

~~42.env~~219

~~43.env.local~~220* [Third-party API providers](/en/third-party-integrations) (Bedrock, Vertex, Foundry). If you use these, continue using CLI for those projects.

~~44.env.*~~221* [CLI flags](/en/cli-reference) for scripting (`--print`, `--resume`, `--continue`)

~~45**/.claude/settings.local.json~~222* [Programmatic usage](/en/headless) via the Agent SDK

223

224## Troubleshooting

225

226Solutions to common issues with the Claude desktop app. For CLI issues, see [CLI troubleshooting](/en/troubleshooting).

227

228### Check your version

229

230To see which version of the desktop app you're running:

231

232* **macOS**: Click **Claude** in the menu bar, then **About Claude**

233* **Windows**: Click **Help**, then **About**

234

235Click the version number to copy it to your clipboard.

236

237### "Branch doesn't exist yet" when opening in CLI

238

239Remote sessions can create branches that don't exist on your local machine. Click the branch name in the session toolbar to copy it, then fetch it locally:

240

241```bash theme={null}

242git fetch origin <branch-name>

243git checkout <branch-name>

46```244```

47 245

~~48The file uses `.gitignore`-style patterns. When a worktree is created, files matching these patterns that are also in your `.gitignore` will be copied from your main repository to the worktree.~~246### "Failed to load session" error

49 247

~~50<Tip>~~248This error can occur for several reasons:

~~51 Only files that are both matched by `.worktreeinclude` AND listed in `.gitignore` are copied. This prevents accidentally duplicating tracked files.~~

~~52</Tip>~~

53 249

~~54### Launch Claude Code on the web~~250* The selected folder no longer exists or is inaccessible

251* A Git repository requires Git LFS but it's not installed (see [Git LFS errors](#git-lfs-errors))

252* File permissions prevent access to the project directory

55 253

~~56From the desktop app, you can kick off Claude Code sessions that run on Anthropic's secure cloud infrastructure. This is useful for:~~254Try selecting a different folder or restarting the desktop app.

57 255

~~58To start a web session from desktop, select a remote environment when creating a new session.~~256### App won't quit

59 257

~~60For more details, see [Claude Code on the web](/en/claude-code-on-the-web).~~258If the desktop app doesn't close properly:

61 259

~~62## Bundled Claude Code version~~260* **macOS**: Press Cmd+Q. If the app doesn't respond, use Force Quit (Cmd+Option+Esc, select Claude, click Force Quit).

261* **Windows**: Use Task Manager (Ctrl+Shift+Esc) to end the Claude process.

63 262

64Claude Code on desktop includes a bundled, stable version of Claude Code to ensure a consistent experience for all desktop users. The bundled version is required and downloaded on first launch even if a version of Claude Code exists on the computer. Desktop automatically manages version updates and cleans up old versions.263### Windows installation issues

65 264

~~66<Note>~~265If the installer fails silently or doesn't complete properly:

~~67 The bundled Claude Code version in Desktop may differ from the latest CLI version. Desktop prioritizes stability while the CLI may have newer features.~~

~~68</Note>~~

69 266

~~70## Environment configuration~~2671. **PATH not updated**: After installation, open a new terminal window. The PATH updates only apply to new terminal sessions.

2682. **Concurrent installation error**: If you see an error about another installation in progress but there isn't one, try running the installer as Administrator.

71 269

72For local environments, Claude Code on desktop automatically extracts your `$PATH` environment variable from your shell configuration. This allows local sessions to access development tools like `yarn`, `npm`, `node`, and other commands available in your terminal without additional setup.270### Session not finding installed tools

73 271

~~74### Custom environment variables~~272If Claude can't find tools like `npm`, `node`, or other CLI commands:

75 273

76Select "Local" environment, then to the right, select the settings button. This will open a dialog where you can update local environment variables. This is useful for setting project-specific variables or API keys that your development workflows require. Environment variable values are masked in the UI for security reasons.2741. Verify the tools work in your regular terminal

2752. Check that your shell profile (`~/.zshrc`, `~/.bashrc`) properly sets up PATH

2763. Restart the desktop app to reload environment variables

77 277

~~78<Note>~~278### MCP servers not working (Windows)

~~79 Environment variables must be specified as key-value pairs, in [`.env` format](https://www.dotenv.org/). For example:~~

80 279

~~81 ```~~280If MCP server toggles don't respond or servers fail to connect on Windows:

~~82 API_KEY=your_api_key~~

~~83 DEBUG=true~~

84 281

~~85 # Multiline values - wrap in quotes~~2821. Check that the MCP server is properly configured in your settings

~~86 CERT="-----BEGIN CERT-----~~2832. Restart the desktop app after making changes

~~87 MIIE...~~2843. Verify the MCP server process is running (check Task Manager)

~~88 -----END CERT-----"~~2854. Review the server logs for connection errors

~~89 ```~~

~~90</Note>~~

91 286

~~92## Enterprise configuration~~287### Git LFS errors

93 288

94Organizations can disable local Claude Code use in the desktop application with the `isClaudeCodeForDesktopEnabled` [enterprise policy option](https://support.claude.com/en/articles/12622667-enterprise-configuration#h_003283c7cb). Additionally, Claude Code on the web can be disabled in your [admin settings](https://claude.ai/admin-settings/claude-code).289If you see "Git LFS is required by this repository but is not installed," your repository uses Git Large File Storage for large binary files. Install Git LFS before opening this repository:

95 290

~~96## Related resources~~2911. Install Git LFS from [git-lfs.com](https://git-lfs.com/)

2922. Run `git lfs install` in your terminal

2933. Restart the desktop app

97 294

~~98* [Claude Code on the web](/en/claude-code-on-the-web)~~295## Enterprise configuration

~~99* [Claude Desktop support articles](https://support.claude.com/en/collections/16163169-claude-desktop)~~

100* [Enterprise Configuration](https://support.claude.com/en/articles/12622667-enterprise-configuration)

101* [Common workflows](/en/common-workflows)

102* [Settings reference](/en/settings)

103 296

297Organizations can disable local Claude Code use in the desktop application with the `isClaudeCodeForDesktopEnabled` [enterprise policy option](https://support.claude.com/en/articles/12622667-enterprise-configuration#h_003283c7cb). Additionally, Claude Code on the web can be disabled in your [admin settings](https://claude.ai/admin-settings/claude-code).

104 298

299## Related resources

105 300

106> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt301* [Claude Code on the web](/en/claude-code-on-the-web): Run remote sessions that continue in the cloud

302* [CLI reference](/en/cli-reference): Use Claude Code in your terminal with flags and scripting

303* [Common workflows](/en/common-workflows): Tutorials for debugging, refactoring, testing, and more

304* [Settings reference](/en/settings): Configure Claude Code behavior with settings files

305* [Claude Desktop support](https://support.claude.com/en/collections/16163169-claude-desktop): Help articles for the Chat tab and general desktop app usage

306* [Enterprise configuration](https://support.claude.com/en/articles/12622667-enterprise-configuration): Admin policies for organizational deployments

devcontainer.md +4 −4

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Development containers5# Development containers

2 6

3> Learn about the Claude Code development container for teams that need consistent, secure environments.7> Learn about the Claude Code development container for teams that need consistent, secure environments.

75* [VS Code devcontainers documentation](https://code.visualstudio.com/docs/devcontainers/containers)79* [VS Code devcontainers documentation](https://code.visualstudio.com/docs/devcontainers/containers)

76* [Claude Code security best practices](/en/security)80* [Claude Code security best practices](/en/security)

77* [Enterprise network configuration](/en/network-config)81* [Enterprise network configuration](/en/network-config)

~~81> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt~~

discover-plugins.md +42 −11

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Discover and install prebuilt plugins through marketplaces5# Discover and install prebuilt plugins through marketplaces

2 6

3> Find and install plugins from marketplaces to extend Claude Code with new commands, agents, and capabilities.7> Find and install plugins from marketplaces to extend Claude Code with new commands, agents, and capabilities.

4 8

~~5Plugins extend Claude Code with custom commands, agents, hooks, and MCP servers. Plugin marketplaces are catalogs that help you discover and install these extensions without building them yourself.~~9Plugins extend Claude Code with skills, agents, hooks, and MCP servers. Plugin marketplaces are catalogs that help you discover and install these extensions without building them yourself.

6 10

7Looking to create and distribute your own marketplace? See [Create and distribute a plugin marketplace](/en/plugin-marketplaces).11Looking to create and distribute your own marketplace? See [Create and distribute a plugin marketplace](/en/plugin-marketplaces).

8 12

40 44

41### Code intelligence45### Code intelligence

42 46

43Code intelligence plugins help Claude understand your codebase more deeply. With these plugins installed, Claude can jump to definitions, find references, and see type errors immediately after edits. These plugins use the [Language Server Protocol](https://microsoft.github.io/language-server-protocol/) (LSP), the same technology that powers VS Code's code intelligence.47Code intelligence plugins enable Claude Code's built-in LSP tool, giving Claude the ability to jump to definitions, find references, and see type errors immediately after edits. These plugins configure [Language Server Protocol](https://microsoft.github.io/language-server-protocol/) connections, the same technology that powers VS Code's code intelligence.

44 48

45These plugins require the language server binary to be installed on your system. If you already have a language server installed, Claude may prompt you to install the corresponding plugin when you open a project.49These plugins require the language server binary to be installed on your system. If you already have a language server installed, Claude may prompt you to install the corresponding plugin when you open a project.

46 50

50| C# | `csharp-lsp` | `csharp-ls` |54| C# | `csharp-lsp` | `csharp-ls` |

51| Go | `gopls-lsp` | `gopls` |55| Go | `gopls-lsp` | `gopls` |

57| Kotlin | `kotlin-lsp` | `kotlin-language-server` |

53| Lua | `lua-lsp` | `lua-language-server` |58| Lua | `lua-lsp` | `lua-language-server` |

54| PHP | `php-lsp` | `intelephense` |59| PHP | `php-lsp` | `intelephense` |

63 If you see `Executable not found in $PATH` in the `/plugin` Errors tab after installing a plugin, install the required binary from the table above.68 If you see `Executable not found in $PATH` in the `/plugin` Errors tab after installing a plugin, install the required binary from the table above.

64</Note>69</Note>

65 70

71#### What Claude gains from code intelligence plugins

73Once a code intelligence plugin is installed and its language server binary is available, Claude gains two capabilities:

75* **Automatic diagnostics**: after every file edit Claude makes, the language server analyzes the changes and reports errors and warnings back automatically. Claude sees type errors, missing imports, and syntax issues without needing to run a compiler or linter. If Claude introduces an error, it notices and fixes the issue in the same turn. This requires no configuration beyond installing the plugin. You can see diagnostics inline by pressing **Ctrl+O** when the "diagnostics found" indicator appears.

76* **Code navigation**: Claude can use the language server to jump to definitions, find references, get type info on hover, list symbols, find implementations, and trace call hierarchies. These operations give Claude more precise navigation than grep-based search, though availability may vary by language and environment.

78If you run into issues, see [Code intelligence troubleshooting](#code-intelligence-issues).

66### External integrations80### External integrations

67 81

68These plugins bundle pre-configured [MCP servers](/en/mcp) so you can connect Claude to external services without manual setup:82These plugins bundle pre-configured [MCP servers](/en/mcp) so you can connect Claude to external services without manual setup:

210/plugin marketplace add ./path/to/marketplace.json224/plugin marketplace add ./path/to/marketplace.json

211```225```

212 226

213Or add a remote `marketplace.json` file via URL:227### Add from remote URLs

228

229Add a remote `marketplace.json` file via URL:

214 230

215```shell theme={null}231```shell theme={null}

216/plugin marketplace add https://example.com/marketplace.json232/plugin marketplace add https://example.com/marketplace.json

217```233```

218 234

235<Note>

236 URL-based marketplaces have some limitations compared to Git-based marketplaces. If you encounter "path not found" errors when installing plugins, see [Troubleshooting](/en/plugin-marketplaces#plugins-with-relative-paths-fail-in-url-based-marketplaces).

237</Note>

238

219## Install plugins239## Install plugins

220 240

221Once you've added marketplaces, you can install plugins directly (installs to user scope by default):241Once you've added marketplaces, you can install plugins directly (installs to user scope by default):

230* **Project scope**: install for all collaborators on this repository (adds to `.claude/settings.json`)250* **Project scope**: install for all collaborators on this repository (adds to `.claude/settings.json`)

231* **Local scope**: install for yourself in this repository only (not shared with collaborators)251* **Local scope**: install for yourself in this repository only (not shared with collaborators)

232 252

233You may also see plugins with **managed** scope—these are installed by enterprise administrators via [managed settings](/en/settings#enterprise-managed-policy-settings) and cannot be modified.253You may also see plugins with **managed** scope—these are installed by administrators via [managed settings](/en/settings#settings-files) and cannot be modified.

234 254

235Run `/plugin` and go to the **Installed** tab to see your plugins grouped by scope.255Run `/plugin` and go to the **Installed** tab to see your plugins grouped by scope.

236 256

240 260

241## Manage installed plugins261## Manage installed plugins

242 262

243Run `/plugin` and go to the **Installed** tab to view, enable, disable, or uninstall your plugins.263Run `/plugin` and go to the **Installed** tab to view, enable, disable, or uninstall your plugins. Type to filter the list by plugin name or description.

244 264

245You can also manage plugins with direct commands.265You can also manage plugins with direct commands.

246 266

323 343

324To disable all automatic updates entirely for both Claude Code and all plugins, set the `DISABLE_AUTOUPDATER` environment variable. See [Auto updates](/en/setup#auto-updates) for details.344To disable all automatic updates entirely for both Claude Code and all plugins, set the `DISABLE_AUTOUPDATER` environment variable. See [Auto updates](/en/setup#auto-updates) for details.

325 345

346To keep plugin auto-updates enabled while disabling Claude Code auto-updates, set `FORCE_AUTOUPDATE_PLUGINS=true` along with `DISABLE_AUTOUPDATER`:

347

348```shell theme={null}

349export DISABLE_AUTOUPDATER=true

350export FORCE_AUTOUPDATE_PLUGINS=true

351```

352

353This is useful when you want to manage Claude Code updates manually but still receive automatic plugin updates.

354

326## Configure team marketplaces355## Configure team marketplaces

327 356

328Team admins can set up automatic marketplace installation for projects by adding marketplace configuration to `.claude/settings.json`. When team members trust the repository folder, Claude Code prompts them to install these marketplaces and plugins.357Team admins can set up automatic marketplace installation for projects by adding marketplace configuration to `.claude/settings.json`. When team members trust the repository folder, Claude Code prompts them to install these marketplaces and plugins.

347* **Marketplace not loading**: Verify the URL is accessible and that `.claude-plugin/marketplace.json` exists at the path376* **Marketplace not loading**: Verify the URL is accessible and that `.claude-plugin/marketplace.json` exists at the path

348* **Plugin installation failures**: Check that plugin source URLs are accessible and repositories are public (or you have access)377* **Plugin installation failures**: Check that plugin source URLs are accessible and repositories are public (or you have access)

349* **Files not found after installation**: Plugins are copied to a cache, so paths referencing files outside the plugin directory won't work378* **Files not found after installation**: Plugins are copied to a cache, so paths referencing files outside the plugin directory won't work

350* **Plugin Skills not appearing**: Clear the cache with `rm -rf ~/.claude/plugins/cache`, restart Claude Code, and reinstall the plugin. See [Plugin Skills not appearing](/en/skills#plugin-skills-not-appearing-after-installation) for details.379* **Plugin skills not appearing**: Clear the cache with `rm -rf ~/.claude/plugins/cache`, restart Claude Code, and reinstall the plugin.

351 380

352For detailed troubleshooting with solutions, see [Troubleshooting](/en/plugin-marketplaces#troubleshooting) in the marketplace guide. For debugging tools, see [Debugging and development tools](/en/plugins-reference#debugging-and-development-tools).381For detailed troubleshooting with solutions, see [Troubleshooting](/en/plugin-marketplaces#troubleshooting) in the marketplace guide. For debugging tools, see [Debugging and development tools](/en/plugins-reference#debugging-and-development-tools).

353 382

383### Code intelligence issues

384

385* **Language server not starting**: verify the binary is installed and available in your `$PATH`. Check the `/plugin` Errors tab for details.

386* **High memory usage**: language servers like `rust-analyzer` and `pyright` can consume significant memory on large projects. If you experience memory issues, disable the plugin with `/plugin disable <plugin-name>` and rely on Claude's built-in search tools instead.

387* **False positive diagnostics in monorepos**: language servers may report unresolved import errors for internal packages if the workspace isn't configured correctly. These don't affect Claude's ability to edit code.

388

354## Next steps389## Next steps

355 390

356* **Build your own plugins**: See [Plugins](/en/plugins) to create custom commands, agents, and hooks391* **Build your own plugins**: See [Plugins](/en/plugins) to create skills, agents, and hooks

357* **Create a marketplace**: See [Create a plugin marketplace](/en/plugin-marketplaces) to distribute plugins to your team or community392* **Create a marketplace**: See [Create a plugin marketplace](/en/plugin-marketplaces) to distribute plugins to your team or community

358* **Technical reference**: See [Plugins reference](/en/plugins-reference) for complete specifications393* **Technical reference**: See [Plugins reference](/en/plugins-reference) for complete specifications

~~359~~

~~360~~

~~361~~

362> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

fast-mode.md +131 −0 added

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# Speed up responses with fast mode

7> Get faster Opus 4.6 responses in Claude Code by toggling fast mode.

9<Note>

10 Fast mode is in [research preview](#research-preview). The feature, pricing, and availability may change based on feedback.

11</Note>

13Fast mode is a high-speed configuration for Claude Opus 4.6, making the model 2.5x faster at a higher cost per token. Toggle it on with `/fast` when you need speed for interactive work like rapid iteration or live debugging, and toggle it off when cost matters more than latency.

15Fast mode is not a different model. It uses the same Opus 4.6 with a different API configuration that prioritizes speed over cost efficiency. You get identical quality and capabilities, just faster responses.

17What to know:

19* Use `/fast` to toggle on fast mode in Claude Code CLI. Also available via `/fast` in Claude Code VS Code Extension.

20* Fast mode for Opus 4.6 pricing starts at \$30/150 MTok. Fast mode is available at a 50% discount for all plans until 11:59pm PT on February 16.

21* Available to all Claude Code users on subscription plans (Pro/Max/Team/Enterprise) and Claude Console.

22* For Claude Code users on subscription plans (Pro/Max/Team/Enterprise), fast mode is available via extra usage only and not included in the subscription rate limits.

24This page covers how to [toggle fast mode](#toggle-fast-mode), its [cost tradeoff](#understand-the-cost-tradeoff), [when to use it](#decide-when-to-use-fast-mode), [requirements](#requirements), and [rate limit behavior](#handle-rate-limits).

26## Toggle fast mode

28Toggle fast mode in either of these ways:

30* Type `/fast` and press Tab to toggle on or off

31* Set `"fastMode": true` in your [user settings file](/en/settings)

33Fast mode persists across sessions. For the best cost efficiency, enable fast mode at the start of a session rather than switching mid-conversation. See [understand the cost tradeoff](#understand-the-cost-tradeoff) for details.

35When you enable fast mode:

37* If you're on a different model, Claude Code automatically switches to Opus 4.6

38* You'll see a confirmation message: "Fast mode ON"

39* A small `↯` icon appears next to the prompt while fast mode is active

40* Run `/fast` again at any time to check whether fast mode is on or off

42When you disable fast mode with `/fast` again, you remain on Opus 4.6. The model does not revert to your previous model. To switch to a different model, use `/model`.

44## Understand the cost tradeoff

46Fast mode has higher per-token pricing than standard Opus 4.6:

48| Mode | Input (MTok) | Output (MTok) |

49| ------------------------------ | ------------ | ------------- |

50| Fast mode on Opus 4.6 (\<200K) | \$30 | \$150 |

51| Fast mode on Opus 4.6 (>200K) | \$60 | \$225 |

53Fast mode is compatible with the 1M token extended context window.

55When you switch into fast mode mid-conversation, you pay the full fast mode uncached input token price for the entire conversation context. This costs more than if you had enabled fast mode from the start.

57## Decide when to use fast mode

59Fast mode is best for interactive work where response latency matters more than cost:

61* Rapid iteration on code changes

62* Live debugging sessions

63* Time-sensitive work with tight deadlines

65Standard mode is better for:

67* Long autonomous tasks where speed matters less

68* Batch processing or CI/CD pipelines

69* Cost-sensitive workloads

71### Fast mode vs effort level

73Fast mode and effort level both affect response speed, but differently:

75| Setting | Effect |

76| ---------------------- | -------------------------------------------------------------------------------- |

77| **Fast mode** | Same model quality, lower latency, higher cost |

78| **Lower effort level** | Less thinking time, faster responses, potentially lower quality on complex tasks |

80You can combine both: use fast mode with a lower [effort level](/en/model-config#adjust-effort-level) for maximum speed on straightforward tasks.

82## Requirements

84Fast mode requires all of the following:

86* **Not available on third-party cloud providers**: fast mode is not available on Amazon Bedrock, Google Vertex AI, or Microsoft Azure Foundry. Fast mode is available through the Anthropic Console API and for Claude subscription plans using extra usage.

87* **Extra usage enabled**: your account must have extra usage enabled, which allows billing beyond your plan's included usage. For individual accounts, enable this in your [Console billing settings](https://platform.claude.com/settings/organization/billing). For Teams and Enterprise, an admin must enable extra usage for the organization.

89<Note>

90 Fast mode usage is billed directly to extra usage, even if you have remaining usage on your plan. This means fast mode tokens do not count against your plan's included usage and are charged at the fast mode rate from the first token.

91</Note>

93* **Admin enablement for Teams and Enterprise**: fast mode is disabled by default for Teams and Enterprise organizations. An admin must explicitly [enable fast mode](#enable-fast-mode-for-your-organization) before users can access it.

95<Note>

96 If your admin has not enabled fast mode for your organization, the `/fast` command will show "Fast mode has been disabled by your organization."

97</Note>

99### Enable fast mode for your organization

100

101Admins can enable fast mode in:

102

103* **Console** (API customers): [Claude Code preferences](https://platform.claude.com/claude-code/preferences)

104* **Claude AI** (Teams and Enterprise): [Admin Settings > Claude Code](https://claude.ai/admin-settings/claude-code)

105

106## Handle rate limits

107

108Fast mode has separate rate limits from standard Opus 4.6. When you hit the fast mode rate limit or run out of extra usage credits:

109

1101. Fast mode automatically falls back to standard Opus 4.6

1112. The `↯` icon turns gray to indicate cooldown

1123. You continue working at standard speed and pricing

1134. When the cooldown expires, fast mode automatically re-enables

114

115To disable fast mode manually instead of waiting for cooldown, run `/fast` again.

116

117## Research preview

118

119Fast mode is a research preview feature. This means:

120

121* The feature may change based on feedback

122* Availability and pricing are subject to change

123* The underlying API configuration may evolve

124

125Report issues or feedback through your usual Anthropic support channels.

126

127## See also

128

129* [Model configuration](/en/model-config): switch models and adjust effort levels

130* [Manage costs effectively](/en/costs): track token usage and reduce costs

131* [Status line configuration](/en/statusline): display model and context information

features-overview.md +278 −0 added

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# Extend Claude Code

7> Understand when to use CLAUDE.md, Skills, subagents, hooks, MCP, and plugins.

9Claude Code combines a model that reasons about your code with [built-in tools](/en/how-claude-code-works#tools) for file operations, search, execution, and web access. The built-in tools cover most coding tasks. This guide covers the extension layer: features you add to customize what Claude knows, connect it to external services, and automate workflows.

11<Note>

12 For how the core agentic loop works, see [How Claude Code works](/en/how-claude-code-works).

13</Note>

15**New to Claude Code?** Start with [CLAUDE.md](/en/memory) for project conventions. Add other extensions as you need them.

17## Overview

19Extensions plug into different parts of the agentic loop:

21* **[CLAUDE.md](/en/memory)** adds persistent context Claude sees every session

22* **[Skills](/en/skills)** add reusable knowledge and invocable workflows

23* **[MCP](/en/mcp)** connects Claude to external services and tools

24* **[Subagents](/en/sub-agents)** run their own loops in isolated context, returning summaries

25* **[Agent teams](/en/agent-teams)** coordinate multiple independent sessions with shared tasks and peer-to-peer messaging

26* **[Hooks](/en/hooks)** run outside the loop entirely as deterministic scripts

27* **[Plugins](/en/plugins)** and **[marketplaces](/en/plugin-marketplaces)** package and distribute these features

29[Skills](/en/skills) are the most flexible extension. A skill is a markdown file containing knowledge, workflows, or instructions. You can invoke skills with a slash command like `/deploy`, or Claude can load them automatically when relevant. Skills can run in your current conversation or in an isolated context via subagents.

31## Match features to your goal

33Features range from always-on context that Claude sees every session, to on-demand capabilities you or Claude can invoke, to background automation that runs on specific events. The table below shows what's available and when each one makes sense.

36| ---------------------------------- | ---------------------------------------------------------- | ------------------------------------------------------------------------------- | -------------------------------------------------------------------------------- |

41| **MCP** | Connect to external services | External data or actions | Query your database, post to Slack, control a browser |

44**[Plugins](/en/plugins)** are the packaging layer. A plugin bundles skills, hooks, subagents, and MCP servers into a single installable unit. Plugin skills are namespaced (like `/my-plugin:review`) so multiple plugins can coexist. Use plugins when you want to reuse the same setup across multiple repositories or distribute to others via a **[marketplace](/en/plugin-marketplaces)**.

46### Compare similar features

48Some features can seem similar. Here's how to tell them apart.

50<Tabs>

51 <Tab title="Skill vs Subagent">

52 Skills and subagents solve different problems:

54 * **Skills** are reusable content you can load into any context

55 * **Subagents** are isolated workers that run separately from your main conversation

57 | Aspect | Skill | Subagent |

58 | --------------- | ---------------------------------------------- | ---------------------------------------------------------------- |

59 | **What it is** | Reusable instructions, knowledge, or workflows | Isolated worker with its own context |

60 | **Key benefit** | Share content across contexts | Context isolation. Work happens separately, only summary returns |

61 | **Best for** | Reference material, invocable workflows | Tasks that read many files, parallel work, specialized workers |

63 **Skills can be reference or action.** Reference skills provide knowledge Claude uses throughout your session (like your API style guide). Action skills tell Claude to do something specific (like `/deploy` that runs your deployment workflow).

65 **Use a subagent** when you need context isolation or when your context window is getting full. The subagent might read dozens of files or run extensive searches, but your main conversation only receives a summary. Since subagent work doesn't consume your main context, this is also useful when you don't need the intermediate work to remain visible. Custom subagents can have their own instructions and can preload skills.

67 **They can combine.** A subagent can preload specific skills (`skills:` field). A skill can run in isolated context using `context: fork`. See [Skills](/en/skills) for details.

68 </Tab>

70 <Tab title="CLAUDE.md vs Skill">

71 Both store instructions, but they load differently and serve different purposes.

73 | Aspect | CLAUDE.md | Skill |

74 | ------------------------- | ---------------------------- | --------------------------------------- |

75 | **Loads** | Every session, automatically | On demand |

76 | **Can include files** | Yes, with `@path` imports | Yes, with `@path` imports |

77 | **Can trigger workflows** | No | Yes, with `/<name>` |

78 | **Best for** | "Always do X" rules | Reference material, invocable workflows |

80 **Put it in CLAUDE.md** if Claude should always know it: coding conventions, build commands, project structure, "never do X" rules.

82 **Put it in a skill** if it's reference material Claude needs sometimes (API docs, style guides) or a workflow you trigger with `/<name>` (deploy, review, release).

84 **Rule of thumb:** Keep CLAUDE.md under \~500 lines. If it's growing, move reference content to skills.

85 </Tab>

87 <Tab title="Subagent vs Agent team">

88 Both parallelize work, but they're architecturally different:

90 * **Subagents** run inside your session and report results back to your main context

91 * **Agent teams** are independent Claude Code sessions that communicate with each other

93 | Aspect | Subagent | Agent team |

94 | ----------------- | ------------------------------------------------ | --------------------------------------------------- |

95 | **Context** | Own context window; results return to the caller | Own context window; fully independent |

96 | **Communication** | Reports results back to the main agent only | Teammates message each other directly |

97 | **Coordination** | Main agent manages all work | Shared task list with self-coordination |

98 | **Best for** | Focused tasks where only the result matters | Complex work requiring discussion and collaboration |

99 | **Token cost** | Lower: results summarized back to main context | Higher: each teammate is a separate Claude instance |

100

101 **Use a subagent** when you need a quick, focused worker: research a question, verify a claim, review a file. The subagent does the work and returns a summary. Your main conversation stays clean.

102

103 **Use an agent team** when teammates need to share findings, challenge each other, and coordinate independently. Agent teams are best for research with competing hypotheses, parallel code review, and new feature development where each teammate owns a separate piece.

104

105 **Transition point:** If you're running parallel subagents but hitting context limits, or if your subagents need to communicate with each other, agent teams are the natural next step.

106

107 <Note>

108 Agent teams are experimental and disabled by default. See [agent teams](/en/agent-teams) for setup and current limitations.

109 </Note>

110 </Tab>

111

112 <Tab title="MCP vs Skill">

113 MCP connects Claude to external services. Skills extend what Claude knows, including how to use those services effectively.

114

115 | Aspect | MCP | Skill |

116 | -------------- | ---------------------------------------------------- | ------------------------------------------------------- |

117 | **What it is** | Protocol for connecting to external services | Knowledge, workflows, and reference material |

118 | **Provides** | Tools and data access | Knowledge, workflows, reference material |

119 | **Examples** | Slack integration, database queries, browser control | Code review checklist, deploy workflow, API style guide |

120

121 These solve different problems and work well together:

122

123 **MCP** gives Claude the ability to interact with external systems. Without MCP, Claude can't query your database or post to Slack.

124

125 **Skills** give Claude knowledge about how to use those tools effectively, plus workflows you can trigger with `/<name>`. A skill might include your team's database schema and query patterns, or a `/post-to-slack` workflow with your team's message formatting rules.

126

127 Example: An MCP server connects Claude to your database. A skill teaches Claude your data model, common query patterns, and which tables to use for different tasks.

128 </Tab>

129</Tabs>

130

131### Understand how features layer

132

133Features can be defined at multiple levels: user-wide, per-project, via plugins, or through managed policies. You can also nest CLAUDE.md files in subdirectories or place skills in specific packages of a monorepo. When the same feature exists at multiple levels, here's how they layer:

134

135* **CLAUDE.md files** are additive: all levels contribute content to Claude's context simultaneously. Files from your working directory and above load at launch; subdirectories load as you work in them. When instructions conflict, Claude uses judgment to reconcile them, with more specific instructions typically taking precedence. See [how Claude looks up memories](/en/memory#how-claude-looks-up-memories).

136* **Skills and subagents** override by name: when the same name exists at multiple levels, one definition wins based on priority (managed > user > project for skills; managed > CLI flag > project > user > plugin for subagents). Plugin skills are [namespaced](/en/plugins#add-skills-to-your-plugin) to avoid conflicts. See [skill discovery](/en/skills#where-skills-live) and [subagent scope](/en/sub-agents#choose-the-subagent-scope).

137* **MCP servers** override by name: local > project > user. See [MCP scope](/en/mcp#scope-hierarchy-and-precedence).

138* **Hooks** merge: all registered hooks fire for their matching events regardless of source. See [hooks](/en/hooks).

139

140### Combine features

141

142Each extension solves a different problem: CLAUDE.md handles always-on context, skills handle on-demand knowledge and workflows, MCP handles external connections, subagents handle isolation, and hooks handle automation. Real setups combine them based on your workflow.

143

144For example, you might use CLAUDE.md for project conventions, a skill for your deployment workflow, MCP to connect to your database, and a hook to run linting after every edit. Each feature handles what it's best at.

145

146| Pattern | How it works | Example |

147| ---------------------- | -------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- |

148| **Skill + MCP** | MCP provides the connection; a skill teaches Claude how to use it well | MCP connects to your database, a skill documents your schema and query patterns |

149| **Skill + Subagent** | A skill spawns subagents for parallel work | `/review` skill kicks off security, performance, and style subagents that work in isolated context |

150| **CLAUDE.md + Skills** | CLAUDE.md holds always-on rules; skills hold reference material loaded on demand | CLAUDE.md says "follow our API conventions," a skill contains the full API style guide |

151| **Hook + MCP** | A hook triggers external actions through MCP | Post-edit hook sends a Slack notification when Claude modifies critical files |

152

153## Understand context costs

154

155Every feature you add consumes some of Claude's context. Too much can fill up your context window, but it can also add noise that makes Claude less effective; skills may not trigger correctly, or Claude may lose track of your conventions. Understanding these trade-offs helps you build an effective setup.

156

157### Context cost by feature

158

159Each feature has a different loading strategy and context cost:

160

162| --------------- | ------------------------- | --------------------------------------------- | -------------------------------------------- |

168

169\*By default, skill descriptions load at session start so Claude can decide when to use them. Set `disable-model-invocation: true` in a skill's frontmatter to hide it from Claude entirely until you invoke it manually. This reduces context cost to zero for skills you only trigger yourself.

170

171### Understand how features load

172

173Each feature loads at different points in your session. The tabs below explain when each one loads and what goes into context.

174

175<img src="https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/context-loading.svg?fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=bd2e24b8e6a99b31ecfffb63f5b23bf5" alt="Context loading: CLAUDE.md and MCP load at session start and stay in every request. Skills load descriptions at start, full content on invocation. Subagents get isolated context. Hooks run externally." data-og-width="720" width="720" data-og-height="410" height="410" data-path="images/context-loading.svg" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/context-loading.svg?w=280&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=aebaadd1f484f285dd9cb4e0ea6d49b9 280w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/context-loading.svg?w=560&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=030c9b46126d750de315612560082727 560w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/context-loading.svg?w=840&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=6c73f8b0389da4f3190843140c810fe9 840w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/context-loading.svg?w=1100&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=9844c55d08d2c386672447f2e8518669 1100w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/context-loading.svg?w=1650&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=21a9522d0e4bd10ced146aab850ede76 1650w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/context-loading.svg?w=2500&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=d318525915aee1a1a6a4215cfaa61fb9 2500w" />

176

177<Tabs>

178 <Tab title="CLAUDE.md">

179 **When:** Session start

180

181 **What loads:** Full content of all CLAUDE.md files (managed, user, and project levels).

182

183 **Inheritance:** Claude reads CLAUDE.md files from your working directory up to the root, and discovers nested ones in subdirectories as it accesses those files. See [How Claude looks up memories](/en/memory#how-claude-looks-up-memories) for details.

184

185 <Tip>Keep CLAUDE.md under \~500 lines. Move reference material to skills, which load on-demand.</Tip>

186 </Tab>

187

188 <Tab title="Skills">

189 Skills are extra capabilities in Claude's toolkit. They can be reference material (like an API style guide) or invocable workflows you trigger with `/<name>` (like `/deploy`). Some are built-in; you can also create your own. Claude uses skills when appropriate, or you can invoke one directly.

190

191 **When:** Depends on the skill's configuration. By default, descriptions load at session start and full content loads when used. For user-only skills (`disable-model-invocation: true`), nothing loads until you invoke them.

192

193 **What loads:** For model-invocable skills, Claude sees names and descriptions in every request. When you invoke a skill with `/<name>` or Claude loads it automatically, the full content loads into your conversation.

194

195 **How Claude chooses skills:** Claude matches your task against skill descriptions to decide which are relevant. If descriptions are vague or overlap, Claude may load the wrong skill or miss one that would help. To tell Claude to use a specific skill, invoke it with `/<name>`. Skills with `disable-model-invocation: true` are invisible to Claude until you invoke them.

196

197 **Context cost:** Low until used. User-only skills have zero cost until invoked.

198

199 **In subagents:** Skills work differently in subagents. Instead of on-demand loading, skills passed to a subagent are fully preloaded into its context at launch. Subagents don't inherit skills from the main session; you must specify them explicitly.

200

201 <Tip>Use `disable-model-invocation: true` for skills with side effects. This saves context and ensures only you trigger them.</Tip>

202 </Tab>

203

204 <Tab title="MCP servers">

205 **When:** Session start.

206

207 **What loads:** All tool definitions and JSON schemas from connected servers.

208

209 **Context cost:** [Tool search](/en/mcp#scale-with-mcp-tool-search) (enabled by default) loads MCP tools up to 10% of context and defers the rest until needed.

210

211 **Reliability note:** MCP connections can fail silently mid-session. If a server disconnects, its tools disappear without warning. Claude may try to use a tool that no longer exists. If you notice Claude failing to use an MCP tool it previously could access, check the connection with `/mcp`.

212

213 <Tip>Run `/mcp` to see token costs per server. Disconnect servers you're not actively using.</Tip>

214 </Tab>

215

216 <Tab title="Subagents">

217 **When:** On demand, when you or Claude spawns one for a task.

218

219 **What loads:** Fresh, isolated context containing:

220

221 * The system prompt (shared with parent for cache efficiency)

222 * Full content of skills listed in the agent's `skills:` field

223 * CLAUDE.md and git status (inherited from parent)

224 * Whatever context the lead agent passes in the prompt

225

226 **Context cost:** Isolated from main session. Subagents don't inherit your conversation history or invoked skills.

227

228 <Tip>Use subagents for work that doesn't need your full conversation context. Their isolation prevents bloating your main session.</Tip>

229 </Tab>

230

231 <Tab title="Hooks">

232 **When:** On trigger. Hooks fire at specific lifecycle events like tool execution, session boundaries, prompt submission, permission requests, and compaction. See [Hooks](/en/hooks) for the full list.

233

234 **What loads:** Nothing by default. Hooks run as external scripts.

235

236 **Context cost:** Zero, unless the hook returns output that gets added as messages to your conversation.

237

238 <Tip>Hooks are ideal for side effects (linting, logging) that don't need to affect Claude's context.</Tip>

239 </Tab>

240</Tabs>

241

242## Learn more

243

244Each feature has its own guide with setup instructions, examples, and configuration options.

245

246<CardGroup cols={2}>

247 <Card title="CLAUDE.md" icon="file-lines" href="/en/memory">

248 Store project context, conventions, and instructions

249 </Card>

250

251 <Card title="Skills" icon="brain" href="/en/skills">

252 Give Claude domain expertise and reusable workflows

253 </Card>

254

255 <Card title="Subagents" icon="users" href="/en/sub-agents">

256 Offload work to isolated context

257 </Card>

258

259 <Card title="Agent teams" icon="network" href="/en/agent-teams">

260 Coordinate multiple sessions working in parallel

261 </Card>

262

263 <Card title="MCP" icon="plug" href="/en/mcp">

264 Connect Claude to external services

265 </Card>

266

267 <Card title="Hooks" icon="bolt" href="/en/hooks-guide">

268 Automate workflows with hooks

269 </Card>

270

271 <Card title="Plugins" icon="puzzle-piece" href="/en/plugins">

272 Bundle and share feature sets

273 </Card>

274

275 <Card title="Marketplaces" icon="store" href="/en/plugin-marketplaces">

276 Host and distribute plugin collections

277 </Card>

278</CardGroup>

github-actions.md +15 −15

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code GitHub Actions5# Claude Code GitHub Actions

2 6

3> Learn about integrating Claude Code into your development workflow with Claude Code GitHub Actions7> Learn about integrating Claude Code into your development workflow with Claude Code GitHub Actions

5Claude Code GitHub Actions brings AI-powered automation to your GitHub workflow. With a simple `@claude` mention in any PR or issue, Claude can analyze your code, create pull requests, implement features, and fix bugs - all while following your project's standards.9Claude Code GitHub Actions brings AI-powered automation to your GitHub workflow. With a simple `@claude` mention in any PR or issue, Claude can analyze your code, create pull requests, implement features, and fix bugs - all while following your project's standards.

6 10

7<Note>11<Note>

~~8 Claude Code GitHub Actions is built on top of the [Claude Code~~12 Claude Code GitHub Actions is built on top of the [Claude

~~9 SDK](https://docs.claude.com/en/docs/agent-sdk), which enables programmatic integration of~~13 Agent SDK](https://platform.claude.com/docs/en/agent-sdk/overview), which enables programmatic integration of

10 Claude Code into your applications. You can use the SDK to build custom14 Claude Code into your applications. You can use the SDK to build custom

11 automation workflows beyond GitHub Actions.15 automation workflows beyond GitHub Actions.

12</Note>16</Note>

13 17

14<Info>18<Info>

~~15 **Claude Opus 4.5 is now available.** Claude Code GitHub Actions default to Sonnet. To use Opus 4.5, configure the [model parameter](#breaking-changes-reference) to use `claude-opus-4-5-20251101`.~~19 **Claude Opus 4.6 is now available.** Claude Code GitHub Actions default to Sonnet. To use Opus 4.6, configure the [model parameter](#breaking-changes-reference) to use `claude-opus-4-6`.

16</Info>20</Info>

17 21

18## Why use Claude Code GitHub Actions?22## Why use Claude Code GitHub Actions?

90### Breaking Changes Reference94### Breaking Changes Reference

91 95

92| Old Beta Input | New v1.0 Input |96| Old Beta Input | New v1.0 Input |

~~93| --------------------- | -------------------------------- |~~97| --------------------- | ------------------------------------- |

94| `mode` | *(Removed - auto-detected)* |98| `mode` | *(Removed - auto-detected)* |

95| `direct_prompt` | `prompt` |99| `direct_prompt` | `prompt` |

124 prompt: "Review this PR for security issues"128 prompt: "Review this PR for security issues"

125 anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}129 anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}

126 claude_args: |130 claude_args: |

127 --system-prompt "Follow our coding standards"131 --append-system-prompt "Follow our coding standards"

128 --max-turns 10132 --max-turns 10

129 --model claude-sonnet-4-5-20250929133 --model claude-sonnet-4-5-20250929

130```134```

156 # Responds to @claude mentions in comments160 # Responds to @claude mentions in comments

157```161```

158 162

159### Using slash commands163### Using skills

160 164

161```yaml theme={null}165```yaml theme={null}

162name: Code Review166name: Code Review

189 with:193 with:

190 anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}194 anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}

191 prompt: "Generate a summary of yesterday's commits and open issues"195 prompt: "Generate a summary of yesterday's commits and open issues"

192 claude_args: "--model claude-opus-4-5-20251101"196 claude_args: "--model opus"

193```197```

194 198

195### Common use cases199### Common use cases

266Key features:270Key features:

267 271

268* **Unified prompt interface** - Use `prompt` for all instructions272* **Unified prompt interface** - Use `prompt` for all instructions

269* **Slash commands** - Pre-built prompts like `/review` or `/fix`273* **Commands** - Prebuilt prompts like `/review` or `/fix`

270* **CLI passthrough** - Any Claude Code CLI argument via `claude_args`274* **CLI passthrough** - Any Claude Code CLI argument via `claude_args`

271* **Flexible triggers** - Works with any GitHub event275* **Flexible triggers** - Works with any GitHub event

272 276

624The Claude Code Action v1 uses a simplified configuration:628The Claude Code Action v1 uses a simplified configuration:

625 629

627| ------------------- | ----------------------------------------------- | -------- |631| ------------------- | ------------------------------------------------------ | -------- |

628| `prompt` | Instructions for Claude (text or slash command) | No\* |632| `prompt` | Instructions for Claude (text or skill like `/review`) | No\* |

629| `claude_args` | CLI arguments passed to Claude Code | No |633| `claude_args` | CLI arguments passed to Claude Code | No |

630| `anthropic_api_key` | Claude API key | Yes\*\* |634| `anthropic_api_key` | Claude API key | Yes\*\* |

631| `github_token` | GitHub token for API access | No |635| `github_token` | GitHub token for API access | No |

6702. **Custom prompts**: Use the `prompt` parameter in the workflow file to provide workflow-specific instructions. This allows you to customize Claude's behavior for different workflows or tasks.6742. **Custom prompts**: Use the `prompt` parameter in the workflow file to provide workflow-specific instructions. This allows you to customize Claude's behavior for different workflows or tasks.

671 675

672Claude will follow these guidelines when creating PRs and responding to requests.676Claude will follow these guidelines when creating PRs and responding to requests.

~~673~~

~~674~~

~~675~~

676> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

gitlab-ci-cd.md +15 −15

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code GitLab CI/CD5# Claude Code GitLab CI/CD

2 6

3> Learn about integrating Claude Code into your development workflow with GitLab CI/CD7> Learn about integrating Claude Code into your development workflow with GitLab CI/CD

9</Info>13</Info>

10 14

11<Note>15<Note>

12 This integration is built on top of the [Claude Code CLI and SDK](https://docs.claude.com/en/docs/agent-sdk), enabling programmatic use of Claude in your CI/CD jobs and custom automation workflows.16 This integration is built on top of the [Claude Code CLI and Agent SDK](https://platform.claude.com/docs/en/agent-sdk/overview), enabling programmatic use of Claude in your CI/CD jobs and custom automation workflows.

13</Note>17</Note>

14 18

15## Why use Claude Code with GitLab?19## Why use Claude Code with GitLab?

77 before_script:81 before_script:

78 - apk update82 - apk update

79 - apk add --no-cache git curl bash83 - apk add --no-cache git curl bash

~~80 - npm install -g @anthropic-ai/claude-code~~84 - curl -fsSL https://claude.ai/install.sh | bash

81 script:85 script:

82 # Optional: start a GitLab MCP server if your setup provides one86 # Optional: start a GitLab MCP server if your setup provides one

83 - /bin/gitlab-mcp-server || true87 - /bin/gitlab-mcp-server || true

87 claude91 claude

88 -p "${AI_FLOW_INPUT:-'Review this MR and implement the requested changes'}"92 -p "${AI_FLOW_INPUT:-'Review this MR and implement the requested changes'}"

89 --permission-mode acceptEdits93 --permission-mode acceptEdits

~~90 --allowedTools "Bash(*) Read(*) Edit(*) Write(*) mcp__gitlab"~~94 --allowedTools "Bash Read Edit Write mcp__gitlab"

91 --debug95 --debug

92```96```

93 97

255 before_script:259 before_script:

256 - apk update260 - apk update

257 - apk add --no-cache git curl bash261 - apk add --no-cache git curl bash

258 - npm install -g @anthropic-ai/claude-code262 - curl -fsSL https://claude.ai/install.sh | bash

259 script:263 script:

260 - /bin/gitlab-mcp-server || true264 - /bin/gitlab-mcp-server || true

261 - >265 - >

262 claude266 claude

263 -p "${AI_FLOW_INPUT:-'Summarize recent changes and suggest improvements'}"267 -p "${AI_FLOW_INPUT:-'Summarize recent changes and suggest improvements'}"

264 --permission-mode acceptEdits268 --permission-mode acceptEdits

265 --allowedTools "Bash(*) Read(*) Edit(*) Write(*) mcp__gitlab"269 --allowedTools "Bash Read Edit Write mcp__gitlab"

266 --debug270 --debug

267 # Claude Code will use ANTHROPIC_API_KEY from CI/CD variables271 # Claude Code will use ANTHROPIC_API_KEY from CI/CD variables

268```272```

289 before_script:293 before_script:

290 - apk add --no-cache bash curl jq git python3 py3-pip294 - apk add --no-cache bash curl jq git python3 py3-pip

291 - pip install --no-cache-dir awscli295 - pip install --no-cache-dir awscli

292 - npm install -g @anthropic-ai/claude-code296 - curl -fsSL https://claude.ai/install.sh | bash

293 # Exchange GitLab OIDC token for AWS credentials297 # Exchange GitLab OIDC token for AWS credentials

294 - export AWS_WEB_IDENTITY_TOKEN_FILE="${CI_JOB_JWT_FILE:-/tmp/oidc_token}"298 - export AWS_WEB_IDENTITY_TOKEN_FILE="${CI_JOB_JWT_FILE:-/tmp/oidc_token}"

295 - if [ -n "${CI_JOB_JWT_V2}" ]; then printf "%s" "$CI_JOB_JWT_V2" > "$AWS_WEB_IDENTITY_TOKEN_FILE"; fi299 - if [ -n "${CI_JOB_JWT_V2}" ]; then printf "%s" "$CI_JOB_JWT_V2" > "$AWS_WEB_IDENTITY_TOKEN_FILE"; fi

308 claude312 claude

309 -p "${AI_FLOW_INPUT:-'Implement the requested changes and open an MR'}"313 -p "${AI_FLOW_INPUT:-'Implement the requested changes and open an MR'}"

310 --permission-mode acceptEdits314 --permission-mode acceptEdits

311 --allowedTools "Bash(*) Read(*) Edit(*) Write(*) mcp__gitlab"315 --allowedTools "Bash Read Edit Write mcp__gitlab"

312 --debug316 --debug

313 variables:317 variables:

314 AWS_REGION: "us-west-2"318 AWS_REGION: "us-west-2"

339 rules:343 rules:

340 - if: '$CI_PIPELINE_SOURCE == "web"'344 - if: '$CI_PIPELINE_SOURCE == "web"'

341 before_script:345 before_script:

342 - apt-get update && apt-get install -y git nodejs npm && apt-get clean346 - apt-get update && apt-get install -y git && apt-get clean

343 - npm install -g @anthropic-ai/claude-code347 - curl -fsSL https://claude.ai/install.sh | bash

344 # Authenticate to Google Cloud via WIF (no downloaded keys)348 # Authenticate to Google Cloud via WIF (no downloaded keys)

345 - >349 - >

346 gcloud auth login --cred-file=<(cat <<EOF350 gcloud auth login --cred-file=<(cat <<EOF

361 claude365 claude

362 -p "${AI_FLOW_INPUT:-'Review and update code as requested'}"366 -p "${AI_FLOW_INPUT:-'Review and update code as requested'}"

363 --permission-mode acceptEdits367 --permission-mode acceptEdits

364 --allowedTools "Bash(*) Read(*) Edit(*) Write(*) mcp__gitlab"368 --allowedTools "Bash Read Edit Write mcp__gitlab"

365 --debug369 --debug

366 variables:370 variables:

367 CLOUD_ML_REGION: "us-east5"371 CLOUD_ML_REGION: "us-east5"

404* **API costs**:408* **API costs**:

405 * Each Claude interaction consumes tokens based on prompt and response size409 * Each Claude interaction consumes tokens based on prompt and response size

406 * Token usage varies by task complexity and codebase size410 * Token usage varies by task complexity and codebase size

407 * See [Anthropic pricing](https://docs.claude.com/en/docs/about-claude/pricing) for details411 * See [Anthropic pricing](https://platform.claude.com/docs/en/about-claude/pricing) for details

408 412

409* **Cost optimization tips**:413* **Cost optimization tips**:

410 * Use specific `@claude` commands to reduce unnecessary turns414 * Use specific `@claude` commands to reduce unnecessary turns

460 464

4611. **CLAUDE.md**: Define coding standards, security requirements, and project conventions. Claude reads this during runs and follows your rules.4651. **CLAUDE.md**: Define coding standards, security requirements, and project conventions. Claude reads this during runs and follows your rules.

4622. **Custom prompts**: Pass task-specific instructions via `prompt`/`prompt_file` in the job. Use different prompts for different jobs (for example, review, implement, refactor).4662. **Custom prompts**: Pass task-specific instructions via `prompt`/`prompt_file` in the job. Use different prompts for different jobs (for example, review, implement, refactor).

~~463~~

~~464~~

~~465~~

466> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

google-vertex-ai.md +7 −7

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code on Google Vertex AI5# Claude Code on Google Vertex AI

2 6

3> Learn about configuring Claude Code through Google Vertex AI, including setup, IAM configuration, and troubleshooting.7> Learn about configuring Claude Code through Google Vertex AI, including setup, IAM configuration, and troubleshooting.

82```86```

83 87

84<Note>88<Note>

85 [Prompt caching](https://docs.claude.com/en/docs/build-with-claude/prompt-caching) is automatically supported when you specify the `cache_control` ephemeral flag. To disable it, set `DISABLE_PROMPT_CACHING=1`. For heightened rate limits, contact Google Cloud support.89 [Prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching) is automatically supported when you specify the `cache_control` ephemeral flag. To disable it, set `DISABLE_PROMPT_CACHING=1`. For heightened rate limits, contact Google Cloud support.

86</Note>90</Note>

87 91

88<Note>92<Note>

105To customize models:109To customize models:

106 110

107```bash theme={null}111```bash theme={null}

108export ANTHROPIC_MODEL='claude-opus-4-1@20250805'112export ANTHROPIC_MODEL='claude-opus-4-6'

109export ANTHROPIC_SMALL_FAST_MODEL='claude-haiku-4-5@20251001'113export ANTHROPIC_SMALL_FAST_MODEL='claude-haiku-4-5@20251001'

110```114```

111 115

127 131

128## 1M token context window132## 1M token context window

129 133

130Claude Sonnet 4 and Sonnet 4.5 support the [1M token context window](https://docs.claude.com/en/docs/build-with-claude/context-windows#1m-token-context-window) on Vertex AI.134Claude Sonnet 4 and Sonnet 4.5 support the [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows#1m-token-context-window) on Vertex AI.

131 135

132<Note>136<Note>

133 The 1M token context window is currently in beta. To use the extended context window, include the `context-1m-2025-08-07` beta header in your Vertex AI requests.137 The 1M token context window is currently in beta. To use the extended context window, include the `context-1m-2025-08-07` beta header in your Vertex AI requests.

157* [Vertex AI documentation](https://cloud.google.com/vertex-ai/docs)161* [Vertex AI documentation](https://cloud.google.com/vertex-ai/docs)

158* [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing)162* [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing)

159* [Vertex AI quotas and limits](https://cloud.google.com/vertex-ai/docs/quotas)163* [Vertex AI quotas and limits](https://cloud.google.com/vertex-ai/docs/quotas)

~~160~~

~~161~~

~~162~~

163> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

headless.md +25 −6

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Run Claude Code programmatically5# Run Claude Code programmatically

2 6

3> Use the Agent SDK to run Claude Code programmatically from the CLI, Python, or TypeScript.7> Use the Agent SDK to run Claude Code programmatically from the CLI, Python, or TypeScript.

73 ```77 ```

74</Tip>78</Tip>

75 79

80### Stream responses

82Use `--output-format stream-json` with `--verbose` and `--include-partial-messages` to receive tokens as they're generated. Each line is a JSON object representing an event:

84```bash theme={null}

85claude -p "Explain recursion" --output-format stream-json --verbose --include-partial-messages

86```

88The following example uses [jq](https://jqlang.github.io/jq/) to filter for text deltas and display just the streaming text. The `-r` flag outputs raw strings (no quotes) and `-j` joins without newlines so tokens stream continuously:

90```bash theme={null}

91claude -p "Write a poem" --output-format stream-json --verbose --include-partial-messages | \

92 jq -rj 'select(.type == "stream_event" and .event.delta.type? == "text_delta") | .event.delta.text'

93```

95For programmatic streaming with callbacks and message objects, see [Stream responses in real-time](https://platform.claude.com/docs/en/agent-sdk/streaming-output) in the Agent SDK documentation.

76### Auto-approve tools97### Auto-approve tools

77 98

78Use `--allowedTools` to let Claude use certain tools without prompting. This example runs a test suite and fixes failures, allowing Claude to execute Bash commands and read/edit files without asking for permission:99Use `--allowedTools` to let Claude use certain tools without prompting. This example runs a test suite and fixes failures, allowing Claude to execute Bash commands and read/edit files without asking for permission:

88 109

89```bash theme={null}110```bash theme={null}

90claude -p "Look at my staged changes and create an appropriate commit" \111claude -p "Look at my staged changes and create an appropriate commit" \

~~91 --allowedTools "Bash(git diff:*),Bash(git log:*),Bash(git status:*),Bash(git commit:*)"~~112 --allowedTools "Bash(git diff *),Bash(git log *),Bash(git status *),Bash(git commit *)"

92```113```

93 114

115The `--allowedTools` flag uses [permission rule syntax](/en/settings#permission-rule-syntax). The trailing ` *` enables prefix matching, so `Bash(git diff *)` allows any command starting with `git diff`. The space before `*` is important: without it, `Bash(git diff*)` would also match `git diff-index`.

116

94<Note>117<Note>

~~95 [Slash commands](/en/slash-commands) like `/commit` are only available in interactive mode. In `-p` mode, describe the task you want to accomplish instead.~~118 User-invoked [skills](/en/skills) like `/commit` and [built-in commands](/en/interactive-mode#built-in-commands) are only available in interactive mode. In `-p` mode, describe the task you want to accomplish instead.

96</Note>119</Note>

97 120

98### Customize the system prompt121### Customize the system prompt

146 Use the Agent SDK in GitLab pipelines169 Use the Agent SDK in GitLab pipelines

147 </Card>170 </Card>

148</CardGroup>171</CardGroup>

~~149~~

~~150~~

~~151~~

152> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

hooks.md +1143 −726

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Hooks reference5# Hooks reference

2 6

~~3> This page provides reference documentation for implementing hooks in Claude Code.~~7> Reference for Claude Code hook events, configuration schema, JSON input/output formats, exit codes, async hooks, prompt hooks, and MCP tool hooks.

4 8

5<Tip>9<Tip>

~~6 For a quickstart guide with examples, see [Get started with Claude Code hooks](/en/hooks-guide).~~10 For a quickstart guide with examples, see [Automate workflows with hooks](/en/hooks-guide).

7</Tip>11</Tip>

8 12

~~9## Configuration~~13Hooks are user-defined shell commands or LLM prompts that execute automatically at specific points in Claude Code's lifecycle. Use this reference to look up event schemas, configuration options, JSON input/output formats, and advanced features like async hooks and MCP tool hooks. If you're setting up hooks for the first time, start with the [guide](/en/hooks-guide) instead.

10 14

~~11Claude Code hooks are configured in your [settings files](/en/settings):~~15## Hook lifecycle

12 16

~~13* `~/.claude/settings.json` - User settings~~17Hooks fire at specific points during a Claude Code session. When an event fires and a matcher matches, Claude Code passes JSON context about the event to your hook handler. For command hooks, this arrives on stdin. Your handler can then inspect the input, take action, and optionally return a decision. Some events fire once per session, while others fire repeatedly inside the agentic loop:

~~14* `.claude/settings.json` - Project settings~~

~~15* `.claude/settings.local.json` - Local project settings (not committed)~~

~~16* Enterprise managed policy settings~~

17 18

~~18<Note>~~19<div style={{maxWidth: "500px", margin: "0 auto"}}>

~~19 Enterprise administrators can use `allowManagedHooksOnly` to block user, project, and plugin hooks. See [Hook configuration](/en/settings#hook-configuration).~~20 <Frame>

~~20</Note>~~21 <img src="https://mintcdn.com/claude-code/tpQvD9DKENFo4zX_/images/hooks-lifecycle.svg?fit=max&auto=format&n=tpQvD9DKENFo4zX_&q=85&s=7a351ea1cc3d5da7a2176bf51196bc1a" alt="Hook lifecycle diagram showing the sequence of hooks from SessionStart through the agentic loop to SessionEnd" data-og-width="520" width="520" data-og-height="960" height="960" data-path="images/hooks-lifecycle.svg" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/tpQvD9DKENFo4zX_/images/hooks-lifecycle.svg?w=280&fit=max&auto=format&n=tpQvD9DKENFo4zX_&q=85&s=8f32c67d025f0a318d5ed10a4f8ff2e6 280w, https://mintcdn.com/claude-code/tpQvD9DKENFo4zX_/images/hooks-lifecycle.svg?w=560&fit=max&auto=format&n=tpQvD9DKENFo4zX_&q=85&s=896fc424e39ff8d590720331a77e3d80 560w, https://mintcdn.com/claude-code/tpQvD9DKENFo4zX_/images/hooks-lifecycle.svg?w=840&fit=max&auto=format&n=tpQvD9DKENFo4zX_&q=85&s=a1c1c9739cde965e1eade843cee567c5 840w, https://mintcdn.com/claude-code/tpQvD9DKENFo4zX_/images/hooks-lifecycle.svg?w=1100&fit=max&auto=format&n=tpQvD9DKENFo4zX_&q=85&s=5bb083988de020e7d568e8dd8f1422fc 1100w, https://mintcdn.com/claude-code/tpQvD9DKENFo4zX_/images/hooks-lifecycle.svg?w=1650&fit=max&auto=format&n=tpQvD9DKENFo4zX_&q=85&s=343e9883c1e3172f08096c352aa46f12 1650w, https://mintcdn.com/claude-code/tpQvD9DKENFo4zX_/images/hooks-lifecycle.svg?w=2500&fit=max&auto=format&n=tpQvD9DKENFo4zX_&q=85&s=4de37b29de0f6df8b0c3e937a76c3bc6 2500w" />

22 </Frame>

23</div>

21 24

~~22### Structure~~25The table below summarizes when each event fires. The [Hook events](#hook-events) section documents the full input schema and decision control options for each one.

23 26

~~24Hooks are organized by matchers, where each matcher can have multiple hooks:~~27| Event | When it fires |

28| :------------------- | :----------------------------------------------------------------- |

29| `SessionStart` | When a session begins or resumes |

30| `UserPromptSubmit` | When you submit a prompt, before Claude processes it |

31| `PreToolUse` | Before a tool call executes. Can block it |

32| `PermissionRequest` | When a permission dialog appears |

33| `PostToolUse` | After a tool call succeeds |

34| `PostToolUseFailure` | After a tool call fails |

35| `Notification` | When Claude Code sends a notification |

36| `SubagentStart` | When a subagent is spawned |

37| `SubagentStop` | When a subagent finishes |

38| `Stop` | When Claude finishes responding |

39| `TeammateIdle` | When an [agent team](/en/agent-teams) teammate is about to go idle |

40| `TaskCompleted` | When a task is being marked as completed |

41| `PreCompact` | Before context compaction |

42| `SessionEnd` | When a session terminates |

44### How a hook resolves

46To see how these pieces fit together, consider this `PreToolUse` hook that blocks destructive shell commands. The hook runs `block-rm.sh` before every Bash tool call:

25 47

26```json theme={null}48```json theme={null}

27{49{

28 "hooks": {50 "hooks": {

~~29 "EventName": [~~51 "PreToolUse": [

30 {52 {

~~31 "matcher": "ToolPattern",~~53 "matcher": "Bash",

32 "hooks": [54 "hooks": [

33 {55 {

34 "type": "command",56 "type": "command",

~~35 "command": "your-command-here"~~57 "command": ".claude/hooks/block-rm.sh"

36 }58 }

37 ]59 ]

38 }60 }

41}63}

42```64```

43 65

~~44* **matcher**: Pattern to match tool names, case-sensitive (only applicable for~~66The script reads the JSON input from stdin, extracts the command, and returns a `permissionDecision` of `"deny"` if it contains `rm -rf`:

~~45 `PreToolUse`, `PermissionRequest`, and `PostToolUse`)~~

~~46 * Simple strings match exactly: `Write` matches only the Write tool~~

~~47 * Supports regex: `Edit|Write` or `Notebook.*`~~

~~48 * Use `*` to match all tools. You can also use empty string (`""`) or leave~~

~~49 `matcher` blank.~~

~~50* **hooks**: Array of hooks to execute when the pattern matches~~

~~51 * `type`: Hook execution type - `"command"` for bash commands or `"prompt"` for LLM-based evaluation~~

~~52 * `command`: (For `type: "command"`) The bash command to execute (can use `$CLAUDE_PROJECT_DIR` environment variable)~~

~~53 * `prompt`: (For `type: "prompt"`) The prompt to send to the LLM for evaluation~~

~~54 * `timeout`: (Optional) How long a hook should run, in seconds, before canceling that specific hook~~

~~56For events like `UserPromptSubmit`, `Stop`, and `SubagentStop`~~

~~57that don't use matchers, you can omit the matcher field:~~

58 67

~~59```json theme={null}~~68```bash theme={null}

~~60{~~69#!/bin/bash

~~61 "hooks": {~~70# .claude/hooks/block-rm.sh

~~62 "UserPromptSubmit": [~~71COMMAND=$(jq -r '.tool_input.command')

~~63 {~~72

~~64 "hooks": [~~73if echo "$COMMAND" | grep -q 'rm -rf'; then

~~65 {~~74 jq -n '{

~~66 "type": "command",~~75 hookSpecificOutput: {

~~67 "command": "/path/to/prompt-validator.py"~~76 hookEventName: "PreToolUse",

~~68 }~~77 permissionDecision: "deny",

~~69 ]~~78 permissionDecisionReason: "Destructive command blocked by hook"

~~70 }~~

~~71 ]~~

72 }79 }

~~73}~~80 }'

81else

82 exit 0 # allow the command

83fi

74```84```

75 85

~~76### Project-Specific Hook Scripts~~86Now suppose Claude Code decides to run `Bash "rm -rf /tmp/build"`. Here's what happens:

77 87

~~78You can use the environment variable `CLAUDE_PROJECT_DIR` (only available when~~88<Frame>

~~79Claude Code spawns the hook command) to reference scripts stored in your project,~~89 <img src="https://mintcdn.com/claude-code/s7NM0vfd_wres2nf/images/hook-resolution.svg?fit=max&auto=format&n=s7NM0vfd_wres2nf&q=85&s=7c13f51ffcbc37d22a593b27e2f2de72" alt="Hook resolution flow: PreToolUse event fires, matcher checks for Bash match, hook handler runs, result returns to Claude Code" data-og-width="780" width="780" data-og-height="290" height="290" data-path="images/hook-resolution.svg" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/s7NM0vfd_wres2nf/images/hook-resolution.svg?w=280&fit=max&auto=format&n=s7NM0vfd_wres2nf&q=85&s=36a39a07e8bc1995dcb4639e09846905 280w, https://mintcdn.com/claude-code/s7NM0vfd_wres2nf/images/hook-resolution.svg?w=560&fit=max&auto=format&n=s7NM0vfd_wres2nf&q=85&s=6568d90c596c7605bbac2c325b0a0c86 560w, https://mintcdn.com/claude-code/s7NM0vfd_wres2nf/images/hook-resolution.svg?w=840&fit=max&auto=format&n=s7NM0vfd_wres2nf&q=85&s=255a6f68b9475a0e41dbde7b88002dad 840w, https://mintcdn.com/claude-code/s7NM0vfd_wres2nf/images/hook-resolution.svg?w=1100&fit=max&auto=format&n=s7NM0vfd_wres2nf&q=85&s=dcecf8d5edc88cd2bc49deb006d5760d 1100w, https://mintcdn.com/claude-code/s7NM0vfd_wres2nf/images/hook-resolution.svg?w=1650&fit=max&auto=format&n=s7NM0vfd_wres2nf&q=85&s=04fe51bf69ae375e9fd517f18674e35f 1650w, https://mintcdn.com/claude-code/s7NM0vfd_wres2nf/images/hook-resolution.svg?w=2500&fit=max&auto=format&n=s7NM0vfd_wres2nf&q=85&s=b1b76e0b77fddb5c7fa7bf302dacd80b 2500w" />

~~80ensuring they work regardless of Claude's current directory:~~90</Frame>

81 91

~~82```json theme={null}~~92<Steps>

~~83{~~93 <Step title="Event fires">

~~84 "hooks": {~~94 The `PreToolUse` event fires. Claude Code sends the tool input as JSON on stdin to the hook:

~~85 "PostToolUse": [~~95

~~86 {~~96 ```json theme={null}

~~87 "matcher": "Write|Edit",~~97 { "tool_name": "Bash", "tool_input": { "command": "rm -rf /tmp/build" }, ... }

~~88 "hooks": [~~98 ```

99 </Step>

100

101 <Step title="Matcher checks">

102 The matcher `"Bash"` matches the tool name, so `block-rm.sh` runs. If you omit the matcher or use `"*"`, the hook runs on every occurrence of the event. Hooks only skip when a matcher is defined and doesn't match.

103 </Step>

104

105 <Step title="Hook handler runs">

106 The script extracts `"rm -rf /tmp/build"` from the input and finds `rm -rf`, so it prints a decision to stdout:

107

108 ```json theme={null}

89 {109 {

~~90 "type": "command",~~110 "hookSpecificOutput": {

~~91 "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/check-style.sh"~~111 "hookEventName": "PreToolUse",

~~92 }~~112 "permissionDecision": "deny",

~~93 ]~~113 "permissionDecisionReason": "Destructive command blocked by hook"

94 }114 }

~~95 ]~~

96 }115 }

~~97}~~116 ```

~~98```~~117

118 If the command had been safe (like `npm test`), the script would hit `exit 0` instead, which tells Claude Code to allow the tool call with no further action.

119 </Step>

99 120

100### Plugin hooks121 <Step title="Claude Code acts on the result">

122 Claude Code reads the JSON decision, blocks the tool call, and shows Claude the reason.

123 </Step>

124</Steps>

101 125

102[Plugins](/en/plugins) can provide hooks that integrate seamlessly with your user and project hooks. Plugin hooks are automatically merged with your configuration when plugins are enabled.126The [Configuration](#configuration) section below documents the full schema, and each [hook event](#hook-events) section documents what input your command receives and what output it can return.

103 127

104**How plugin hooks work**:128## Configuration

105 129

106* Plugin hooks are defined in the plugin's `hooks/hooks.json` file or in a file given by a custom path to the `hooks` field.130Hooks are defined in JSON settings files. The configuration has three levels of nesting:

107* When a plugin is enabled, its hooks are merged with user and project hooks

108* Multiple hooks from different sources can respond to the same event

109* Plugin hooks use the `${CLAUDE_PLUGIN_ROOT}` environment variable to reference plugin files

110 131

111**Example plugin hook configuration**:1321. Choose a [hook event](#hook-events) to respond to, like `PreToolUse` or `Stop`

1332. Add a [matcher group](#matcher-patterns) to filter when it fires, like "only for the Bash tool"

1343. Define one or more [hook handlers](#hook-handler-fields) to run when matched

135

136See [How a hook resolves](#how-a-hook-resolves) above for a complete walkthrough with an annotated example.

137

138<Note>

139 This page uses specific terms for each level: **hook event** for the lifecycle point, **matcher group** for the filter, and **hook handler** for the shell command, prompt, or agent that runs. "Hook" on its own refers to the general feature.

140</Note>

141

142### Hook locations

143

144Where you define a hook determines its scope:

145

146| Location | Scope | Shareable |

147| :--------------------------------------------------------- | :---------------------------- | :--------------------------------- |

148| `~/.claude/settings.json` | All your projects | No, local to your machine |

149| `.claude/settings.json` | Single project | Yes, can be committed to the repo |

150| `.claude/settings.local.json` | Single project | No, gitignored |

151| Managed policy settings | Organization-wide | Yes, admin-controlled |

152| [Plugin](/en/plugins) `hooks/hooks.json` | When plugin is enabled | Yes, bundled with the plugin |

153| [Skill](/en/skills) or [agent](/en/sub-agents) frontmatter | While the component is active | Yes, defined in the component file |

154

155For details on settings file resolution, see [settings](/en/settings). Enterprise administrators can use `allowManagedHooksOnly` to block user, project, and plugin hooks. See [Hook configuration](/en/settings#hook-configuration).

156

157### Matcher patterns

158

159The `matcher` field is a regex string that filters when hooks fire. Use `"*"`, `""`, or omit `matcher` entirely to match all occurrences. Each event type matches on a different field:

160

161| Event | What the matcher filters | Example matcher values |

162| :--------------------------------------------------------------------- | :------------------------ | :----------------------------------------------------------------------------- |

164| `SessionStart` | how the session started | `startup`, `resume`, `clear`, `compact` |

165| `SessionEnd` | why the session ended | `clear`, `logout`, `prompt_input_exit`, `bypass_permissions_disabled`, `other` |

166| `Notification` | notification type | `permission_prompt`, `idle_prompt`, `auth_success`, `elicitation_dialog` |

167| `SubagentStart` | agent type | `Bash`, `Explore`, `Plan`, or custom agent names |

168| `PreCompact` | what triggered compaction | `manual`, `auto` |

169| `SubagentStop` | agent type | same values as `SubagentStart` |

170| `UserPromptSubmit`, `Stop`, `TeammateIdle`, `TaskCompleted` | no matcher support | always fires on every occurrence |

171

172The matcher is a regex, so `Edit|Write` matches either tool and `Notebook.*` matches any tool starting with Notebook. The matcher runs against a field from the [JSON input](#hook-input-and-output) that Claude Code sends to your hook on stdin. For tool events, that field is `tool_name`. Each [hook event](#hook-events) section lists the full set of matcher values and the input schema for that event.

173

174This example runs a linting script only when Claude writes or edits a file:

112 175

113```json theme={null}176```json theme={null}

114{177{

115 "description": "Automatic code formatting",

116 "hooks": {178 "hooks": {

117 "PostToolUse": [179 "PostToolUse": [

118 {180 {

119 "matcher": "Write|Edit",181 "matcher": "Edit|Write",

120 "hooks": [182 "hooks": [

121 {183 {

122 "type": "command",184 "type": "command",

123 "command": "${CLAUDE_PLUGIN_ROOT}/scripts/format.sh",185 "command": "/path/to/lint-check.sh"

124 "timeout": 30

125 }186 }

126 ]187 ]

127 }188 }

130}191}

131```192```

132 193

133<Note>194`UserPromptSubmit` and `Stop` don't support matchers and always fire on every occurrence. If you add a `matcher` field to these events, it is silently ignored.

134 Plugin hooks use the same format as regular hooks with an optional `description` field to explain the hook's purpose.

135</Note>

136 195

137<Note>196#### Match MCP tools

138 Plugin hooks run alongside your custom hooks. If multiple hooks match an event, they all execute in parallel.

139</Note>

140 197

141**Environment variables for plugins**:198[MCP](/en/mcp) server tools appear as regular tools in tool events (`PreToolUse`, `PostToolUse`, `PostToolUseFailure`, `PermissionRequest`), so you can match them the same way you match any other tool name.

142 199

143* `${CLAUDE_PLUGIN_ROOT}`: Absolute path to the plugin directory200MCP tools follow the naming pattern `mcp__<server>__<tool>`, for example:

144* `${CLAUDE_PROJECT_DIR}`: Project root directory (same as for project hooks)

145* All standard environment variables are available

146 201

147See the [plugin components reference](/en/plugins-reference#hooks) for details on creating plugin hooks.202* `mcp__memory__create_entities`: Memory server's create entities tool

203* `mcp__filesystem__read_file`: Filesystem server's read file tool

204* `mcp__github__search_repositories`: GitHub server's search tool

148 205

149## Prompt-Based Hooks206Use regex patterns to target specific MCP tools or groups of tools:

150 207

151In addition to bash command hooks (`type: "command"`), Claude Code supports prompt-based hooks (`type: "prompt"`) that use an LLM to evaluate whether to allow or block an action. Prompt-based hooks are currently only supported for `Stop` and `SubagentStop` hooks, where they enable intelligent, context-aware decisions.208* `mcp__memory__.*` matches all tools from the `memory` server

209* `mcp__.*__write.*` matches any tool containing "write" from any server

152 210

153### How prompt-based hooks work211This example logs all memory server operations and validates write operations from any MCP server:

~~154~~

155Instead of executing a bash command, prompt-based hooks:

~~156~~

1571. Send the hook input and your prompt to a fast LLM (Haiku)

1582. The LLM responds with structured JSON containing a decision

1593. Claude Code processes the decision automatically

~~160~~

161### Configuration

162 212

163```json theme={null}213```json theme={null}

164{214{

165 "hooks": {215 "hooks": {

166 "Stop": [216 "PreToolUse": [

167 {217 {

218 "matcher": "mcp__memory__.*",

168 "hooks": [219 "hooks": [

169 {220 {

170 "type": "prompt",221 "type": "command",

171 "prompt": "Evaluate if Claude should stop: $ARGUMENTS. Check if all tasks are complete."222 "command": "echo 'Memory operation initiated' >> ~/mcp-operations.log"

223 }

224 ]

225 },

226 {

227 "matcher": "mcp__.*__write.*",

228 "hooks": [

229 {

230 "type": "command",

231 "command": "/home/user/scripts/validate-mcp-write.py"

172 }232 }

173 ]233 ]

174 }234 }

177}237}

178```238```

179 239

180**Fields:**240### Hook handler fields

181 241

182* `type`: Must be `"prompt"`242Each object in the inner `hooks` array is a hook handler: the shell command, LLM prompt, or agent that runs when the matcher matches. There are three types:

183* `prompt`: The prompt text to send to the LLM

184 * Use `$ARGUMENTS` as a placeholder for the hook input JSON

185 * If `$ARGUMENTS` is not present, input JSON is appended to the prompt

186* `timeout`: (Optional) Timeout in seconds (default: 30 seconds)

187 243

188### Response schema244* **[Command hooks](#command-hook-fields)** (`type: "command"`): run a shell command. Your script receives the event's [JSON input](#hook-input-and-output) on stdin and communicates results back through exit codes and stdout.

245* **[Prompt hooks](#prompt-and-agent-hook-fields)** (`type: "prompt"`): send a prompt to a Claude model for single-turn evaluation. The model returns a yes/no decision as JSON. See [Prompt-based hooks](#prompt-based-hooks).

246* **[Agent hooks](#prompt-and-agent-hook-fields)** (`type: "agent"`): spawn a subagent that can use tools like Read, Grep, and Glob to verify conditions before returning a decision. See [Agent-based hooks](#agent-based-hooks).

189 247

190The LLM must respond with JSON containing:248#### Common fields

191 249

192```json theme={null}250These fields apply to all hook types:

193{

194 "decision": "approve" | "block",

195 "reason": "Explanation for the decision",

196 "continue": false, // Optional: stops Claude entirely

197 "stopReason": "Message shown to user", // Optional: custom stop message

198 "systemMessage": "Warning or context" // Optional: shown to user

199}

200```

201 251

202**Response fields:**252| Field | Required | Description |

253| :-------------- | :------- | :-------------------------------------------------------------------------------------------------------------------------------------------- |

254| `type` | yes | `"command"`, `"prompt"`, or `"agent"` |

255| `timeout` | no | Seconds before canceling. Defaults: 600 for command, 30 for prompt, 60 for agent |

256| `statusMessage` | no | Custom spinner message displayed while the hook runs |

257| `once` | no | If `true`, runs only once per session then is removed. Skills only, not agents. See [Hooks in skills and agents](#hooks-in-skills-and-agents) |

203 258

204* `decision`: `"approve"` allows the action, `"block"` prevents it259#### Command hook fields

205* `reason`: Explanation shown to Claude when decision is `"block"`

206* `continue`: (Optional) If `false`, stops Claude's execution entirely

207* `stopReason`: (Optional) Message shown when `continue` is false

208* `systemMessage`: (Optional) Additional message shown to the user

209 260

210### Supported hook events261In addition to the [common fields](#common-fields), command hooks accept these fields:

211 262

212Prompt-based hooks work with any hook event, but are most useful for:263| Field | Required | Description |

264| :-------- | :------- | :------------------------------------------------------------------------------------------------------------------ |

265| `command` | yes | Shell command to execute |

266| `async` | no | If `true`, runs in the background without blocking. See [Run hooks in the background](#run-hooks-in-the-background) |

213 267

214* **Stop**: Intelligently decide if Claude should continue working268#### Prompt and agent hook fields

215* **SubagentStop**: Evaluate if a subagent has completed its task

216* **UserPromptSubmit**: Validate user prompts with LLM assistance

217* **PreToolUse**: Make context-aware permission decisions

218* **PermissionRequest**: Intelligently allow or deny permission dialogs

219 269

220### Example: Intelligent Stop hook270In addition to the [common fields](#common-fields), prompt and agent hooks accept these fields:

221 271

222```json theme={null}272| Field | Required | Description |

223{273| :------- | :------- | :------------------------------------------------------------------------------------------ |

274| `prompt` | yes | Prompt text to send to the model. Use `$ARGUMENTS` as a placeholder for the hook input JSON |

275| `model` | no | Model to use for evaluation. Defaults to a fast model |

276

277All matching hooks run in parallel, and identical handlers are deduplicated automatically. Handlers run in the current directory with Claude Code's environment. The `$CLAUDE_CODE_REMOTE` environment variable is set to `"true"` in remote web environments and not set in the local CLI.

278

279### Reference scripts by path

280

281Use environment variables to reference hook scripts relative to the project or plugin root, regardless of the working directory when the hook runs:

282

283* `$CLAUDE_PROJECT_DIR`: the project root. Wrap in quotes to handle paths with spaces.

284* `${CLAUDE_PLUGIN_ROOT}`: the plugin's root directory, for scripts bundled with a [plugin](/en/plugins).

285

286<Tabs>

287 <Tab title="Project scripts">

288 This example uses `$CLAUDE_PROJECT_DIR` to run a style checker from the project's `.claude/hooks/` directory after any `Write` or `Edit` tool call:

289

290 ```json theme={null}

291 {

224 "hooks": {292 "hooks": {

225 "Stop": [293 "PostToolUse": [

226 {294 {

295 "matcher": "Write|Edit",

227 "hooks": [296 "hooks": [

228 {297 {

229 "type": "prompt",298 "type": "command",

230 "prompt": "You are evaluating whether Claude should stop working. Context: $ARGUMENTS\n\nAnalyze the conversation and determine if:\n1. All user-requested tasks are complete\n2. Any errors need to be addressed\n3. Follow-up work is needed\n\nRespond with JSON: {\"decision\": \"approve\" or \"block\", \"reason\": \"your explanation\"}",299 "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/check-style.sh"

231 "timeout": 30

232 }300 }

233 ]301 ]

234 }302 }

235 ]303 ]

236 }304 }

237}305 }

238```306 ```

307 </Tab>

239 308

240### Example: SubagentStop with custom logic309 <Tab title="Plugin scripts">

310 Define plugin hooks in `hooks/hooks.json` with an optional top-level `description` field. When a plugin is enabled, its hooks merge with your user and project hooks.

241 311

242```json theme={null}312 This example runs a formatting script bundled with the plugin:

243{313

314 ```json theme={null}

315 {

316 "description": "Automatic code formatting",

244 "hooks": {317 "hooks": {

245 "SubagentStop": [318 "PostToolUse": [

246 {319 {

320 "matcher": "Write|Edit",

247 "hooks": [321 "hooks": [

248 {322 {

249 "type": "prompt",323 "type": "command",

250 "prompt": "Evaluate if this subagent should stop. Input: $ARGUMENTS\n\nCheck if:\n- The subagent completed its assigned task\n- Any errors occurred that need fixing\n- Additional context gathering is needed\n\nReturn: {\"decision\": \"approve\" or \"block\", \"reason\": \"explanation\"}"324 "command": "${CLAUDE_PLUGIN_ROOT}/scripts/format.sh",

325 "timeout": 30

251 }326 }

252 ]327 ]

253 }328 }

254 ]329 ]

255 }330 }

256}331 }

332 ```

333

334 See the [plugin components reference](/en/plugins-reference#hooks) for details on creating plugin hooks.

335 </Tab>

336</Tabs>

337

338### Hooks in skills and agents

339

340In addition to settings files and plugins, hooks can be defined directly in [skills](/en/skills) and [subagents](/en/sub-agents) using frontmatter. These hooks are scoped to the component's lifecycle and only run when that component is active.

341

342All hook events are supported. For subagents, `Stop` hooks are automatically converted to `SubagentStop` since that is the event that fires when a subagent completes.

343

344Hooks use the same configuration format as settings-based hooks but are scoped to the component's lifetime and cleaned up when it finishes.

345

346This skill defines a `PreToolUse` hook that runs a security validation script before each `Bash` command:

347

348```yaml theme={null}

349---

350name: secure-operations

351description: Perform operations with security checks

352hooks:

353 PreToolUse:

354 - matcher: "Bash"

355 hooks:

356 - type: command

357 command: "./scripts/security-check.sh"

358---

257```359```

258 360

259### Comparison with bash command hooks361Agents use the same format in their YAML frontmatter.

260 362

261| Feature | Bash Command Hooks | Prompt-Based Hooks |363### The `/hooks` menu

262| --------------------- | ----------------------- | ------------------------------ |

263| **Execution** | Runs bash script | Queries LLM |

264| **Decision logic** | You implement in code | LLM evaluates context |

265| **Setup complexity** | Requires script file | Configure prompt |

266| **Context awareness** | Limited to script logic | Natural language understanding |

267| **Performance** | Fast (local execution) | Slower (API call) |

268| **Use case** | Deterministic rules | Context-aware decisions |

269 364

270### Best practices365Type `/hooks` in Claude Code to open the interactive hooks manager, where you can view, add, and delete hooks without editing settings files directly. For a step-by-step walkthrough, see [Set up your first hook](/en/hooks-guide#set-up-your-first-hook) in the guide.

271 366

272* **Be specific in prompts**: Clearly state what you want the LLM to evaluate367Each hook in the menu is labeled with a bracket prefix indicating its source:

273* **Include decision criteria**: List the factors the LLM should consider

274* **Test your prompts**: Verify the LLM makes correct decisions for your use cases

275* **Set appropriate timeouts**: Default is 30 seconds, adjust if needed

276* **Use for complex decisions**: Bash hooks are better for simple, deterministic rules

277 368

278See the [plugin components reference](/en/plugins-reference#hooks) for details on creating plugin hooks.369* `[User]`: from `~/.claude/settings.json`

370* `[Project]`: from `.claude/settings.json`

371* `[Local]`: from `.claude/settings.local.json`

372* `[Plugin]`: from a plugin's `hooks/hooks.json`, read-only

279 373

280## Hook Events374### Disable or remove hooks

281 375

282### PreToolUse376To remove a hook, delete its entry from the settings JSON file, or use the `/hooks` menu and select the hook to delete it.

283 377

284Runs after Claude creates tool parameters and before processing the tool call.378To temporarily disable all hooks without removing them, set `"disableAllHooks": true` in your settings file or use the toggle in the `/hooks` menu. There is no way to disable an individual hook while keeping it in the configuration.

285 379

286**Common matchers:**380Direct edits to hooks in settings files don't take effect immediately. Claude Code captures a snapshot of hooks at startup and uses it throughout the session. This prevents malicious or accidental hook modifications from taking effect mid-session without your review. If hooks are modified externally, Claude Code warns you and requires review in the `/hooks` menu before changes apply.

287 381

288* `Task` - Subagent tasks (see [subagents documentation](/en/sub-agents))382## Hook input and output

289* `Bash` - Shell commands

290* `Glob` - File pattern matching

291* `Grep` - Content search

292* `Read` - File reading

293* `Edit` - File editing

294* `Write` - File writing

295* `WebFetch`, `WebSearch` - Web operations

296 383

297Use [PreToolUse decision control](#pretooluse-decision-control) to allow, deny, or ask for permission to use the tool.384Hooks receive JSON data via stdin and communicate results through exit codes, stdout, and stderr. This section covers fields and behavior common to all events. Each event's section under [Hook events](#hook-events) includes its specific input schema and decision control options.

298 385

299### PermissionRequest386### Common input fields

300 387

301Runs when the user is shown a permission dialog.388All hook events receive these fields via stdin as JSON, in addition to event-specific fields documented in each [hook event](#hook-events) section:

302Use [PermissionRequest decision control](#permissionrequest-decision-control) to allow or deny on behalf of the user.

303 389

304Recognizes the same matcher values as PreToolUse.390| Field | Description |

391| :---------------- | :----------------------------------------------------------------------------------------------------------------------------------------- |

392| `session_id` | Current session identifier |

393| `transcript_path` | Path to conversation JSON |

394| `cwd` | Current working directory when the hook is invoked |

395| `permission_mode` | Current [permission mode](/en/permissions#permission-modes): `"default"`, `"plan"`, `"acceptEdits"`, `"dontAsk"`, or `"bypassPermissions"` |

396| `hook_event_name` | Name of the event that fired |

305 397

306### PostToolUse398For example, a `PreToolUse` hook for a Bash command receives this on stdin:

307 399

308Runs immediately after a tool completes successfully.400```json theme={null}

401{

402 "session_id": "abc123",

403 "transcript_path": "/home/user/.claude/projects/.../transcript.jsonl",

404 "cwd": "/home/user/my-project",

405 "permission_mode": "default",

406 "hook_event_name": "PreToolUse",

407 "tool_name": "Bash",

408 "tool_input": {

409 "command": "npm test"

410 }

411}

412```

309 413

310Recognizes the same matcher values as PreToolUse.414The `tool_name` and `tool_input` fields are event-specific. Each [hook event](#hook-events) section documents the additional fields for that event.

311 415

312### Notification416### Exit code output

417

418The exit code from your hook command tells Claude Code whether the action should proceed, be blocked, or be ignored.

419

420**Exit 0** means success. Claude Code parses stdout for [JSON output fields](#json-output). JSON output is only processed on exit 0. For most events, stdout is only shown in verbose mode (`Ctrl+O`). The exceptions are `UserPromptSubmit` and `SessionStart`, where stdout is added as context that Claude can see and act on.

313 421

314Runs when Claude Code sends notifications. Supports matchers to filter by notification type.422**Exit 2** means a blocking error. Claude Code ignores stdout and any JSON in it. Instead, stderr text is fed back to Claude as an error message. The effect depends on the event: `PreToolUse` blocks the tool call, `UserPromptSubmit` rejects the prompt, and so on. See [exit code 2 behavior](#exit-code-2-behavior-per-event) for the full list.

315 423

316**Common matchers:**424**Any other exit code** is a non-blocking error. stderr is shown in verbose mode (`Ctrl+O`) and execution continues.

317 425

318* `permission_prompt` - Permission requests from Claude Code426For example, a hook command script that blocks dangerous Bash commands:

319* `idle_prompt` - When Claude is waiting for user input (after 60+ seconds of idle time)427

320* `auth_success` - Authentication success notifications428```bash theme={null}

321* `elicitation_dialog` - When Claude Code needs input for MCP tool elicitation429#!/bin/bash

430# Reads JSON input from stdin, checks the command

431command=$(jq -r '.tool_input.command' < /dev/stdin)

432

433if [[ "$command" == rm* ]]; then

434 echo "Blocked: rm commands are not allowed" >&2

435 exit 2 # Blocking error: tool call is prevented

436fi

437

438exit 0 # Success: tool call proceeds

439```

322 440

323You can use matchers to run different hooks for different notification types, or omit the matcher to run hooks for all notifications.441#### Exit code 2 behavior per event

324 442

325**Example: Different notifications for different types**443Exit code 2 is the way a hook signals "stop, don't do this." The effect depends on the event, because some events represent actions that can be blocked (like a tool call that hasn't happened yet) and others represent things that already happened or can't be prevented.

444

445| Hook event | Can block? | What happens on exit 2 |

446| :------------------- | :--------- | :----------------------------------------------------------------- |

447| `PreToolUse` | Yes | Blocks the tool call |

448| `PermissionRequest` | Yes | Denies the permission |

449| `UserPromptSubmit` | Yes | Blocks prompt processing and erases the prompt |

450| `Stop` | Yes | Prevents Claude from stopping, continues the conversation |

451| `SubagentStop` | Yes | Prevents the subagent from stopping |

452| `TeammateIdle` | Yes | Prevents the teammate from going idle (teammate continues working) |

453| `TaskCompleted` | Yes | Prevents the task from being marked as completed |

454| `PostToolUse` | No | Shows stderr to Claude (tool already ran) |

455| `PostToolUseFailure` | No | Shows stderr to Claude (tool already failed) |

456| `Notification` | No | Shows stderr to user only |

457| `SubagentStart` | No | Shows stderr to user only |

458| `SessionStart` | No | Shows stderr to user only |

459| `SessionEnd` | No | Shows stderr to user only |

460| `PreCompact` | No | Shows stderr to user only |

461

462### JSON output

463

464Exit codes let you allow or block, but JSON output gives you finer-grained control. Instead of exiting with code 2 to block, exit 0 and print a JSON object to stdout. Claude Code reads specific fields from that JSON to control behavior, including [decision control](#decision-control) for blocking, allowing, or escalating to the user.

465

466<Note>

467 You must choose one approach per hook, not both: either use exit codes alone for signaling, or exit 0 and print JSON for structured control. Claude Code only processes JSON on exit 0. If you exit 2, any JSON is ignored.

468</Note>

469

470Your hook's stdout must contain only the JSON object. If your shell profile prints text on startup, it can interfere with JSON parsing. See [JSON validation failed](/en/hooks-guide#json-validation-failed) in the troubleshooting guide.

471

472The JSON object supports three kinds of fields:

473

474* **Universal fields** like `continue` work across all events. These are listed in the table below.

475* **Top-level `decision` and `reason`** are used by some events to block or provide feedback.

476* **`hookSpecificOutput`** is a nested object for events that need richer control. It requires a `hookEventName` field set to the event name.

477

478| Field | Default | Description |

479| :--------------- | :------ | :------------------------------------------------------------------------------------------------------------------------- |

480| `continue` | `true` | If `false`, Claude stops processing entirely after the hook runs. Takes precedence over any event-specific decision fields |

481| `stopReason` | none | Message shown to the user when `continue` is `false`. Not shown to Claude |

482| `suppressOutput` | `false` | If `true`, hides stdout from verbose mode output |

483| `systemMessage` | none | Warning message shown to the user |

484

485To stop Claude entirely regardless of event type:

326 486

327```json theme={null}487```json theme={null}

328{488{ "continue": false, "stopReason": "Build failed, fix errors before continuing" }

329 "hooks": {489```

330 "Notification": [490

331 {491#### Decision control

332 "matcher": "permission_prompt",492

333 "hooks": [493Not every event supports blocking or controlling behavior through JSON. The events that do each use a different set of fields to express that decision. Use this table as a quick reference before writing a hook:

494

495| Events | Decision pattern | Key fields |

496| :-------------------------------------------------------------------- | :------------------- | :---------------------------------------------------------------- |

497| UserPromptSubmit, PostToolUse, PostToolUseFailure, Stop, SubagentStop | Top-level `decision` | `decision: "block"`, `reason` |

498| TeammateIdle, TaskCompleted | Exit code only | Exit code 2 blocks the action, stderr is fed back as feedback |

499| PreToolUse | `hookSpecificOutput` | `permissionDecision` (allow/deny/ask), `permissionDecisionReason` |

500| PermissionRequest | `hookSpecificOutput` | `decision.behavior` (allow/deny) |

501

502Here are examples of each pattern in action:

503

504<Tabs>

505 <Tab title="Top-level decision">

506 Used by `UserPromptSubmit`, `PostToolUse`, `PostToolUseFailure`, `Stop`, and `SubagentStop`. The only value is `"block"`. To allow the action to proceed, omit `decision` from your JSON, or exit 0 without any JSON at all:

507

508 ```json theme={null}

334 {509 {

335 "type": "command",510 "decision": "block",

336 "command": "/path/to/permission-alert.sh"511 "reason": "Test suite must pass before proceeding"

337 }512 }

338 ]513 ```

339 },514 </Tab>

515

516 <Tab title="PreToolUse">

517 Uses `hookSpecificOutput` for richer control: allow, deny, or escalate to the user. You can also modify tool input before it runs or inject additional context for Claude. See [PreToolUse decision control](#pretooluse-decision-control) for the full set of options.

518

519 ```json theme={null}

340 {520 {

341 "matcher": "idle_prompt",521 "hookSpecificOutput": {

342 "hooks": [522 "hookEventName": "PreToolUse",

523 "permissionDecision": "deny",

524 "permissionDecisionReason": "Database writes are not allowed"

525 }

526 }

527 ```

528 </Tab>

529

530 <Tab title="PermissionRequest">

531 Uses `hookSpecificOutput` to allow or deny a permission request on behalf of the user. When allowing, you can also modify the tool's input or apply permission rules so the user isn't prompted again. See [PermissionRequest decision control](#permissionrequest-decision-control) for the full set of options.

532

533 ```json theme={null}

343 {534 {

344 "type": "command",535 "hookSpecificOutput": {

345 "command": "/path/to/idle-notification.sh"536 "hookEventName": "PermissionRequest",

537 "decision": {

538 "behavior": "allow",

539 "updatedInput": {

540 "command": "npm run lint"

346 }541 }

347 ]

348 }542 }

349 ]

350 }543 }

351}544 }

352```545 ```

546 </Tab>

547</Tabs>

353 548

354### UserPromptSubmit549For extended examples including Bash command validation, prompt filtering, and auto-approval scripts, see [What you can automate](/en/hooks-guide#what-you-can-automate) in the guide and the [Bash command validator reference implementation](https://github.com/anthropics/claude-code/blob/main/examples/hooks/bash_command_validator_example.py).

355 550

356Runs when the user submits a prompt, before Claude processes it. This allows you551## Hook events

357to add additional context based on the prompt/conversation, validate prompts, or

358block certain types of prompts.

359 552

360### Stop553Each event corresponds to a point in Claude Code's lifecycle where hooks can run. The sections below are ordered to match the lifecycle: from session setup through the agentic loop to session end. Each section describes when the event fires, what matchers it supports, the JSON input it receives, and how to control behavior through output.

361 554

362Runs when the main Claude Code agent has finished responding. Does not run if555### SessionStart

363the stoppage occurred due to a user interrupt.

364 556

365### SubagentStop557Runs when Claude Code starts a new session or resumes an existing session. Useful for loading development context like existing issues or recent changes to your codebase, or setting up environment variables. For static context that does not require a script, use [CLAUDE.md](/en/memory) instead.

366 558

367Runs when a Claude Code subagent (Task tool call) has finished responding.559SessionStart runs on every session, so keep these hooks fast.

368 560

369### PreCompact561The matcher value corresponds to how the session was initiated:

370 562

371Runs before Claude Code is about to run a compact operation.563| Matcher | When it fires |

564| :-------- | :------------------------------------- |

565| `startup` | New session |

566| `resume` | `--resume`, `--continue`, or `/resume` |

567| `clear` | `/clear` |

568| `compact` | Auto or manual compaction |

372 569

373**Matchers:**570#### SessionStart input

374 571

375* `manual` - Invoked from `/compact`572In addition to the [common input fields](#common-input-fields), SessionStart hooks receive `source`, `model`, and optionally `agent_type`. The `source` field indicates how the session started: `"startup"` for new sessions, `"resume"` for resumed sessions, `"clear"` after `/clear`, or `"compact"` after compaction. The `model` field contains the model identifier. If you start Claude Code with `claude --agent <name>`, an `agent_type` field contains the agent name.

376* `auto` - Invoked from auto-compact (due to full context window)

377 573

378### SessionStart574```json theme={null}

575{

576 "session_id": "abc123",

577 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

578 "cwd": "/Users/...",

579 "permission_mode": "default",

580 "hook_event_name": "SessionStart",

581 "source": "startup",

582 "model": "claude-sonnet-4-5-20250929"

583}

584```

379 585

380Runs when Claude Code starts a new session or resumes an existing session (which586#### SessionStart decision control

381currently does start a new session under the hood). Useful for loading in

382development context like existing issues or recent changes to your codebase, installing dependencies, or setting up environment variables.

383 587

384**Matchers:**588Any text your hook script prints to stdout is added as context for Claude. In addition to the [JSON output fields](#json-output) available to all hooks, you can return these event-specific fields:

385 589

386* `startup` - Invoked from startup590| Field | Description |

387* `resume` - Invoked from `--resume`, `--continue`, or `/resume`591| :------------------ | :------------------------------------------------------------------------ |

388* `clear` - Invoked from `/clear`592| `additionalContext` | String added to Claude's context. Multiple hooks' values are concatenated |

389* `compact` - Invoked from auto or manual compact.

390 593

391#### Persisting environment variables594```json theme={null}

595{

596 "hookSpecificOutput": {

597 "hookEventName": "SessionStart",

598 "additionalContext": "My additional context here"

599 }

600}

601```

392 602

393SessionStart hooks have access to the `CLAUDE_ENV_FILE` environment variable, which provides a file path where you can persist environment variables for subsequent bash commands.603#### Persist environment variables

394 604

395**Example: Setting individual environment variables**605SessionStart hooks have access to the `CLAUDE_ENV_FILE` environment variable, which provides a file path where you can persist environment variables for subsequent Bash commands.

606

607To set individual environment variables, write `export` statements to `CLAUDE_ENV_FILE`. Use append (`>>`) to preserve variables set by other hooks:

396 608

397```bash theme={null}609```bash theme={null}

398#!/bin/bash610#!/bin/bash

399 611

400if [ -n "$CLAUDE_ENV_FILE" ]; then612if [ -n "$CLAUDE_ENV_FILE" ]; then

401 echo 'export NODE_ENV=production' >> "$CLAUDE_ENV_FILE"613 echo 'export NODE_ENV=production' >> "$CLAUDE_ENV_FILE"

402 echo 'export API_KEY=your-api-key' >> "$CLAUDE_ENV_FILE"614 echo 'export DEBUG_LOG=true' >> "$CLAUDE_ENV_FILE"

403 echo 'export PATH="$PATH:./node_modules/.bin"' >> "$CLAUDE_ENV_FILE"615 echo 'export PATH="$PATH:./node_modules/.bin"' >> "$CLAUDE_ENV_FILE"

404fi616fi

405 617

406exit 0618exit 0

407```619```

408 620

409**Example: Persisting all environment changes from the hook**621To capture all environment changes from setup commands, compare the exported variables before and after:

~~410~~

411When your setup modifies the environment (for example, `nvm use`), capture and persist all changes by diffing the environment:

412 622

413```bash theme={null}623```bash theme={null}

414#!/bin/bash624#!/bin/bash

427exit 0637exit 0

428```638```

429 639

430Any variables written to this file will be available in all subsequent bash commands that Claude Code executes during the session.640Any variables written to this file will be available in all subsequent Bash commands that Claude Code executes during the session.

431 641

432<Note>642<Note>

433 `CLAUDE_ENV_FILE` is only available for SessionStart hooks. Other hook types do not have access to this variable.643 `CLAUDE_ENV_FILE` is available for SessionStart hooks. Other hook types do not have access to this variable.

434</Note>644</Note>

435 645

436### SessionEnd646### UserPromptSubmit

437 647

438Runs when a Claude Code session ends. Useful for cleanup tasks, logging session648Runs when the user submits a prompt, before Claude processes it. This allows you

439statistics, or saving session state.649to add additional context based on the prompt/conversation, validate prompts, or

650block certain types of prompts.

651

652#### UserPromptSubmit input

653

654In addition to the [common input fields](#common-input-fields), UserPromptSubmit hooks receive the `prompt` field containing the text the user submitted.

655

656```json theme={null}

657{

658 "session_id": "abc123",

659 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

660 "cwd": "/Users/...",

661 "permission_mode": "default",

662 "hook_event_name": "UserPromptSubmit",

663 "prompt": "Write a function to calculate the factorial of a number"

664}

665```

666

667#### UserPromptSubmit decision control

668

669`UserPromptSubmit` hooks can control whether a user prompt is processed and add context. All [JSON output fields](#json-output) are available.

670

671There are two ways to add context to the conversation on exit code 0:

672

673* **Plain text stdout**: any non-JSON text written to stdout is added as context

674* **JSON with `additionalContext`**: use the JSON format below for more control. The `additionalContext` field is added as context

675

676Plain stdout is shown as hook output in the transcript. The `additionalContext` field is added more discretely.

677

678To block a prompt, return a JSON object with `decision` set to `"block"`:

679

680| Field | Description |

681| :------------------ | :----------------------------------------------------------------------------------------------------------------- |

682| `decision` | `"block"` prevents the prompt from being processed and erases it from context. Omit to allow the prompt to proceed |

683| `reason` | Shown to the user when `decision` is `"block"`. Not added to context |

684| `additionalContext` | String added to Claude's context |

685

686```json theme={null}

687{

688 "decision": "block",

689 "reason": "Explanation for decision",

690 "hookSpecificOutput": {

691 "hookEventName": "UserPromptSubmit",

692 "additionalContext": "My additional context here"

693 }

694}

695```

696

697<Note>

698 The JSON format isn't required for simple use cases. To add context, you can print plain text to stdout with exit code 0. Use JSON when you need to

699 block prompts or want more structured control.

700</Note>

701

702### PreToolUse

703

704Runs after Claude creates tool parameters and before processing the tool call. Matches on tool name: `Bash`, `Edit`, `Write`, `Read`, `Glob`, `Grep`, `Task`, `WebFetch`, `WebSearch`, and any [MCP tool names](#match-mcp-tools).

705

706Use [PreToolUse decision control](#pretooluse-decision-control) to allow, deny, or ask for permission to use the tool.

707

708#### PreToolUse input

709

710In addition to the [common input fields](#common-input-fields), PreToolUse hooks receive `tool_name`, `tool_input`, and `tool_use_id`. The `tool_input` fields depend on the tool:

711

712##### Bash

713

714Executes shell commands.

715

717| :------------------ | :------ | :----------------- | :-------------------------------------------- |

722

723##### Write

724

725Creates or overwrites a file.

726

728| :---------- | :----- | :-------------------- | :--------------------------------- |

731

732##### Edit

733

734Replaces a string in an existing file.

735

737| :------------ | :------ | :-------------------- | :--------------------------------- |

742

743##### Read

744

745Reads file contents.

746

748| :---------- | :----- | :-------------------- | :----------------------------------------- |

752

753##### Glob

754

755Finds files matching a glob pattern.

756

758| :-------- | :----- | :--------------- | :--------------------------------------------------------------------- |

759| `pattern` | string | `"**/*.ts"` | Glob pattern to match files against |

761

762##### Grep

763

764Searches file contents with regular expressions.

765

767| :------------ | :------ | :--------------- | :------------------------------------------------------------------------------------ |

774

775##### WebFetch

776

777Fetches and processes web content.

778

780| :------- | :----- | :---------------------------- | :----------------------------------- |

783

784##### WebSearch

785

786Searches the web.

787

789| :---------------- | :----- | :----------------------------- | :------------------------------------------------ |

793

794##### Task

795

796Spawns a [subagent](/en/sub-agents).

440 797

799| :-------------- | :----- | :------------------------- | :------------------------------------------- |

442 804

443* `clear` - Session cleared with /clear command805#### PreToolUse decision control

444* `logout` - User logged out

445* `prompt_input_exit` - User exited while prompt input was visible

446* `other` - Other exit reasons

447 806

448## Hook Input807`PreToolUse` hooks can control whether a tool call proceeds. Unlike other hooks that use a top-level `decision` field, PreToolUse returns its decision inside a `hookSpecificOutput` object. This gives it richer control: three outcomes (allow, deny, or ask) plus the ability to modify tool input before execution.

449 808

450Hooks receive JSON data via stdin containing session information and809| Field | Description |

451event-specific data:810| :------------------------- | :----------------------------------------------------------------------------------------------------------------------------------------------- |

811| `permissionDecision` | `"allow"` bypasses the permission system, `"deny"` prevents the tool call, `"ask"` prompts the user to confirm |

812| `permissionDecisionReason` | For `"allow"` and `"ask"`, shown to the user but not Claude. For `"deny"`, shown to Claude |

813| `updatedInput` | Modifies the tool's input parameters before execution. Combine with `"allow"` to auto-approve, or `"ask"` to show the modified input to the user |

814| `additionalContext` | String added to Claude's context before the tool executes |

452 815

453```typescript theme={null}816```json theme={null}

454{817{

455 // Common fields818 "hookSpecificOutput": {

456 session_id: string819 "hookEventName": "PreToolUse",

457 transcript_path: string // Path to conversation JSON820 "permissionDecision": "allow",

458 cwd: string // The current working directory when the hook is invoked821 "permissionDecisionReason": "My reason here",

459 permission_mode: string // Current permission mode: "default", "plan", "acceptEdits", "dontAsk", or "bypassPermissions"822 "updatedInput": {

~~460~~ 823 "field_to_modify": "new value"

461 // Event-specific fields824 },

462 hook_event_name: string825 "additionalContext": "Current environment: production. Proceed with caution."

463 ...826 }

827}

828```

829

830<Note>

831 PreToolUse previously used top-level `decision` and `reason` fields, but these are deprecated for this event. Use `hookSpecificOutput.permissionDecision` and `hookSpecificOutput.permissionDecisionReason` instead. The deprecated values `"approve"` and `"block"` map to `"allow"` and `"deny"` respectively. Other events like PostToolUse and Stop continue to use top-level `decision` and `reason` as their current format.

832</Note>

833

834### PermissionRequest

835

836Runs when the user is shown a permission dialog.

837Use [PermissionRequest decision control](#permissionrequest-decision-control) to allow or deny on behalf of the user.

838

839Matches on tool name, same values as PreToolUse.

840

841#### PermissionRequest input

842

843PermissionRequest hooks receive `tool_name` and `tool_input` fields like PreToolUse hooks, but without `tool_use_id`. An optional `permission_suggestions` array contains the "always allow" options the user would normally see in the permission dialog. The difference is when the hook fires: PermissionRequest hooks run when a permission dialog is about to be shown to the user, while PreToolUse hooks run before tool execution regardless of permission status.

844

845```json theme={null}

846{

847 "session_id": "abc123",

848 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

849 "cwd": "/Users/...",

850 "permission_mode": "default",

851 "hook_event_name": "PermissionRequest",

852 "tool_name": "Bash",

853 "tool_input": {

854 "command": "rm -rf node_modules",

855 "description": "Remove node_modules directory"

856 },

857 "permission_suggestions": [

858 { "type": "toolAlwaysAllow", "tool": "Bash" }

859 ]

860}

861```

862

863#### PermissionRequest decision control

864

865`PermissionRequest` hooks can allow or deny permission requests. In addition to the [JSON output fields](#json-output) available to all hooks, your hook script can return a `decision` object with these event-specific fields:

866

867| Field | Description |

868| :------------------- | :------------------------------------------------------------------------------------------------------------- |

869| `behavior` | `"allow"` grants the permission, `"deny"` denies it |

870| `updatedInput` | For `"allow"` only: modifies the tool's input parameters before execution |

871| `updatedPermissions` | For `"allow"` only: applies permission rule updates, equivalent to the user selecting an "always allow" option |

872| `message` | For `"deny"` only: tells Claude why the permission was denied |

873| `interrupt` | For `"deny"` only: if `true`, stops Claude |

874

875```json theme={null}

876{

877 "hookSpecificOutput": {

878 "hookEventName": "PermissionRequest",

879 "decision": {

880 "behavior": "allow",

881 "updatedInput": {

882 "command": "npm run lint"

883 }

884 }

885 }

886}

887```

888

889### PostToolUse

890

891Runs immediately after a tool completes successfully.

892

893Matches on tool name, same values as PreToolUse.

894

895#### PostToolUse input

896

897`PostToolUse` hooks fire after a tool has already executed successfully. The input includes both `tool_input`, the arguments sent to the tool, and `tool_response`, the result it returned. The exact schema for both depends on the tool.

898

899```json theme={null}

900{

901 "session_id": "abc123",

902 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

903 "cwd": "/Users/...",

904 "permission_mode": "default",

905 "hook_event_name": "PostToolUse",

906 "tool_name": "Write",

907 "tool_input": {

908 "file_path": "/path/to/file.txt",

909 "content": "file content"

910 },

911 "tool_response": {

912 "filePath": "/path/to/file.txt",

913 "success": true

914 },

915 "tool_use_id": "toolu_01ABC123..."

916}

917```

918

919#### PostToolUse decision control

920

921`PostToolUse` hooks can provide feedback to Claude after tool execution. In addition to the [JSON output fields](#json-output) available to all hooks, your hook script can return these event-specific fields:

922

923| Field | Description |

924| :--------------------- | :----------------------------------------------------------------------------------------- |

925| `decision` | `"block"` prompts Claude with the `reason`. Omit to allow the action to proceed |

926| `reason` | Explanation shown to Claude when `decision` is `"block"` |

927| `additionalContext` | Additional context for Claude to consider |

928| `updatedMCPToolOutput` | For [MCP tools](#match-mcp-tools) only: replaces the tool's output with the provided value |

929

930```json theme={null}

931{

932 "decision": "block",

933 "reason": "Explanation for decision",

934 "hookSpecificOutput": {

935 "hookEventName": "PostToolUse",

936 "additionalContext": "Additional information for Claude"

937 }

938}

939```

940

941### PostToolUseFailure

942

943Runs when a tool execution fails. This event fires for tool calls that throw errors or return failure results. Use this to log failures, send alerts, or provide corrective feedback to Claude.

944

945Matches on tool name, same values as PreToolUse.

946

947#### PostToolUseFailure input

948

949PostToolUseFailure hooks receive the same `tool_name` and `tool_input` fields as PostToolUse, along with error information as top-level fields:

950

951```json theme={null}

952{

953 "session_id": "abc123",

954 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

955 "cwd": "/Users/...",

956 "permission_mode": "default",

957 "hook_event_name": "PostToolUseFailure",

958 "tool_name": "Bash",

959 "tool_input": {

960 "command": "npm test",

961 "description": "Run test suite"

962 },

963 "tool_use_id": "toolu_01ABC123...",

964 "error": "Command exited with non-zero status code 1",

965 "is_interrupt": false

966}

967```

968

969| Field | Description |

970| :------------- | :------------------------------------------------------------------------------ |

971| `error` | String describing what went wrong |

972| `is_interrupt` | Optional boolean indicating whether the failure was caused by user interruption |

973

974#### PostToolUseFailure decision control

975

976`PostToolUseFailure` hooks can provide context to Claude after a tool failure. In addition to the [JSON output fields](#json-output) available to all hooks, your hook script can return these event-specific fields:

977

978| Field | Description |

979| :------------------ | :------------------------------------------------------------ |

980| `additionalContext` | Additional context for Claude to consider alongside the error |

981

982```json theme={null}

983{

984 "hookSpecificOutput": {

985 "hookEventName": "PostToolUseFailure",

986 "additionalContext": "Additional information about the failure for Claude"

987 }

988}

989```

990

991### Notification

992

993Runs when Claude Code sends notifications. Matches on notification type: `permission_prompt`, `idle_prompt`, `auth_success`, `elicitation_dialog`. Omit the matcher to run hooks for all notification types.

994

995Use separate matchers to run different handlers depending on the notification type. This configuration triggers a permission-specific alert script when Claude needs permission approval and a different notification when Claude has been idle:

996

997```json theme={null}

998{

999 "hooks": {

1000 "Notification": [

1001 {

1002 "matcher": "permission_prompt",

1003 "hooks": [

1004 {

1005 "type": "command",

1006 "command": "/path/to/permission-alert.sh"

1007 }

1008 ]

1009 },

1010 {

1011 "matcher": "idle_prompt",

1012 "hooks": [

1013 {

1014 "type": "command",

1015 "command": "/path/to/idle-notification.sh"

1016 }

1017 ]

1018 }

1019 ]

1020 }

464}1021}

465```1022```

466 1023

467### PreToolUse Input1024#### Notification input

468 1025

469The exact schema for `tool_input` depends on the tool.1026In addition to the [common input fields](#common-input-fields), Notification hooks receive `message` with the notification text, an optional `title`, and `notification_type` indicating which type fired.

470 1027

471```json theme={null}1028```json theme={null}

472{1029{

474 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1031 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

475 "cwd": "/Users/...",1032 "cwd": "/Users/...",

476 "permission_mode": "default",1033 "permission_mode": "default",

477 "hook_event_name": "PreToolUse",1034 "hook_event_name": "Notification",

478 "tool_name": "Write",1035 "message": "Claude needs your permission to use Bash",

479 "tool_input": {1036 "title": "Permission needed",

480 "file_path": "/path/to/file.txt",1037 "notification_type": "permission_prompt"

481 "content": "file content"

482 },

483 "tool_use_id": "toolu_01ABC123..."

484}1038}

485```1039```

486 1040

487### PostToolUse Input1041Notification hooks cannot block or modify notifications. In addition to the [JSON output fields](#json-output) available to all hooks, you can return `additionalContext` to add context to the conversation:

1042

1043| Field | Description |

1044| :------------------ | :------------------------------- |

1045| `additionalContext` | String added to Claude's context |

1046

1047### SubagentStart

1048

1049Runs when a Claude Code subagent is spawned via the Task tool. Supports matchers to filter by agent type name (built-in agents like `Bash`, `Explore`, `Plan`, or custom agent names from `.claude/agents/`).

488 1050

489The exact schema for `tool_input` and `tool_response` depends on the tool.1051#### SubagentStart input

1052

1053In addition to the [common input fields](#common-input-fields), SubagentStart hooks receive `agent_id` with the unique identifier for the subagent and `agent_type` with the agent name (built-in agents like `"Bash"`, `"Explore"`, `"Plan"`, or custom agent names).

490 1054

491```json theme={null}1055```json theme={null}

492{1056{

494 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1058 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

495 "cwd": "/Users/...",1059 "cwd": "/Users/...",

496 "permission_mode": "default",1060 "permission_mode": "default",

497 "hook_event_name": "PostToolUse",1061 "hook_event_name": "SubagentStart",

498 "tool_name": "Write",1062 "agent_id": "agent-abc123",

499 "tool_input": {1063 "agent_type": "Explore"

500 "file_path": "/path/to/file.txt",

501 "content": "file content"

502 },

503 "tool_response": {

504 "filePath": "/path/to/file.txt",

505 "success": true

506 },

507 "tool_use_id": "toolu_01ABC123..."

508}1064}

509```1065```

510 1066

511### Notification Input1067SubagentStart hooks cannot block subagent creation, but they can inject context into the subagent. In addition to the [JSON output fields](#json-output) available to all hooks, you can return:

1068

1069| Field | Description |

1070| :------------------ | :------------------------------------- |

1071| `additionalContext` | String added to the subagent's context |

512 1072

513```json theme={null}1073```json theme={null}

514{1074{

515 "session_id": "abc123",1075 "hookSpecificOutput": {

516 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1076 "hookEventName": "SubagentStart",

517 "cwd": "/Users/...",1077 "additionalContext": "Follow security guidelines for this task"

518 "permission_mode": "default",1078 }

519 "hook_event_name": "Notification",

520 "message": "Claude needs your permission to use Bash",

521 "notification_type": "permission_prompt"

522}1079}

523```1080```

524 1081

525### UserPromptSubmit Input1082### SubagentStop

1083

1084Runs when a Claude Code subagent has finished responding. Matches on agent type, same values as SubagentStart.

1085

1086#### SubagentStop input

1087

1088In addition to the [common input fields](#common-input-fields), SubagentStop hooks receive `stop_hook_active`, `agent_id`, `agent_type`, and `agent_transcript_path`. The `agent_type` field is the value used for matcher filtering. The `transcript_path` is the main session's transcript, while `agent_transcript_path` is the subagent's own transcript stored in a nested `subagents/` folder.

526 1089

527```json theme={null}1090```json theme={null}

528{1091{

529 "session_id": "abc123",1092 "session_id": "abc123",

530 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1093 "transcript_path": "~/.claude/projects/.../abc123.jsonl",

531 "cwd": "/Users/...",1094 "cwd": "/Users/...",

532 "permission_mode": "default",1095 "permission_mode": "default",

533 "hook_event_name": "UserPromptSubmit",1096 "hook_event_name": "SubagentStop",

534 "prompt": "Write a function to calculate the factorial of a number"1097 "stop_hook_active": false,

1098 "agent_id": "def456",

1099 "agent_type": "Explore",

1100 "agent_transcript_path": "~/.claude/projects/.../abc123/subagents/agent-def456.jsonl"

535}1101}

536```1102```

537 1103

538### Stop and SubagentStop Input1104SubagentStop hooks use the same decision control format as [Stop hooks](#stop-decision-control).

1105

1106### Stop

1107

1108Runs when the main Claude Code agent has finished responding. Does not run if

1109the stoppage occurred due to a user interrupt.

539 1110

540`stop_hook_active` is true when Claude Code is already continuing as a result of1111#### Stop input

541a stop hook. Check this value or process the transcript to prevent Claude Code1112

542from running indefinitely.1113In addition to the [common input fields](#common-input-fields), Stop hooks receive `stop_hook_active`. This field is `true` when Claude Code is already continuing as a result of a stop hook. Check this value or process the transcript to prevent Claude Code from running indefinitely.

543 1114

544```json theme={null}1115```json theme={null}

545{1116{

546 "session_id": "abc123",1117 "session_id": "abc123",

547 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1118 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

1119 "cwd": "/Users/...",

548 "permission_mode": "default",1120 "permission_mode": "default",

549 "hook_event_name": "Stop",1121 "hook_event_name": "Stop",

550 "stop_hook_active": true1122 "stop_hook_active": true

551}1123}

552```1124```

553 1125

554### PreCompact Input1126#### Stop decision control

555 1127

556For `manual`, `custom_instructions` comes from what the user passes into1128`Stop` and `SubagentStop` hooks can control whether Claude continues. In addition to the [JSON output fields](#json-output) available to all hooks, your hook script can return these event-specific fields:

557`/compact`. For `auto`, `custom_instructions` is empty.1129

1130| Field | Description |

1131| :--------- | :------------------------------------------------------------------------- |

1132| `decision` | `"block"` prevents Claude from stopping. Omit to allow Claude to stop |

1133| `reason` | Required when `decision` is `"block"`. Tells Claude why it should continue |

558 1134

559```json theme={null}1135```json theme={null}

560{1136{

561 "session_id": "abc123",1137 "decision": "block",

562 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1138 "reason": "Must be provided when Claude is blocked from stopping"

563 "permission_mode": "default",

564 "hook_event_name": "PreCompact",

565 "trigger": "manual",

566 "custom_instructions": ""

567}1139}

568```1140```

569 1141

570### SessionStart Input1142### TeammateIdle

571 1143

572```json theme={null}1144Runs when an [agent team](/en/agent-teams) teammate is about to go idle after finishing its turn. Use this to enforce quality gates before a teammate stops working, such as requiring passing lint checks or verifying that output files exist.

573{1145

574 "session_id": "abc123",1146When a `TeammateIdle` hook exits with code 2, the teammate receives the stderr message as feedback and continues working instead of going idle. TeammateIdle hooks do not support matchers and fire on every occurrence.

575 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

576 "permission_mode": "default",

577 "hook_event_name": "SessionStart",

578 "source": "startup"

579}

580```

581 1147

582### SessionEnd Input1148#### TeammateIdle input

1149

1150In addition to the [common input fields](#common-input-fields), TeammateIdle hooks receive `teammate_name` and `team_name`.

583 1151

584```json theme={null}1152```json theme={null}

585{1153{

586 "session_id": "abc123",1154 "session_id": "abc123",

587 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1155 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

588 "cwd": "/Users/...",1156 "cwd": "/Users/...",

589 "permission_mode": "default",1157 "permission_mode": "default",

590 "hook_event_name": "SessionEnd",1158 "hook_event_name": "TeammateIdle",

591 "reason": "exit"1159 "teammate_name": "researcher",

1160 "team_name": "my-project"

592}1161}

593```1162```

594 1163

595## Hook Output1164| Field | Description |

~~596~~ 1165| :-------------- | :-------------------------------------------- |

597There are two mutually exclusive ways for hooks to return output back to Claude Code. The output1166| `teammate_name` | Name of the teammate that is about to go idle |

598communicates whether to block and any feedback that should be shown to Claude1167| `team_name` | Name of the team |

599and the user.

~~600~~

601### Simple: Exit Code

602 1168

603Hooks communicate status through exit codes, stdout, and stderr:1169#### TeammateIdle decision control

604 1170

605* **Exit code 0**: Success. `stdout` is shown to the user in verbose mode1171TeammateIdle hooks use exit codes only, not JSON decision control. This example checks that a build artifact exists before allowing a teammate to go idle:

606 (ctrl+o), except for `UserPromptSubmit` and `SessionStart`, where stdout is

607 added to the context. JSON output in `stdout` is parsed for structured control

608 (see [Advanced: JSON Output](#advanced-json-output)).

609* **Exit code 2**: Blocking error. Only `stderr` is used as the error message

610 and fed back to Claude. The format is `[command]: {stderr}`. JSON in `stdout`

611 is **not** processed for exit code 2. See per-hook-event behavior below.

612* **Other exit codes**: Non-blocking error. `stderr` is shown to the user in verbose mode (ctrl+o) with

613 format `Failed with non-blocking status code: {stderr}`. If `stderr` is empty,

614 it shows `No stderr output`. Execution continues.

615 1172

616<Warning>1173```bash theme={null}

617 Reminder: Claude Code does not see stdout if the exit code is 0, except for1174#!/bin/bash

618 the `UserPromptSubmit` hook where stdout is injected as context.

619</Warning>

620 1175

621#### Exit Code 2 Behavior1176if [ ! -f "./dist/output.js" ]; then

1177 echo "Build artifact missing. Run the build before stopping." >&2

1178 exit 2

1179fi

622 1180

623| Hook Event | Behavior |1181exit 0

624| ------------------- | ------------------------------------------------------------------ |1182```

625| `PreToolUse` | Blocks the tool call, shows stderr to Claude |

626| `PermissionRequest` | Denies the permission, shows stderr to Claude |

627| `PostToolUse` | Shows stderr to Claude (tool already ran) |

628| `Notification` | N/A, shows stderr to user only |

629| `UserPromptSubmit` | Blocks prompt processing, erases prompt, shows stderr to user only |

630| `Stop` | Blocks stoppage, shows stderr to Claude |

631| `SubagentStop` | Blocks stoppage, shows stderr to Claude subagent |

632| `PreCompact` | N/A, shows stderr to user only |

633| `SessionStart` | N/A, shows stderr to user only |

634| `SessionEnd` | N/A, shows stderr to user only |

635 1183

636### Advanced: JSON Output1184### TaskCompleted

637 1185

638Hooks can return structured JSON in `stdout` for more sophisticated control.1186Runs when a task is being marked as completed. This fires in two situations: when any agent explicitly marks a task as completed through the TaskUpdate tool, or when an [agent team](/en/agent-teams) teammate finishes its turn with in-progress tasks. Use this to enforce completion criteria like passing tests or lint checks before a task can close.

639 1187

640<Warning>1188When a `TaskCompleted` hook exits with code 2, the task is not marked as completed and the stderr message is fed back to the model as feedback. TaskCompleted hooks do not support matchers and fire on every occurrence.

641 JSON output is only processed when the hook exits with code 0. If your hook

642 exits with code 2 (blocking error), `stderr` text is used directly—any JSON in `stdout`

643 is ignored. For other non-zero exit codes, only `stderr` is shown to the user in verbose mode (ctrl+o).

644</Warning>

645 1189

646#### Common JSON Fields1190#### TaskCompleted input

647 1191

648All hook types can include these optional fields:1192In addition to the [common input fields](#common-input-fields), TaskCompleted hooks receive `task_id`, `task_subject`, and optionally `task_description`, `teammate_name`, and `team_name`.

649 1193

650```json theme={null}1194```json theme={null}

651{1195{

652 "continue": true, // Whether Claude should continue after hook execution (default: true)1196 "session_id": "abc123",

653 "stopReason": "string", // Message shown when continue is false1197 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

~~654~~ 1198 "cwd": "/Users/...",

655 "suppressOutput": true, // Hide stdout from transcript mode (default: false)1199 "permission_mode": "default",

656 "systemMessage": "string" // Optional warning message shown to the user1200 "hook_event_name": "TaskCompleted",

1201 "task_id": "task-001",

1202 "task_subject": "Implement user authentication",

1203 "task_description": "Add login and signup endpoints",

1204 "teammate_name": "implementer",

1205 "team_name": "my-project"

657}1206}

658```1207```

659 1208

660If `continue` is false, Claude stops processing after the hooks run.1209| Field | Description |

1210| :----------------- | :------------------------------------------------------ |

1211| `task_id` | Identifier of the task being completed |

1212| `task_subject` | Title of the task |

1213| `task_description` | Detailed description of the task. May be absent |

1214| `teammate_name` | Name of the teammate completing the task. May be absent |

1215| `team_name` | Name of the team. May be absent |

661 1216

662* For `PreToolUse`, this is different from `"permissionDecision": "deny"`, which1217#### TaskCompleted decision control

663 only blocks a specific tool call and provides automatic feedback to Claude.

664* For `PostToolUse`, this is different from `"decision": "block"`, which

665 provides automated feedback to Claude.

666* For `UserPromptSubmit`, this prevents the prompt from being processed.

667* For `Stop` and `SubagentStop`, this takes precedence over any

668 `"decision": "block"` output.

669* In all cases, `"continue" = false` takes precedence over any

670 `"decision": "block"` output.

671 1218

672`stopReason` accompanies `continue` with a reason shown to the user, not shown1219TaskCompleted hooks use exit codes only, not JSON decision control. This example runs tests and blocks task completion if they fail:

673to Claude.

674 1220

675#### `PreToolUse` Decision Control1221```bash theme={null}

~~676~~ 1222#!/bin/bash

677`PreToolUse` hooks can control whether a tool call proceeds.1223INPUT=$(cat)

1224TASK_SUBJECT=$(echo "$INPUT" | jq -r '.task_subject')

678 1225

679* `"allow"` bypasses the permission system. `permissionDecisionReason` is shown1226# Run the test suite

680 to the user but not to Claude.1227if ! npm test 2>&1; then

681* `"deny"` prevents the tool call from executing. `permissionDecisionReason` is1228 echo "Tests not passing. Fix failing tests before completing: $TASK_SUBJECT" >&2

682 shown to Claude.1229 exit 2

683* `"ask"` asks the user to confirm the tool call in the UI.1230fi

684 `permissionDecisionReason` is shown to the user but not to Claude.

685 1231

686Additionally, hooks can modify tool inputs before execution using `updatedInput`:1232exit 0

1233```

687 1234

688* `updatedInput` allows you to modify the tool's input parameters before the tool executes.1235### PreCompact

689* This is most useful with `"permissionDecision": "allow"` to modify and approve tool calls.

690 1236

691```json theme={null}1237Runs before Claude Code is about to run a compact operation.

692{

693 "hookSpecificOutput": {

694 "hookEventName": "PreToolUse",

695 "permissionDecision": "allow"

696 "permissionDecisionReason": "My reason here",

697 "updatedInput": {

698 "field_to_modify": "new value"

699 }

700 }

701}

702```

703 1238

704<Note>1239The matcher value indicates whether compaction was triggered manually or automatically:

705 The `decision` and `reason` fields are deprecated for PreToolUse hooks.

706 Use `hookSpecificOutput.permissionDecision` and

707 `hookSpecificOutput.permissionDecisionReason` instead. The deprecated fields

708 `"approve"` and `"block"` map to `"allow"` and `"deny"` respectively.

709</Note>

710 1240

711#### `PermissionRequest` Decision Control1241| Matcher | When it fires |

1242| :------- | :------------------------------------------- |

1243| `manual` | `/compact` |

1244| `auto` | Auto-compact when the context window is full |

712 1245

713`PermissionRequest` hooks can allow or deny permission requests shown to the user.1246#### PreCompact input

714 1247

715* For `"behavior": "allow"` you can also optionally pass in an `"updatedInput"` that modifies the tool's input parameters before the tool executes.1248In addition to the [common input fields](#common-input-fields), PreCompact hooks receive `trigger` and `custom_instructions`. For `manual`, `custom_instructions` contains what the user passes into `/compact`. For `auto`, `custom_instructions` is empty.

716* For `"behavior": "deny"` you can also optionally pass in a `"message"` string that tells the model why the permission was denied, and a boolean `"interrupt"` which will stop Claude.

717 1249

718```json theme={null}1250```json theme={null}

719{1251{

720 "hookSpecificOutput": {1252 "session_id": "abc123",

721 "hookEventName": "PermissionRequest",1253 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

722 "decision": {1254 "cwd": "/Users/...",

723 "behavior": "allow",1255 "permission_mode": "default",

724 "updatedInput": {1256 "hook_event_name": "PreCompact",

725 "command": "npm run lint"1257 "trigger": "manual",

726 }1258 "custom_instructions": ""

727 }

728 }

729}1259}

730```1260```

731 1261

732#### `PostToolUse` Decision Control1262### SessionEnd

1263

1264Runs when a Claude Code session ends. Useful for cleanup tasks, logging session

1265statistics, or saving session state. Supports matchers to filter by exit reason.

1266

1267The `reason` field in the hook input indicates why the session ended:

1268

1269| Reason | Description |

1270| :---------------------------- | :----------------------------------------- |

1271| `clear` | Session cleared with `/clear` command |

1272| `logout` | User logged out |

1273| `prompt_input_exit` | User exited while prompt input was visible |

1274| `bypass_permissions_disabled` | Bypass permissions mode was disabled |

1275| `other` | Other exit reasons |

733 1276

734`PostToolUse` hooks can provide feedback to Claude after tool execution.1277#### SessionEnd input

735 1278

736* `"block"` automatically prompts Claude with `reason`.1279In addition to the [common input fields](#common-input-fields), SessionEnd hooks receive a `reason` field indicating why the session ended. See the [reason table](#sessionend) above for all values.

737* `undefined` does nothing. `reason` is ignored.

738* `"hookSpecificOutput.additionalContext"` adds context for Claude to consider.

739 1280

740```json theme={null}1281```json theme={null}

741{1282{

742 "decision": "block" | undefined,1283 "session_id": "abc123",

743 "reason": "Explanation for decision",1284 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",

744 "hookSpecificOutput": {1285 "cwd": "/Users/...",

745 "hookEventName": "PostToolUse",1286 "permission_mode": "default",

746 "additionalContext": "Additional information for Claude"1287 "hook_event_name": "SessionEnd",

747 }1288 "reason": "other"

748}1289}

749```1290```

750 1291

751#### `UserPromptSubmit` Decision Control1292SessionEnd hooks have no decision control. They cannot block session termination but can perform cleanup tasks.

752 1293

753`UserPromptSubmit` hooks can control whether a user prompt is processed and add context.1294## Prompt-based hooks

754 1295

755**Adding context (exit code 0):**1296In addition to Bash command hooks (`type: "command"`), Claude Code supports prompt-based hooks (`type: "prompt"`) that use an LLM to evaluate whether to allow or block an action. Prompt-based hooks work with the following events: `PreToolUse`, `PostToolUse`, `PostToolUseFailure`, `PermissionRequest`, `UserPromptSubmit`, `Stop`, `SubagentStop`, and `TaskCompleted`. `TeammateIdle` does not support prompt-based or agent-based hooks.

756There are two ways to add context to the conversation:1297

1298### How prompt-based hooks work

757 1299

7581. **Plain text stdout** (simpler): Any non-JSON text written to stdout is added1300Instead of executing a Bash command, prompt-based hooks:

759 as context. This is the easiest way to inject information.

760 1301

7612. **JSON with `additionalContext`** (structured): Use the JSON format below for13021. Send the hook input and your prompt to a Claude model, Haiku by default

762 more control. The `additionalContext` field is added as context.13032. The LLM responds with structured JSON containing a decision

13043. Claude Code processes the decision automatically

763 1305

764Both methods work with exit code 0. Plain stdout is shown as hook output in1306### Prompt hook configuration

765the transcript; `additionalContext` is added more discretely.

766 1307

767**Blocking prompts:**1308Set `type` to `"prompt"` and provide a `prompt` string instead of a `command`. Use the `$ARGUMENTS` placeholder to inject the hook's JSON input data into your prompt text. Claude Code sends the combined prompt and input to a fast Claude model, which returns a JSON decision.

768 1309

769* `"decision": "block"` prevents the prompt from being processed. The submitted1310This `Stop` hook asks the LLM to evaluate whether all tasks are complete before allowing Claude to finish:

770 prompt is erased from context. `"reason"` is shown to the user but not added

771 to context.

772* `"decision": undefined` (or omitted) allows the prompt to proceed normally.

773 1311

774```json theme={null}1312```json theme={null}

775{1313{

776 "decision": "block" | undefined,1314 "hooks": {

777 "reason": "Explanation for decision",1315 "Stop": [

778 "hookSpecificOutput": {1316 {

779 "hookEventName": "UserPromptSubmit",1317 "hooks": [

780 "additionalContext": "My additional context here"1318 {

1319 "type": "prompt",

1320 "prompt": "Evaluate if Claude should stop: $ARGUMENTS. Check if all tasks are complete."

1321 }

1322 ]

1323 }

1324 ]

781 }1325 }

782}1326}

783```1327```

784 1328

785<Note>1329| Field | Required | Description |

786 The JSON format isn't required for simple use cases. To add context, you can print plain text to stdout with exit code 0. Use JSON when you need to1330| :-------- | :------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------ |

787 block prompts or want more structured control.1331| `type` | yes | Must be `"prompt"` |

788</Note>1332| `prompt` | yes | The prompt text to send to the LLM. Use `$ARGUMENTS` as a placeholder for the hook input JSON. If `$ARGUMENTS` is not present, input JSON is appended to the prompt |

~~789~~ 1333| `model` | no | Model to use for evaluation. Defaults to a fast model |

790#### `Stop`/`SubagentStop` Decision Control1334| `timeout` | no | Timeout in seconds. Default: 30 |

791 1335

792`Stop` and `SubagentStop` hooks can control whether Claude must continue.1336### Response schema

793 1337

794* `"block"` prevents Claude from stopping. You must populate `reason` for Claude1338The LLM must respond with JSON containing:

795 to know how to proceed.

796* `undefined` allows Claude to stop. `reason` is ignored.

797 1339

798```json theme={null}1340```json theme={null}

799{1341{

800 "decision": "block" | undefined,1342 "ok": true | false,

801 "reason": "Must be provided when Claude is blocked from stopping"1343 "reason": "Explanation for the decision"

802}1344}

803```1345```

804 1346

805#### `SessionStart` Decision Control1347| Field | Description |

1348| :------- | :--------------------------------------------------------- |

1349| `ok` | `true` allows the action, `false` prevents it |

1350| `reason` | Required when `ok` is `false`. Explanation shown to Claude |

806 1351

807`SessionStart` hooks allow you to load in context at the start of a session.1352### Example: Multi-criteria Stop hook

808 1353

809* `"hookSpecificOutput.additionalContext"` adds the string to the context.1354This `Stop` hook uses a detailed prompt to check three conditions before allowing Claude to stop. If `"ok"` is `false`, Claude continues working with the provided reason as its next instruction. `SubagentStop` hooks use the same format to evaluate whether a [subagent](/en/sub-agents) should stop:

810* Multiple hooks' `additionalContext` values are concatenated.

811 1355

812```json theme={null}1356```json theme={null}

813{1357{

814 "hookSpecificOutput": {1358 "hooks": {

815 "hookEventName": "SessionStart",1359 "Stop": [

816 "additionalContext": "My additional context here"1360 {

1361 "hooks": [

1362 {

1363 "type": "prompt",

1364 "prompt": "You are evaluating whether Claude should stop working. Context: $ARGUMENTS\n\nAnalyze the conversation and determine if:\n1. All user-requested tasks are complete\n2. Any errors need to be addressed\n3. Follow-up work is needed\n\nRespond with JSON: {\"ok\": true} to allow stopping, or {\"ok\": false, \"reason\": \"your explanation\"} to continue working.",

1365 "timeout": 30

1366 }

1367 ]

1368 }

1369 ]

817 }1370 }

818}1371}

819```1372```

820 1373

821#### `SessionEnd` Decision Control1374## Agent-based hooks

~~822~~

823`SessionEnd` hooks run when a session ends. They cannot block session termination

824but can perform cleanup tasks.

~~825~~

826#### Exit Code Example: Bash Command Validation

~~827~~

828```python theme={null}

829#!/usr/bin/env python3

830import json

831import re

832import sys

~~833~~

834# Define validation rules as a list of (regex pattern, message) tuples

835VALIDATION_RULES = [

836 (

837 r"\bgrep\b(?!.*\|)",

838 "Use 'rg' (ripgrep) instead of 'grep' for better performance and features",

839 ),

840 (

841 r"\bfind\s+\S+\s+-name\b",

842 "Use 'rg --files | rg pattern' or 'rg --files -g pattern' instead of 'find -name' for better performance",

843 ),

844]

~~845~~

~~846~~

847def validate_command(command: str) -> list[str]:

848 issues = []

849 for pattern, message in VALIDATION_RULES:

850 if re.search(pattern, command):

851 issues.append(message)

852 return issues

~~853~~

~~854~~

855try:

856 input_data = json.load(sys.stdin)

857except json.JSONDecodeError as e:

858 print(f"Error: Invalid JSON input: {e}", file=sys.stderr)

859 sys.exit(1)

~~860~~

861tool_name = input_data.get("tool_name", "")

862tool_input = input_data.get("tool_input", {})

863command = tool_input.get("command", "")

~~864~~

865if tool_name != "Bash" or not command:

866 sys.exit(1)

~~867~~

868# Validate the command

869issues = validate_command(command)

~~870~~

871if issues:

872 for message in issues:

873 print(f"• {message}", file=sys.stderr)

874 # Exit code 2 blocks tool call and shows stderr to Claude

875 sys.exit(2)

876```

~~877~~

878#### JSON Output Example: UserPromptSubmit to Add Context and Validation

879 1375

880<Note>1376Agent-based hooks (`type: "agent"`) are like prompt-based hooks but with multi-turn tool access. Instead of a single LLM call, an agent hook spawns a subagent that can read files, search code, and inspect the codebase to verify conditions. Agent hooks support the same events as prompt-based hooks.

881 For `UserPromptSubmit` hooks, you can inject context using either method:

882 1377

883 * **Plain text stdout** with exit code 0: Simplest approach, prints text1378### How agent hooks work

884 * **JSON output** with exit code 0: Use `"decision": "block"` to reject prompts,

885 or `additionalContext` for structured context injection

886 1379

887 Remember: Exit code 2 only uses `stderr` for the error message. To block using1380When an agent hook fires:

888 JSON (with a custom reason), use `"decision": "block"` with exit code 0.

889</Note>

890 1381

891```python theme={null}13821. Claude Code spawns a subagent with your prompt and the hook's JSON input

892#!/usr/bin/env python313832. The subagent can use tools like Read, Grep, and Glob to investigate

893import json13843. After up to 50 turns, the subagent returns a structured `{ "ok": true/false }` decision

894import sys13854. Claude Code processes the decision the same way as a prompt hook

895import re

896import datetime

~~897~~

898# Load input from stdin

899try:

900 input_data = json.load(sys.stdin)

901except json.JSONDecodeError as e:

902 print(f"Error: Invalid JSON input: {e}", file=sys.stderr)

903 sys.exit(1)

~~904~~

905prompt = input_data.get("prompt", "")

~~906~~

907# Check for sensitive patterns

908sensitive_patterns = [

909 (r"(?i)\b(password|secret|key|token)\s*[:=]", "Prompt contains potential secrets"),

910]

~~911~~

912for pattern, message in sensitive_patterns:

913 if re.search(pattern, prompt):

914 # Use JSON output to block with a specific reason

915 output = {

916 "decision": "block",

917 "reason": f"Security policy violation: {message}. Please rephrase your request without sensitive information."

918 }

919 print(json.dumps(output))

920 sys.exit(0)

921 1386

922# Add current time to context1387Agent hooks are useful when verification requires inspecting actual files or test output, not just evaluating the hook input data alone.

923context = f"Current time: {datetime.datetime.now()}"

924print(context)

925 1388

926"""1389### Agent hook configuration

927The following is also equivalent:

928print(json.dumps({

929 "hookSpecificOutput": {

930 "hookEventName": "UserPromptSubmit",

931 "additionalContext": context,

932 },

933}))

934"""

935 1390

936# Allow the prompt to proceed with the additional context1391Set `type` to `"agent"` and provide a `prompt` string. The configuration fields are the same as [prompt hooks](#prompt-hook-configuration), with a longer default timeout:

937sys.exit(0)

938```

939 1392

940#### JSON Output Example: PreToolUse with Approval1393| Field | Required | Description |

~~941~~ 1394| :-------- | :------- | :------------------------------------------------------------------------------------------ |

942```python theme={null}1395| `type` | yes | Must be `"agent"` |

943#!/usr/bin/env python31396| `prompt` | yes | Prompt describing what to verify. Use `$ARGUMENTS` as a placeholder for the hook input JSON |

944import json1397| `model` | no | Model to use. Defaults to a fast model |

945import sys1398| `timeout` | no | Timeout in seconds. Default: 60 |

~~946~~

947# Load input from stdin

948try:

949 input_data = json.load(sys.stdin)

950except json.JSONDecodeError as e:

951 print(f"Error: Invalid JSON input: {e}", file=sys.stderr)

952 sys.exit(1)

~~953~~

954tool_name = input_data.get("tool_name", "")

955tool_input = input_data.get("tool_input", {})

~~956~~

957# Example: Auto-approve file reads for documentation files

958if tool_name == "Read":

959 file_path = tool_input.get("file_path", "")

960 if file_path.endswith((".md", ".mdx", ".txt", ".json")):

961 # Use JSON output to auto-approve the tool call

962 output = {

963 "decision": "approve",

964 "reason": "Documentation file auto-approved",

965 "suppressOutput": True # Don't show in verbose mode

966 }

967 print(json.dumps(output))

968 sys.exit(0)

969 1399

970# For other cases, let the normal permission flow proceed1400The response schema is the same as prompt hooks: `{ "ok": true }` to allow or `{ "ok": false, "reason": "..." }` to block.

971sys.exit(0)

972```

973 1401

974## Working with MCP Tools1402This `Stop` hook verifies that all unit tests pass before allowing Claude to finish:

975 1403

976Claude Code hooks work seamlessly with1404```json theme={null}

977[Model Context Protocol (MCP) tools](/en/mcp). When MCP servers1405{

978provide tools, they appear with a special naming pattern that you can match in1406 "hooks": {

979your hooks.1407 "Stop": [

1408 {

1409 "hooks": [

1410 {

1411 "type": "agent",

1412 "prompt": "Verify that all unit tests pass. Run the test suite and check the results. $ARGUMENTS",

1413 "timeout": 120

1414 }

1415 ]

1416 }

1417 ]

1418 }

1419}

1420```

980 1421

981### MCP Tool Naming1422## Run hooks in the background

982 1423

983MCP tools follow the pattern `mcp__<server>__<tool>`, for example:1424By default, hooks block Claude's execution until they complete. For long-running tasks like deployments, test suites, or external API calls, set `"async": true` to run the hook in the background while Claude continues working. Async hooks cannot block or control Claude's behavior: response fields like `decision`, `permissionDecision`, and `continue` have no effect, because the action they would have controlled has already completed.

984 1425

985* `mcp__memory__create_entities` - Memory server's create entities tool1426### Configure an async hook

986* `mcp__filesystem__read_file` - Filesystem server's read file tool

987* `mcp__github__search_repositories` - GitHub server's search tool

988 1427

989### Configuring Hooks for MCP Tools1428Add `"async": true` to a command hook's configuration to run it in the background without blocking Claude. This field is only available on `type: "command"` hooks.

990 1429

991You can target specific MCP tools or entire MCP servers:1430This hook runs a test script after every `Write` tool call. Claude continues working immediately while `run-tests.sh` executes for up to 120 seconds. When the script finishes, its output is delivered on the next conversation turn:

992 1431

993```json theme={null}1432```json theme={null}

994{1433{

995 "hooks": {1434 "hooks": {

996 "PreToolUse": [1435 "PostToolUse": [

997 {

998 "matcher": "mcp__memory__.*",

999 "hooks": [

1000 {

1001 "type": "command",

1002 "command": "echo 'Memory operation initiated' >> ~/mcp-operations.log"

1003 }

1004 ]

1005 },

1006 {1436 {

1007 "matcher": "mcp__.*__write.*",1437 "matcher": "Write",

1008 "hooks": [1438 "hooks": [

1009 {1439 {

1010 "type": "command",1440 "type": "command",

1011 "command": "/home/user/scripts/validate-mcp-write.py"1441 "command": "/path/to/run-tests.sh",

1442 "async": true,

1443 "timeout": 120

1012 }1444 }

1013 ]1445 ]

1014 }1446 }

1017}1449}

1018```1450```

1019 1451

1020## Examples1452The `timeout` field sets the maximum time in seconds for the background process. If not specified, async hooks use the same 10-minute default as sync hooks.

~~1021~~

1022<Tip>

1023 For practical examples including code formatting, notifications, and file protection, see [More Examples](/en/hooks-guide#more-examples) in the get started guide.

1024</Tip>

~~1025~~

1026## Security Considerations

1027 1453

1028### Disclaimer1454### How async hooks execute

1029 1455

1030**USE AT YOUR OWN RISK**: Claude Code hooks execute arbitrary shell commands on1456When an async hook fires, Claude Code starts the hook process and immediately continues without waiting for it to finish. The hook receives the same JSON input via stdin as a synchronous hook.

1031your system automatically. By using hooks, you acknowledge that:

1032 1457

1033* You are solely responsible for the commands you configure1458After the background process exits, if the hook produced a JSON response with a `systemMessage` or `additionalContext` field, that content is delivered to Claude as context on the next conversation turn.

1034* Hooks can modify, delete, or access any files your user account can access

1035* Malicious or poorly written hooks can cause data loss or system damage

1036* Anthropic provides no warranty and assumes no liability for any damages

1037 resulting from hook usage

1038* You should thoroughly test hooks in a safe environment before production use

1039 1459

1040Always review and understand any hook commands before adding them to your1460### Example: run tests after file changes

1041configuration.

1042 1461

1043### Security Best Practices1462This hook starts a test suite in the background whenever Claude writes a file, then reports the results back to Claude when the tests finish. Save this script to `.claude/hooks/run-tests-async.sh` in your project and make it executable with `chmod +x`:

1044 1463

1045Here are some key practices for writing more secure hooks:1464```bash theme={null}

1465#!/bin/bash

1466# run-tests-async.sh

1046 1467

10471. **Validate and sanitize inputs** - Never trust input data blindly1468# Read hook input from stdin

10482. **Always quote shell variables** - Use `"$VAR"` not `$VAR`1469INPUT=$(cat)

10493. **Block path traversal** - Check for `..` in file paths1470FILE_PATH=$(echo "$INPUT" | jq -r '.tool_input.file_path // empty')

10504. **Use absolute paths** - Specify full paths for scripts (use

1051 "\$CLAUDE\_PROJECT\_DIR" for the project path)

10525. **Skip sensitive files** - Avoid `.env`, `.git/`, keys, etc.

1053 1471

1054### Configuration Safety1472# Only run tests for source files

1473if [[ "$FILE_PATH" != *.ts && "$FILE_PATH" != *.js ]]; then

1474 exit 0

1475fi

1055 1476

1056Direct edits to hooks in settings files don't take effect immediately. Claude1477# Run tests and report results via systemMessage

1057Code:1478RESULT=$(npm test 2>&1)

1479EXIT_CODE=$?

1058 1480

10591. Captures a snapshot of hooks at startup1481if [ $EXIT_CODE -eq 0 ]; then

10602. Uses this snapshot throughout the session1482 echo "{\"systemMessage\": \"Tests passed after editing $FILE_PATH\"}"

10613. Warns if hooks are modified externally1483else

10624. Requires review in `/hooks` menu for changes to apply1484 echo "{\"systemMessage\": \"Tests failed after editing $FILE_PATH: $RESULT\"}"

1485fi

1486```

1063 1487

1064This prevents malicious hook modifications from affecting your current session.1488Then add this configuration to `.claude/settings.json` in your project root. The `async: true` flag lets Claude keep working while tests run:

1065 1489

1066## Hook Execution Details1490```json theme={null}

1491{

1492 "hooks": {

1493 "PostToolUse": [

1494 {

1495 "matcher": "Write|Edit",

1496 "hooks": [

1497 {

1498 "type": "command",

1499 "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/run-tests-async.sh",

1500 "async": true,

1501 "timeout": 300

1502 }

1503 ]

1504 }

1505 ]

1506 }

1507}

1508```

1067 1509

1068* **Timeout**: 60-second execution limit by default, configurable per command.1510### Limitations

1069 * A timeout for an individual command does not affect the other commands.

1070* **Parallelization**: All matching hooks run in parallel

1071* **Deduplication**: Multiple identical hook commands are deduplicated automatically

1072* **Environment**: Runs in current directory with Claude Code's environment

1073 * The `CLAUDE_PROJECT_DIR` environment variable is available and contains the

1074 absolute path to the project root directory (where Claude Code was started)

1075 * The `CLAUDE_CODE_REMOTE` environment variable indicates whether the hook is running in a remote (web) environment (`"true"`) or local CLI environment (not set or empty). Use this to run different logic based on execution context.

1076* **Input**: JSON via stdin

1077* **Output**:

1078 * PreToolUse/PermissionRequest/PostToolUse/Stop/SubagentStop: Progress shown in verbose mode (ctrl+o)

1079 * Notification/SessionEnd: Logged to debug only (`--debug`)

1080 * UserPromptSubmit/SessionStart: stdout added as context for Claude

1081 1511

1082## Debugging1512Async hooks have several constraints compared to synchronous hooks:

1083 1513

1084### Basic Troubleshooting1514* Only `type: "command"` hooks support `async`. Prompt-based hooks cannot run asynchronously.

1515* Async hooks cannot block tool calls or return decisions. By the time the hook completes, the triggering action has already proceeded.

1516* Hook output is delivered on the next conversation turn. If the session is idle, the response waits until the next user interaction.

1517* Each execution creates a separate background process. There is no deduplication across multiple firings of the same async hook.

1085 1518

1086If your hooks aren't working:1519## Security considerations

1087 1520

10881. **Check configuration** - Run `/hooks` to see if your hook is registered1521### Disclaimer

10892. **Verify syntax** - Ensure your JSON settings are valid

10903. **Test commands** - Run hook commands manually first

10914. **Check permissions** - Make sure scripts are executable

10925. **Review logs** - Use `claude --debug` to see hook execution details

1093 1522

1094Common issues:1523Hooks run with your system user's full permissions.

1095 1524

1096* **Quotes not escaped** - Use `\"` inside JSON strings1525<Warning>

1097* **Wrong matcher** - Check tool names match exactly (case-sensitive)1526 Hooks execute shell commands with your full user permissions. They can modify, delete, or access any files your user account can access. Review and test all hook commands before adding them to your configuration.

1098* **Command not found** - Use full paths for scripts1527</Warning>

1099 1528

1100### Advanced Debugging1529### Security best practices

1101 1530

1102For complex hook issues:1531Keep these practices in mind when writing hooks:

1103 1532

11041. **Inspect hook execution** - Use `claude --debug` to see detailed hook1533* **Validate and sanitize inputs**: never trust input data blindly

1105 execution1534* **Always quote shell variables**: use `"$VAR"` not `$VAR`

11062. **Validate JSON schemas** - Test hook input/output with external tools1535* **Block path traversal**: check for `..` in file paths

11073. **Check environment variables** - Verify Claude Code's environment is correct1536* **Use absolute paths**: specify full paths for scripts, using `"$CLAUDE_PROJECT_DIR"` for the project root

11084. **Test edge cases** - Try hooks with unusual file paths or inputs1537* **Skip sensitive files**: avoid `.env`, `.git/`, keys, etc.

11095. **Monitor system resources** - Check for resource exhaustion during hook

1110 execution

11116. **Use structured logging** - Implement logging in your hook scripts

1112 1538

1113### Debug Output Example1539## Debug hooks

1114 1540

1115Use `claude --debug` to see hook execution details:1541Run `claude --debug` to see hook execution details, including which hooks matched, their exit codes, and output. Toggle verbose mode with `Ctrl+O` to see hook progress in the transcript.

1116 1542

1117```1543```

1118[DEBUG] Executing hooks for PostToolUse:Write1544[DEBUG] Executing hooks for PostToolUse:Write

1120[DEBUG] Found 1 hook matchers in settings1546[DEBUG] Found 1 hook matchers in settings

1121[DEBUG] Matched 1 hooks for query "Write"1547[DEBUG] Matched 1 hooks for query "Write"

1122[DEBUG] Found 1 hook commands to execute1548[DEBUG] Found 1 hook commands to execute

1123[DEBUG] Executing hook command: <Your command> with timeout 60000ms1549[DEBUG] Executing hook command: <Your command> with timeout 600000ms

1124[DEBUG] Hook command completed with status 0: <Your stdout>1550[DEBUG] Hook command completed with status 0: <Your stdout>

1125```1551```

1126 1552

1127Progress messages appear in verbose mode (ctrl+o) showing:1553For troubleshooting common issues like hooks not firing, infinite Stop hook loops, or configuration errors, see [Limitations and troubleshooting](/en/hooks-guide#limitations-and-troubleshooting) in the guide.

~~1128~~

1129* Which hook is running

1130* Command being executed

1131* Success/failure status

1132* Output or error messages

~~1133~~

~~1134~~

~~1135~~

1136> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

hooks-guide.md +502 −206

Details

~~1# Get started with Claude Code hooks~~1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

2 4

~~3> Learn how to customize and extend Claude Code's behavior by registering shell commands~~5# Automate workflows with hooks

4 6

~~5Claude Code hooks are user-defined shell commands that execute at various points~~7> Run shell commands automatically when Claude Code edits files, finishes tasks, or needs input. Format code, send notifications, validate commands, and enforce project rules.

~~6in Claude Code's lifecycle. Hooks provide deterministic control over Claude~~8

~~7Code's behavior, ensuring certain actions always happen rather than relying on~~9Hooks are user-defined shell commands that execute at specific points in Claude Code's lifecycle. They provide deterministic control over Claude Code's behavior, ensuring certain actions always happen rather than relying on the LLM to choose to run them. Use hooks to enforce project rules, automate repetitive tasks, and integrate Claude Code with your existing tools.

~~8the LLM to choose to run them.~~10

11For decisions that require judgment rather than deterministic rules, you can also use [prompt-based hooks](#prompt-based-hooks) or [agent-based hooks](#agent-based-hooks) that use a Claude model to evaluate conditions.

13For other ways to extend Claude Code, see [skills](/en/skills) for giving Claude additional instructions and executable commands, [subagents](/en/sub-agents) for running tasks in isolated contexts, and [plugins](/en/plugins) for packaging extensions to share across projects.

9 14

10<Tip>15<Tip>

~~11 For reference documentation on hooks, see [Hooks reference](/en/hooks).~~16 This guide covers common use cases and how to get started. For full event schemas, JSON input/output formats, and advanced features like async hooks and MCP tool hooks, see the [Hooks reference](/en/hooks).

12</Tip>17</Tip>

13 18

~~14Example use cases for hooks include:~~19## Set up your first hook

15 20

~~16* **Notifications**: Customize how you get notified when Claude Code is awaiting~~21The fastest way to create a hook is through the `/hooks` interactive menu in Claude Code. This walkthrough creates a desktop notification hook, so you get alerted whenever Claude is waiting for your input instead of watching the terminal.

~~17 your input or permission to run something.~~

~~18* **Automatic formatting**: Run `prettier` on .ts files, `gofmt` on .go files,~~

~~19 etc. after every file edit.~~

~~20* **Logging**: Track and count all executed commands for compliance or~~

~~21 debugging.~~

~~22* **Feedback**: Provide automated feedback when Claude Code produces code that~~

~~23 does not follow your codebase conventions.~~

~~24* **Custom permissions**: Block modifications to production files or sensitive~~

~~25 directories.~~

26 22

~~27By encoding these rules as hooks rather than prompting instructions, you turn~~23<Steps>

~~28suggestions into app-level code that executes every time it is expected to run.~~24 <Step title="Open the hooks menu">

25 Type `/hooks` in the Claude Code CLI. You'll see a list of all available hook events, plus an option to disable all hooks. Each event corresponds to a point in Claude's lifecycle where you can run custom code. Select `Notification` to create a hook that fires when Claude needs your attention.

26 </Step>

29 27

~~30<Warning>~~28 <Step title="Configure the matcher">

~~31 You must consider the security implication of hooks as you add them, because hooks run automatically during the agent loop with your current environment's credentials.~~29 The menu shows a list of matchers, which filter when the hook fires. Set the matcher to `*` to fire on all notification types. You can narrow it later by changing the matcher to a specific value like `permission_prompt` or `idle_prompt`.

~~32 For example, malicious hooks code can exfiltrate your data. Always review your hooks implementation before registering them.~~30 </Step>

33 31

~~34 For full security best practices, see [Security Considerations](/en/hooks#security-considerations) in the hooks reference documentation.~~32 <Step title="Add your command">

~~35</Warning>~~33 Select `+ Add new hook…`. The menu prompts you for a shell command to run when the event fires. Hooks run any shell command you provide, so you can use your platform's built-in notification tool. Copy the command for your OS:

36 34

~~37## Hook Events Overview~~35 <Tabs>

36 <Tab title="macOS">

37 Uses [`osascript`](https://ss64.com/mac/osascript.html) to trigger a native macOS notification through AppleScript:

38 38

~~39Claude Code provides several hook events that run at different points in the~~39 ```

~~40workflow:~~40 osascript -e 'display notification "Claude Code needs your attention" with title "Claude Code"'

41 ```

42 </Tab>

41 43

~~42* **PreToolUse**: Runs before tool calls (can block them)~~44 <Tab title="Linux">

~~43* **PermissionRequest**: Runs when a permission dialog is shown (can allow or deny)~~45 Uses `notify-send`, which is pre-installed on most Linux desktops with a notification daemon:

~~44* **PostToolUse**: Runs after tool calls complete~~

~~45* **UserPromptSubmit**: Runs when the user submits a prompt, before Claude processes it~~

~~46* **Notification**: Runs when Claude Code sends notifications~~

~~47* **Stop**: Runs when Claude Code finishes responding~~

~~48* **SubagentStop**: Runs when subagent tasks complete~~

~~49* **PreCompact**: Runs before Claude Code is about to run a compact operation~~

~~50* **SessionStart**: Runs when Claude Code starts a new session or resumes an existing session~~

~~51* **SessionEnd**: Runs when Claude Code session ends~~

52 46

~~53Each event receives different data and can control Claude's behavior in~~47 ```

~~54different ways.~~48 notify-send 'Claude Code' 'Claude Code needs your attention'

49 ```

50 </Tab>

55 51

~~56## Quickstart~~52 <Tab title="Windows (PowerShell)">

53 Uses PowerShell to show a native message box through .NET's Windows Forms:

57 54

~~58In this quickstart, you'll add a hook that logs the shell commands that Claude~~55 ```

~~59Code runs.~~56 powershell.exe -Command "[System.Reflection.Assembly]::LoadWithPartialName('System.Windows.Forms'); [System.Windows.Forms.MessageBox]::Show('Claude Code needs your attention', 'Claude Code')"

57 ```

58 </Tab>

59 </Tabs>

60 </Step>

60 61

~~61### Prerequisites~~62 <Step title="Choose a storage location">

63 The menu asks where to save the hook configuration. Select `User settings` to store it in `~/.claude/settings.json`, which applies the hook to all your projects. You could also choose `Project settings` to scope it to the current project. See [Configure hook location](#configure-hook-location) for all available scopes.

64 </Step>

62 65

~~63Install `jq` for JSON processing in the command line.~~66 <Step title="Test the hook">

67 Press `Esc` to return to the CLI. Ask Claude to do something that requires permission, then switch away from the terminal. You should receive a desktop notification.

68 </Step>

69</Steps>

64 70

~~65### Step 1: Open hooks configuration~~71## What you can automate

66 72

~~67Run the `/hooks` [slash command](/en/slash-commands) and select~~73Hooks let you run code at key points in Claude Code's lifecycle: format files after edits, block commands before they execute, send notifications when Claude needs input, inject context at session start, and more. For the full list of hook events, see the [Hooks reference](/en/hooks#hook-lifecycle).

~~68the `PreToolUse` hook event.~~

69 74

~~70`PreToolUse` hooks run before tool calls and can block them while providing~~75Each example includes a ready-to-use configuration block that you add to a [settings file](#configure-hook-location). The most common patterns:

~~71Claude feedback on what to do differently.~~

72 76

~~73### Step 2: Add a matcher~~77* [Get notified when Claude needs input](#get-notified-when-claude-needs-input)

78* [Auto-format code after edits](#auto-format-code-after-edits)

79* [Block edits to protected files](#block-edits-to-protected-files)

80* [Re-inject context after compaction](#re-inject-context-after-compaction)

74 81

~~75Select `+ Add new matcher…` to run your hook only on Bash tool calls.~~82### Get notified when Claude needs input

76 83

~~77Type `Bash` for the matcher.~~84Get a desktop notification whenever Claude finishes working and needs your input, so you can switch to other tasks without checking the terminal.

78 85

~~79<Note>You can use `*` to match all tools.</Note>~~86This hook uses the `Notification` event, which fires when Claude is waiting for input or permission. Each tab below uses the platform's native notification command. Add this to `~/.claude/settings.json`, or use the [interactive walkthrough](#set-up-your-first-hook) above to configure it with `/hooks`:

80 87

~~81### Step 3: Add the hook~~88<Tabs>

89 <Tab title="macOS">

90 ```json theme={null}

91 {

92 "hooks": {

93 "Notification": [

94 {

95 "matcher": "",

96 "hooks": [

97 {

98 "type": "command",

99 "command": "osascript -e 'display notification \"Claude Code needs your attention\" with title \"Claude Code\"'"

100 }

101 ]

102 }

103 ]

104 }

105 }

106 ```

107 </Tab>

82 108

~~83Select `+ Add new hook…` and enter this command:~~109 <Tab title="Linux">

110 ```json theme={null}

111 {

112 "hooks": {

113 "Notification": [

114 {

115 "matcher": "",

116 "hooks": [

117 {

118 "type": "command",

119 "command": "notify-send 'Claude Code' 'Claude Code needs your attention'"

120 }

121 ]

122 }

123 ]

124 }

125 }

126 ```

127 </Tab>

84 128

~~85```bash theme={null}~~129 <Tab title="Windows (PowerShell)">

~~86jq -r '"\(.tool_input.command) - \(.tool_input.description // "No description")"' >> ~/.claude/bash-command-log.txt~~130 ```json theme={null}

131 {

132 "hooks": {

133 "Notification": [

134 {

135 "matcher": "",

136 "hooks": [

137 {

138 "type": "command",

139 "command": "powershell.exe -Command \"[System.Reflection.Assembly]::LoadWithPartialName('System.Windows.Forms'); [System.Windows.Forms.MessageBox]::Show('Claude Code needs your attention', 'Claude Code')\""

140 }

141 ]

142 }

143 ]

144 }

145 }

146 ```

147 </Tab>

148</Tabs>

149

150### Auto-format code after edits

151

152Automatically run [Prettier](https://prettier.io/) on every file Claude edits, so formatting stays consistent without manual intervention.

153

154This hook uses the `PostToolUse` event with an `Edit|Write` matcher, so it runs only after file-editing tools. The command extracts the edited file path with [`jq`](https://jqlang.github.io/jq/) and passes it to Prettier. Add this to `.claude/settings.json` in your project root:

155

156```json theme={null}

157{

158 "hooks": {

159 "PostToolUse": [

160 {

161 "matcher": "Edit|Write",

162 "hooks": [

163 {

164 "type": "command",

165 "command": "jq -r '.tool_input.file_path' | xargs npx prettier --write"

166 }

167 ]

168 }

169 ]

170 }

171}

87```172```

88 173

~~89### Step 4: Save your configuration~~174<Note>

175 The Bash examples on this page use `jq` for JSON parsing. Install it with `brew install jq` (macOS), `apt-get install jq` (Debian/Ubuntu), or see [`jq` downloads](https://jqlang.github.io/jq/download/).

176</Note>

177

178### Block edits to protected files

179

180Prevent Claude from modifying sensitive files like `.env`, `package-lock.json`, or anything in `.git/`. Claude receives feedback explaining why the edit was blocked, so it can adjust its approach.

90 181

~~91For storage location, select `User settings` since you're logging to your home~~182This example uses a separate script file that the hook calls. The script checks the target file path against a list of protected patterns and exits with code 2 to block the edit.

~~92directory. This hook will then apply to all projects, not just your current~~

~~93project.~~

94 183

~~95Then press `Esc` until you return to the REPL. Your hook is now registered.~~184<Steps>

185 <Step title="Create the hook script">

186 Save this to `.claude/hooks/protect-files.sh`:

96 187

~~97### Step 5: Verify your hook~~188 ```bash theme={null}

189 #!/bin/bash

190 # protect-files.sh

98 191

~~99Run `/hooks` again or check `~/.claude/settings.json` to see your configuration:~~192 INPUT=$(cat)

193 FILE_PATH=$(echo "$INPUT" | jq -r '.tool_input.file_path // empty')

194

195 PROTECTED_PATTERNS=(".env" "package-lock.json" ".git/")

196

197 for pattern in "${PROTECTED_PATTERNS[@]}"; do

198 if [[ "$FILE_PATH" == *"$pattern"* ]]; then

199 echo "Blocked: $FILE_PATH matches protected pattern '$pattern'" >&2

200 exit 2

201 fi

202 done

203

204 exit 0

205 ```

206 </Step>

207

208 <Step title="Make the script executable (macOS/Linux)">

209 Hook scripts must be executable for Claude Code to run them:

210

211 ```bash theme={null}

212 chmod +x .claude/hooks/protect-files.sh

213 ```

214 </Step>

215

216 <Step title="Register the hook">

217 Add a `PreToolUse` hook to `.claude/settings.json` that runs the script before any `Edit` or `Write` tool call:

218

219 ```json theme={null}

220 {

221 "hooks": {

222 "PreToolUse": [

223 {

224 "matcher": "Edit|Write",

225 "hooks": [

226 {

227 "type": "command",

228 "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/protect-files.sh"

229 }

230 ]

231 }

232 ]

233 }

234 }

235 ```

236 </Step>

237</Steps>

238

239### Re-inject context after compaction

240

241When Claude's context window fills up, compaction summarizes the conversation to free space. This can lose important details. Use a `SessionStart` hook with a `compact` matcher to re-inject critical context after every compaction.

242

243Any text your command writes to stdout is added to Claude's context. This example reminds Claude of project conventions and recent work. Add this to `.claude/settings.json` in your project root:

100 244

101```json theme={null}245```json theme={null}

102{246{

103 "hooks": {247 "hooks": {

104 "PreToolUse": [248 "SessionStart": [

105 {249 {

106 "matcher": "Bash",250 "matcher": "compact",

107 "hooks": [251 "hooks": [

108 {252 {

109 "type": "command",253 "type": "command",

110 "command": "jq -r '\"\\(.tool_input.command) - \\(.tool_input.description // \"No description\")\"' >> ~/.claude/bash-command-log.txt"254 "command": "echo 'Reminder: use Bun, not npm. Run bun test before committing. Current sprint: auth refactor.'"

111 }255 }

112 ]256 ]

113 }257 }

116}260}

117```261```

118 262

119### Step 6: Test your hook263You can replace the `echo` with any command that produces dynamic output, like `git log --oneline -5` to show recent commits. For injecting context on every session start, consider using [CLAUDE.md](/en/memory) instead. For environment variables, see [`CLAUDE_ENV_FILE`](/en/hooks#persist-environment-variables) in the reference.

120 264

121Ask Claude to run a simple command like `ls` and check your log file:265## How hooks work

122 266

123```bash theme={null}267Hook events fire at specific lifecycle points in Claude Code. When an event fires, all matching hooks run in parallel, and identical hook commands are automatically deduplicated. The table below shows each event and when it triggers:

124cat ~/.claude/bash-command-log.txt

125```

126 268

127You should see entries like:269| Event | When it fires |

270| :------------------- | :----------------------------------------------------------------- |

271| `SessionStart` | When a session begins or resumes |

272| `UserPromptSubmit` | When you submit a prompt, before Claude processes it |

273| `PreToolUse` | Before a tool call executes. Can block it |

274| `PermissionRequest` | When a permission dialog appears |

275| `PostToolUse` | After a tool call succeeds |

276| `PostToolUseFailure` | After a tool call fails |

277| `Notification` | When Claude Code sends a notification |

278| `SubagentStart` | When a subagent is spawned |

279| `SubagentStop` | When a subagent finishes |

280| `Stop` | When Claude finishes responding |

281| `TeammateIdle` | When an [agent team](/en/agent-teams) teammate is about to go idle |

282| `TaskCompleted` | When a task is being marked as completed |

283| `PreCompact` | Before context compaction |

284| `SessionEnd` | When a session terminates |

128 285

286Each hook has a `type` that determines how it runs. Most hooks use `"type": "command"`, which runs a shell command. Two other options use a Claude model to make decisions: `"type": "prompt"` for single-turn evaluation and `"type": "agent"` for multi-turn verification with tool access. See [Prompt-based hooks](#prompt-based-hooks) and [Agent-based hooks](#agent-based-hooks) for details.

287

288### Read input and return output

289

290Hooks communicate with Claude Code through stdin, stdout, stderr, and exit codes. When an event fires, Claude Code passes event-specific data as JSON to your script's stdin. Your script reads that data, does its work, and tells Claude Code what to do next via the exit code.

291

292#### Hook input

293

294Every event includes common fields like `session_id` and `cwd`, but each event type adds different data. For example, when Claude runs a Bash command, a `PreToolUse` hook receives something like this on stdin:

295

296```json theme={null}

297{

298 "session_id": "abc123", // unique ID for this session

299 "cwd": "/Users/sarah/myproject", // working directory when the event fired

300 "hook_event_name": "PreToolUse", // which event triggered this hook

301 "tool_name": "Bash", // the tool Claude is about to use

302 "tool_input": { // the arguments Claude passed to the tool

303 "command": "npm test" // for Bash, this is the shell command

304 }

305}

129```306```

130ls - Lists files and directories307

308Your script can parse that JSON and act on any of those fields. `UserPromptSubmit` hooks get the `prompt` text instead, `SessionStart` hooks get the `source` (startup, resume, compact), and so on. See [Common input fields](/en/hooks#common-input-fields) in the reference for shared fields, and each event's section for event-specific schemas.

309

310#### Hook output

311

312Your script tells Claude Code what to do next by writing to stdout or stderr and exiting with a specific code. For example, a `PreToolUse` hook that wants to block a command:

313

314```bash theme={null}

315#!/bin/bash

316INPUT=$(cat)

317COMMAND=$(echo "$INPUT" | jq -r '.tool_input.command')

318

319if echo "$COMMAND" | grep -q "drop table"; then

320 echo "Blocked: dropping tables is not allowed" >&2 # stderr becomes Claude's feedback

321 exit 2 # exit 2 = block the action

322fi

323

324exit 0 # exit 0 = let it proceed

131```325```

132 326

133## More Examples327The exit code determines what happens next:

328

329* **Exit 0**: the action proceeds. For `UserPromptSubmit` and `SessionStart` hooks, anything you write to stdout is added to Claude's context.

330* **Exit 2**: the action is blocked. Write a reason to stderr, and Claude receives it as feedback so it can adjust.

331* **Any other exit code**: the action proceeds. Stderr is logged but not shown to Claude. Toggle verbose mode with `Ctrl+O` to see these messages in the transcript.

332

333#### Structured JSON output

334

335Exit codes give you two options: allow or block. For more control, exit 0 and print a JSON object to stdout instead.

134 336

135<Note>337<Note>

136 For a complete example implementation, see the [bash command validator example](https://github.com/anthropics/claude-code/blob/main/examples/hooks/bash_command_validator_example.py) in our public codebase.338 Use exit 2 to block with a stderr message, or exit 0 with JSON for structured control. Don't mix them: Claude Code ignores JSON when you exit 2.

137</Note>339</Note>

138 340

139### Code Formatting Hook341For example, a `PreToolUse` hook can deny a tool call and tell Claude why, or escalate it to the user for approval:

342

343```json theme={null}

344{

345 "hookSpecificOutput": {

346 "hookEventName": "PreToolUse",

347 "permissionDecision": "deny",

348 "permissionDecisionReason": "Use rg instead of grep for better performance"

349 }

350}

351```

352

353Claude Code reads `permissionDecision` and cancels the tool call, then feeds `permissionDecisionReason` back to Claude as feedback. These three options are specific to `PreToolUse`:

354

355* `"allow"`: proceed without showing a permission prompt

356* `"deny"`: cancel the tool call and send the reason to Claude

357* `"ask"`: show the permission prompt to the user as normal

140 358

141Automatically format TypeScript files after editing:359Other events use different decision patterns. For example, `PostToolUse` and `Stop` hooks use a top-level `decision: "block"` field, while `PermissionRequest` uses `hookSpecificOutput.decision.behavior`. See the [summary table](/en/hooks#decision-control) in the reference for a full breakdown by event.

360

361For `UserPromptSubmit` hooks, use `additionalContext` instead to inject text into Claude's context. Prompt-based hooks (`type: "prompt"`) handle output differently: see [Prompt-based hooks](#prompt-based-hooks).

362

363### Filter hooks with matchers

364

365Without a matcher, a hook fires on every occurrence of its event. Matchers let you narrow that down. For example, if you want to run a formatter only after file edits (not after every tool call), add a matcher to your `PostToolUse` hook:

142 366

143```json theme={null}367```json theme={null}

144{368{

147 {371 {

148 "matcher": "Edit|Write",372 "matcher": "Edit|Write",

149 "hooks": [373 "hooks": [

150 {374 { "type": "command", "command": "prettier --write ..." }

151 "type": "command",

152 "command": "jq -r '.tool_input.file_path' | { read file_path; if echo \"$file_path\" | grep -q '\\.ts$'; then npx prettier --write \"$file_path\"; fi; }"

153 }

154 ]375 ]

155 }376 }

156 ]377 ]

158}379}

159```380```

160 381

161### Markdown Formatting Hook382The `"Edit|Write"` matcher is a regex pattern that matches the tool name. The hook only fires when Claude uses the `Edit` or `Write` tool, not when it uses `Bash`, `Read`, or any other tool.

162 383

163Automatically fix missing language tags and formatting issues in markdown files:384Each event type matches on a specific field. Matchers support exact strings and regex patterns:

164 385

165```json theme={null}386| Event | What the matcher filters | Example matcher values |

166{387| :--------------------------------------------------------------------- | :------------------------ | :----------------------------------------------------------------------------- |

389| `SessionStart` | how the session started | `startup`, `resume`, `clear`, `compact` |

390| `SessionEnd` | why the session ended | `clear`, `logout`, `prompt_input_exit`, `bypass_permissions_disabled`, `other` |

391| `Notification` | notification type | `permission_prompt`, `idle_prompt`, `auth_success`, `elicitation_dialog` |

392| `SubagentStart` | agent type | `Bash`, `Explore`, `Plan`, or custom agent names |

393| `PreCompact` | what triggered compaction | `manual`, `auto` |

394| `UserPromptSubmit`, `Stop`, `TeammateIdle`, `TaskCompleted` | no matcher support | always fires on every occurrence |

395| `SubagentStop` | agent type | same values as `SubagentStart` |

396

397A few more examples showing matchers on different event types:

398

399<Tabs>

400 <Tab title="Log every Bash command">

401 Match only `Bash` tool calls and log each command to a file. The `PostToolUse` event fires after the command completes, so `tool_input.command` contains what ran. The hook receives the event data as JSON on stdin, and `jq -r '.tool_input.command'` extracts just the command string, which `>>` appends to the log file:

402

403 ```json theme={null}

404 {

167 "hooks": {405 "hooks": {

168 "PostToolUse": [406 "PostToolUse": [

169 {407 {

170 "matcher": "Edit|Write",408 "matcher": "Bash",

171 "hooks": [409 "hooks": [

172 {410 {

173 "type": "command",411 "type": "command",

174 "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/markdown_formatter.py"412 "command": "jq -r '.tool_input.command' >> ~/.claude/command-log.txt"

175 }413 }

176 ]414 ]

177 }415 }

178 ]416 ]

179 }417 }

180}418 }

181```419 ```

420 </Tab>

182 421

183Create `.claude/hooks/markdown_formatter.py` with this content:422 <Tab title="Match MCP tools">

~~184~~ 423 MCP tools use a different naming convention than built-in tools: `mcp__<server>__<tool>`, where `<server>` is the MCP server name and `<tool>` is the tool it provides. For example, `mcp__github__search_repositories` or `mcp__filesystem__read_file`. Use a regex matcher to target all tools from a specific server, or match across servers with a pattern like `mcp__.*__write.*`. See [Match MCP tools](/en/hooks#match-mcp-tools) in the reference for the full list of examples.

185````python theme={null}

186#!/usr/bin/env python3

187"""

188Markdown formatter for Claude Code output.

189Fixes missing language tags and spacing issues while preserving code content.

190"""

191import json

192import sys

193import re

194import os

~~195~~

196def detect_language(code):

197 """Best-effort language detection from code content."""

198 s = code.strip()

~~199~~

200 # JSON detection

201 if re.search(r'^\s*[{\[]', s):

202 try:

203 json.loads(s)

204 return 'json'

205 except:

206 pass

~~207~~

208 # Python detection

209 if re.search(r'^\s*def\s+\w+\s*\(', s, re.M) or \

210 re.search(r'^\s*(import|from)\s+\w+', s, re.M):

211 return 'python'

~~212~~

213 # JavaScript detection

214 if re.search(r'\b(function\s+\w+\s*\(|const\s+\w+\s*=)', s) or \

215 re.search(r'=>|console\.(log|error)', s):

216 return 'javascript'

~~217~~

218 # Bash detection

219 if re.search(r'^#!.*\b(bash|sh)\b', s, re.M) or \

220 re.search(r'\b(if|then|fi|for|in|do|done)\b', s):

221 return 'bash'

~~222~~

223 # SQL detection

224 if re.search(r'\b(SELECT|INSERT|UPDATE|DELETE|CREATE)\s+', s, re.I):

225 return 'sql'

~~226~~

227 return 'text'

~~228~~

229def format_markdown(content):

230 """Format markdown content with language detection."""

231 # Fix unlabeled code fences

232 def add_lang_to_fence(match):

233 indent, info, body, closing = match.groups()

234 if not info.strip():

235 lang = detect_language(body)

236 return f"{indent}```{lang}\n{body}{closing}\n"

237 return match.group(0)

~~238~~

239 fence_pattern = r'(?ms)^([ \t]{0,3})```([^\n]*)\n(.*?)(\n\1```)\s*$'

240 content = re.sub(fence_pattern, add_lang_to_fence, content)

~~241~~

242 # Fix excessive blank lines (only outside code fences)

243 content = re.sub(r'\n{3,}', '\n\n', content)

~~244~~

245 return content.rstrip() + '\n'

~~246~~

247# Main execution

248try:

249 input_data = json.load(sys.stdin)

250 file_path = input_data.get('tool_input', {}).get('file_path', '')

~~251~~

252 if not file_path.endswith(('.md', '.mdx')):

253 sys.exit(0) # Not a markdown file

~~254~~

255 if os.path.exists(file_path):

256 with open(file_path, 'r', encoding='utf-8') as f:

257 content = f.read()

~~258~~

259 formatted = format_markdown(content)

~~260~~

261 if formatted != content:

262 with open(file_path, 'w', encoding='utf-8') as f:

263 f.write(formatted)

264 print(f"✓ Fixed markdown formatting in {file_path}")

~~265~~

266except Exception as e:

267 print(f"Error formatting markdown: {e}", file=sys.stderr)

268 sys.exit(1)

269````

~~270~~

271Make the script executable:

272 424

273```bash theme={null}425 The command below extracts the tool name from the hook's JSON input with `jq` and writes it to stderr, where it shows up in verbose mode (`Ctrl+O`):

274chmod +x .claude/hooks/markdown_formatter.py

275```

276 426

277This hook automatically:427 ```json theme={null}

428 {

429 "hooks": {

430 "PreToolUse": [

431 {

432 "matcher": "mcp__github__.*",

433 "hooks": [

434 {

435 "type": "command",

436 "command": "echo \"GitHub tool called: $(jq -r '.tool_name')\" >&2"

437 }

438 ]

439 }

440 ]

441 }

442 }

443 ```

444 </Tab>

278 445

279* Detects programming languages in unlabeled code blocks446 <Tab title="Clean up on session end">

280* Adds appropriate language tags for syntax highlighting447 The `SessionEnd` event supports matchers on the reason the session ended. This hook only fires on `clear` (when you run `/clear`), not on normal exits:

281* Fixes excessive blank lines while preserving code content

282* Only processes markdown files (`.md`, `.mdx`)

283 448

284### Custom Notification Hook449 ```json theme={null}

450 {

451 "hooks": {

452 "SessionEnd": [

453 {

454 "matcher": "clear",

455 "hooks": [

456 {

457 "type": "command",

458 "command": "rm -f /tmp/claude-scratch-*.txt"

459 }

460 ]

461 }

462 ]

463 }

464 }

465 ```

466 </Tab>

467</Tabs>

468

469For full matcher syntax, see the [Hooks reference](/en/hooks#configuration).

470

471### Configure hook location

472

473Where you add a hook determines its scope:

474

475| Location | Scope | Shareable |

476| :--------------------------------------------------------- | :--------------------------------- | :--------------------------------- |

477| `~/.claude/settings.json` | All your projects | No, local to your machine |

478| `.claude/settings.json` | Single project | Yes, can be committed to the repo |

479| `.claude/settings.local.json` | Single project | No, gitignored |

480| Managed policy settings | Organization-wide | Yes, admin-controlled |

481| [Plugin](/en/plugins) `hooks/hooks.json` | When plugin is enabled | Yes, bundled with the plugin |

482| [Skill](/en/skills) or [agent](/en/sub-agents) frontmatter | While the skill or agent is active | Yes, defined in the component file |

285 483

286Get desktop notifications when Claude needs input:484You can also use the [`/hooks` menu](/en/hooks#the-hooks-menu) in Claude Code to add, delete, and view hooks interactively. To disable all hooks at once, use the toggle at the bottom of the `/hooks` menu or set `"disableAllHooks": true` in your settings file.

485

486Hooks added through the `/hooks` menu take effect immediately. If you edit settings files directly while Claude Code is running, the changes won't take effect until you review them in the `/hooks` menu or restart your session.

487

488## Prompt-based hooks

489

490For decisions that require judgment rather than deterministic rules, use `type: "prompt"` hooks. Instead of running a shell command, Claude Code sends your prompt and the hook's input data to a Claude model (Haiku by default) to make the decision. You can specify a different model with the `model` field if you need more capability.

491

492The model's only job is to return a yes/no decision as JSON:

493

494* `"ok": true`: the action proceeds

495* `"ok": false`: the action is blocked. The model's `"reason"` is fed back to Claude so it can adjust.

496

497This example uses a `Stop` hook to ask the model whether all requested tasks are complete. If the model returns `"ok": false`, Claude keeps working and uses the `reason` as its next instruction:

287 498

288```json theme={null}499```json theme={null}

289{500{

290 "hooks": {501 "hooks": {

291 "Notification": [502 "Stop": [

292 {503 {

293 "matcher": "",

294 "hooks": [504 "hooks": [

295 {505 {

296 "type": "command",506 "type": "prompt",

297 "command": "notify-send 'Claude Code' 'Awaiting your input'"507 "prompt": "Check if all tasks are complete. If not, respond with {\"ok\": false, \"reason\": \"what remains to be done\"}."

298 }508 }

299 ]509 ]

300 }510 }

303}513}

304```514```

305 515

306### File Protection Hook516For full configuration options, see [Prompt-based hooks](/en/hooks#prompt-based-hooks) in the reference.

517

518## Agent-based hooks

519

520When verification requires inspecting files or running commands, use `type: "agent"` hooks. Unlike prompt hooks which make a single LLM call, agent hooks spawn a subagent that can read files, search code, and use other tools to verify conditions before returning a decision.

307 521

308Block edits to sensitive files:522Agent hooks use the same `"ok"` / `"reason"` response format as prompt hooks, but with a longer default timeout of 60 seconds and up to 50 tool-use turns.

523

524This example verifies that tests pass before allowing Claude to stop:

309 525

310```json theme={null}526```json theme={null}

311{527{

312 "hooks": {528 "hooks": {

313 "PreToolUse": [529 "Stop": [

314 {530 {

315 "matcher": "Edit|Write",

316 "hooks": [531 "hooks": [

317 {532 {

318 "type": "command",533 "type": "agent",

319 "command": "python3 -c \"import json, sys; data=json.load(sys.stdin); path=data.get('tool_input',{}).get('file_path',''); sys.exit(2 if any(p in path for p in ['.env', 'package-lock.json', '.git/']) else 0)\""534 "prompt": "Verify that all unit tests pass. Run the test suite and check the results. $ARGUMENTS",

535 "timeout": 120

320 }536 }

321 ]537 ]

322 }538 }

325}541}

326```542```

327 543

328## Learn more544Use prompt hooks when the hook input data alone is enough to make a decision. Use agent hooks when you need to verify something against the actual state of the codebase.

545

546For full configuration options, see [Agent-based hooks](/en/hooks#agent-based-hooks) in the reference.

547

548## Limitations and troubleshooting

549

550### Limitations

551

552* Hooks communicate through stdout, stderr, and exit codes only. They cannot trigger slash commands or tool calls directly.

553* Hook timeout is 10 minutes by default, configurable per hook with the `timeout` field (in seconds).

554* `PostToolUse` hooks cannot undo actions since the tool has already executed.

555* `PermissionRequest` hooks do not fire in [non-interactive mode](/en/headless) (`-p`). Use `PreToolUse` hooks for automated permission decisions.

556* `Stop` hooks fire whenever Claude finishes responding, not only at task completion. They do not fire on user interrupts.

329 557

330* For reference documentation on hooks, see [Hooks reference](/en/hooks).558### Hook not firing

331* For comprehensive security best practices and safety guidelines, see [Security Considerations](/en/hooks#security-considerations) in the hooks reference documentation.

332* For troubleshooting steps and debugging techniques, see [Debugging](/en/hooks#debugging) in the hooks reference

333 documentation.

334 559

560The hook is configured but never executes.

335 561

562* Run `/hooks` and confirm the hook appears under the correct event

563* Check that the matcher pattern matches the tool name exactly (matchers are case-sensitive)

564* Verify you're triggering the right event type (e.g., `PreToolUse` fires before tool execution, `PostToolUse` fires after)

565* If using `PermissionRequest` hooks in non-interactive mode (`-p`), switch to `PreToolUse` instead

566

567### Hook error in output

568

569You see a message like "PreToolUse hook error: ..." in the transcript.

570

571* Your script exited with a non-zero code unexpectedly. Test it manually by piping sample JSON:

572 ```bash theme={null}

573 echo '{"tool_name":"Bash","tool_input":{"command":"ls"}}' | ./my-hook.sh

574 echo $? # Check the exit code

575 ```

576* If you see "command not found", use absolute paths or `$CLAUDE_PROJECT_DIR` to reference scripts

577* If you see "jq: command not found", install `jq` or use Python/Node.js for JSON parsing

578* If the script isn't running at all, make it executable: `chmod +x ./my-hook.sh`

579

580### `/hooks` shows no hooks configured

581

582You edited a settings file but the hooks don't appear in the menu.

583

584* Restart your session or open `/hooks` to reload. Hooks added through the `/hooks` menu take effect immediately, but manual file edits require a reload.

585* Verify your JSON is valid (trailing commas and comments are not allowed)

586* Confirm the settings file is in the correct location: `.claude/settings.json` for project hooks, `~/.claude/settings.json` for global hooks

587

588### Stop hook runs forever

589

590Claude keeps working in an infinite loop instead of stopping.

591

592Your Stop hook script needs to check whether it already triggered a continuation. Parse the `stop_hook_active` field from the JSON input and exit early if it's `true`:

593

594```bash theme={null}

595#!/bin/bash

596INPUT=$(cat)

597if [ "$(echo "$INPUT" | jq -r '.stop_hook_active')" = "true" ]; then

598 exit 0 # Allow Claude to stop

599fi

600# ... rest of your hook logic

601```

602

603### JSON validation failed

604

605Claude Code shows a JSON parsing error even though your hook script outputs valid JSON.

606

607When Claude Code runs a hook, it spawns a shell that sources your profile (`~/.zshrc` or `~/.bashrc`). If your profile contains unconditional `echo` statements, that output gets prepended to your hook's JSON:

608

609```

610Shell ready on arm64

611{"decision": "block", "reason": "Not allowed"}

612```

613

614Claude Code tries to parse this as JSON and fails. To fix this, wrap echo statements in your shell profile so they only run in interactive shells:

615

616```bash theme={null}

617# In ~/.zshrc or ~/.bashrc

618if [[ $- == *i* ]]; then

619 echo "Shell ready"

620fi

621```

622

623The `$-` variable contains shell flags, and `i` means interactive. Hooks run in non-interactive shells, so the echo is skipped.

624

625### Debug techniques

626

627Toggle verbose mode with `Ctrl+O` to see hook output in the transcript, or run `claude --debug` for full execution details including which hooks matched and their exit codes.

628

629## Learn more

336 630

337> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt631* [Hooks reference](/en/hooks): full event schemas, JSON output format, async hooks, and MCP tool hooks

632* [Security considerations](/en/hooks#security-considerations): review before deploying hooks in shared or production environments

633* [Bash command validator example](https://github.com/anthropics/claude-code/blob/main/examples/hooks/bash_command_validator_example.py): complete reference implementation

how-claude-code-works.md +240 −0 added

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# How Claude Code works

7> Understand the agentic loop, built-in tools, and how Claude Code interacts with your project.

9Claude Code is an agentic assistant that runs in your terminal. While it excels at coding, it can help with anything you can do from the command line: writing docs, running builds, searching files, researching topics, and more.

11This guide covers the core architecture, built-in capabilities, and [tips for working effectively](#work-effectively-with-claude-code). For step-by-step walkthroughs, see [Common workflows](/en/common-workflows). For extensibility features like skills, MCP, and hooks, see [Extend Claude Code](/en/features-overview).

13## The agentic loop

15When you give Claude a task, it works through three phases: **gather context**, **take action**, and **verify results**. These phases blend together. Claude uses tools throughout, whether searching files to understand your code, editing to make changes, or running tests to check its work.

17<img src="https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/agentic-loop.svg?fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=e30acfc80d6ff01ec877dd19c7af58b2" alt="The agentic loop: Your prompt leads to Claude gathering context, taking action, verifying results, and repeating until task complete. You can interrupt at any point." data-og-width="720" width="720" data-og-height="280" height="280" data-path="images/agentic-loop.svg" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/agentic-loop.svg?w=280&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=8620f6ebce761a1e8bbf7f0a0255cc15 280w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/agentic-loop.svg?w=560&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=7b46b5ff4454aa4a03725eee625b39a0 560w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/agentic-loop.svg?w=840&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=7fa0397bc37d147e3bf3bb6296c6477f 840w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/agentic-loop.svg?w=1100&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=73b2a7040c4c93821c4d5bbee9f4a2d4 1100w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/agentic-loop.svg?w=1650&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=17703cbeb6f59b40a00ab24f56d5f8f9 1650w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/agentic-loop.svg?w=2500&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=20dedb60b95d45a1bd60a0cccaf3e1ff 2500w" />

19The loop adapts to what you ask. A question about your codebase might only need context gathering. A bug fix cycles through all three phases repeatedly. A refactor might involve extensive verification. Claude decides what each step requires based on what it learned from the previous step, chaining dozens of actions together and course-correcting along the way.

21You're part of this loop too. You can interrupt at any point to steer Claude in a different direction, provide additional context, or ask it to try a different approach. Claude works autonomously but stays responsive to your input.

23The agentic loop is powered by two components: [models](#models) that reason and [tools](#tools) that act. Claude Code serves as the **agentic harness** around Claude: it provides the tools, context management, and execution environment that turn a language model into a capable coding agent.

25### Models

27Claude Code uses Claude models to understand your code and reason about tasks. Claude can read code in any language, understand how components connect, and figure out what needs to change to accomplish your goal. For complex tasks, it breaks work into steps, executes them, and adjusts based on what it learns.

29[Multiple models](/en/model-config) are available with different tradeoffs. Sonnet handles most coding tasks well. Opus provides stronger reasoning for complex architectural decisions. Switch with `/model` during a session or start with `claude --model <name>`.

31When this guide says "Claude chooses" or "Claude decides," it's the model doing the reasoning.

33### Tools

35Tools are what make Claude Code agentic. Without tools, Claude can only respond with text. With tools, Claude can act: read your code, edit files, run commands, search the web, and interact with external services. Each tool use returns information that feeds back into the loop, informing Claude's next decision.

37The built-in tools generally fall into four categories, each representing a different kind of agency.

39| Category | What Claude can do |

40| --------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------- |

41| **File operations** | Read files, edit code, create new files, rename and reorganize |

42| **Search** | Find files by pattern, search content with regex, explore codebases |

43| **Execution** | Run shell commands, start servers, run tests, use git |

44| **Web** | Search the web, fetch documentation, look up error messages |

45| **Code intelligence** | See type errors and warnings after edits, jump to definitions, find references (requires [code intelligence plugins](/en/discover-plugins#code-intelligence)) |

47These are the primary capabilities. Claude also has tools for spawning subagents, asking you questions, and other orchestration tasks. See [Tools available to Claude](/en/settings#tools-available-to-claude) for the complete list.

49Claude chooses which tools to use based on your prompt and what it learns along the way. When you say "fix the failing tests," Claude might:

511. Run the test suite to see what's failing

522. Read the error output

533. Search for the relevant source files

544. Read those files to understand the code

555. Edit the files to fix the issue

566. Run the tests again to verify

58Each tool use gives Claude new information that informs the next step. This is the agentic loop in action.

60**Extending the base capabilities:** The built-in tools are the foundation. You can extend what Claude knows with [skills](/en/skills), connect to external services with [MCP](/en/mcp), automate workflows with [hooks](/en/hooks), and offload tasks to [subagents](/en/sub-agents). These extensions form a layer on top of the core agentic loop. See [Extend Claude Code](/en/features-overview) for guidance on choosing the right extension for your needs.

62## What Claude can access

64This guide focuses on the terminal. Claude Code also runs in [VS Code, JetBrains IDEs, and other environments](/en/ide-integrations).

66When you run `claude` in a directory, Claude Code gains access to:

68* **Your project.** Files in your directory and subdirectories, plus files elsewhere with your permission.

69* **Your terminal.** Any command you could run: build tools, git, package managers, system utilities, scripts. If you can do it from the command line, Claude can too.

70* **Your git state.** Current branch, uncommitted changes, and recent commit history.

71* **Your [CLAUDE.md](/en/memory).** A markdown file where you store project-specific instructions, conventions, and context that Claude should know every session.

72* **Extensions you configure.** [MCP servers](/en/mcp) for external services, [skills](/en/skills) for workflows, [subagents](/en/sub-agents) for delegated work, and [Claude in Chrome](/en/chrome) for browser interaction.

74Because Claude sees your whole project, it can work across it. When you ask Claude to "fix the authentication bug," it searches for relevant files, reads multiple files to understand context, makes coordinated edits across them, runs tests to verify the fix, and commits the changes if you ask. This is different from inline code assistants that only see the current file.

76## Work with sessions

78Claude Code saves your conversation locally as you work. Each message, tool use, and result is stored, which enables [rewinding](#undo-changes-with-checkpoints), [resuming, and forking](#resume-or-fork-sessions) sessions. Before Claude makes code changes, it also snapshots the affected files so you can revert if needed.

80**Sessions are independent.** Each new session starts with a fresh context window, without the conversation history from previous sessions. Claude can persist learnings across sessions using [auto memory](/en/memory#auto-memory), and you can add your own persistent instructions in [CLAUDE.md](/en/memory).

82### Work across branches

84Each Claude Code conversation is a session tied to your current directory. When you resume, you only see sessions from that directory.

86Claude sees your current branch's files. When you switch branches, Claude sees the new branch's files, but your conversation history stays the same. Claude remembers what you discussed even after switching.

88Since sessions are tied to directories, you can run parallel Claude sessions by using [git worktrees](/en/common-workflows#run-parallel-claude-code-sessions-with-git-worktrees), which create separate directories for individual branches.

90### Resume or fork sessions

92When you resume a session with `claude --continue` or `claude --resume`, you pick up where you left off using the same session ID. New messages append to the existing conversation. Your full conversation history is restored, but session-scoped permissions are not. You'll need to re-approve those.

94<img src="https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/session-continuity.svg?fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=f671b603cc856119c95475b9084ebfef" alt="Session continuity: resume continues the same session, fork creates a new branch with a new ID." data-og-width="560" width="560" data-og-height="280" height="280" data-path="images/session-continuity.svg" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/session-continuity.svg?w=280&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=bddf1f33d419a27d7427acdf06058804 280w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/session-continuity.svg?w=560&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=417478eb9b86003b8eebaac058a8618a 560w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/session-continuity.svg?w=840&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=1d89d26e2c0487f067d187c3fa5f7170 840w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/session-continuity.svg?w=1100&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=8ea739a1f7860e4edbbcf74d444e37b2 1100w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/session-continuity.svg?w=1650&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=9cb5095d6a8920f04c3b78d31a69c809 1650w, https://mintcdn.com/claude-code/ELkJZG54dIaeldDC/images/session-continuity.svg?w=2500&fit=max&auto=format&n=ELkJZG54dIaeldDC&q=85&s=d67e1744e4878813d20c6c3f39d9459d 2500w" />

96To branch off and try a different approach without affecting the original session, use the `--fork-session` flag:

98```bash theme={null}

99claude --continue --fork-session

100```

101

102This creates a new session ID while preserving the conversation history up to that point. The original session remains unchanged. Like resume, forked sessions don't inherit session-scoped permissions.

103

104**Same session in multiple terminals**: If you resume the same session in multiple terminals, both terminals write to the same session file. Messages from both get interleaved, like two people writing in the same notebook. Nothing corrupts, but the conversation becomes jumbled. Each terminal only sees its own messages during the session, but if you resume that session later, you'll see everything interleaved. For parallel work from the same starting point, use `--fork-session` to give each terminal its own clean session.

105

106### The context window

107

108Claude's context window holds your conversation history, file contents, command outputs, [CLAUDE.md](/en/memory), loaded skills, and system instructions. As you work, context fills up. Claude compacts automatically, but instructions from early in the conversation can get lost. Put persistent rules in CLAUDE.md, and run `/context` to see what's using space.

109

110#### When context fills up

111

112Claude Code manages context automatically as you approach the limit. It clears older tool outputs first, then summarizes the conversation if needed. Your requests and key code snippets are preserved; detailed instructions from early in the conversation may be lost. Put persistent rules in CLAUDE.md rather than relying on conversation history.

113

114To control what's preserved during compaction, add a "Compact Instructions" section to CLAUDE.md or run `/compact` with a focus (like `/compact focus on the API changes`).

115

116Run `/context` to see what's using space. MCP servers add tool definitions to every request, so a few servers can consume significant context before you start working. Run `/mcp` to check per-server costs.

117

118#### Manage context with skills and subagents

119

120Beyond compaction, you can use other features to control what loads into context.

121

122[Skills](/en/skills) load on demand. Claude sees skill descriptions at session start, but the full content only loads when a skill is used. For skills you invoke manually, set `disable-model-invocation: true` to keep descriptions out of context until you need them.

123

124[Subagents](/en/sub-agents) get their own fresh context, completely separate from your main conversation. Their work doesn't bloat your context. When done, they return a summary. This isolation is why subagents help with long sessions.

125

126See [context costs](/en/features-overview#understand-context-costs) for what each feature costs, and [reduce token usage](/en/costs#reduce-token-usage) for tips on managing context.

127

128## Stay safe with checkpoints and permissions

129

130Claude has two safety mechanisms: checkpoints let you undo file changes, and permissions control what Claude can do without asking.

131

132### Undo changes with checkpoints

133

134**Every file edit is reversible.** Before Claude edits any file, it snapshots the current contents. If something goes wrong, press `Esc` twice to rewind to a previous state, or ask Claude to undo.

135

136Checkpoints are local to your session, separate from git. They only cover file changes. Actions that affect remote systems (databases, APIs, deployments) can't be checkpointed, which is why Claude asks before running commands with external side effects.

137

138### Control what Claude can do

139

140Press `Shift+Tab` to cycle through permission modes:

141

142* **Default**: Claude asks before file edits and shell commands

143* **Auto-accept edits**: Claude edits files without asking, still asks for commands

144* **Plan mode**: Claude uses read-only tools only, creating a plan you can approve before execution

145* **Delegate mode**: Claude coordinates work through [agent teammates](/en/agent-teams) only, with no direct implementation. Only available when an agent team is active.

146

147You can also allow specific commands in `.claude/settings.json` so Claude doesn't ask each time. This is useful for trusted commands like `npm test` or `git status`. Settings can be scoped from organization-wide policies down to personal preferences. See [Permissions](/en/permissions) for details.

148

149***

150

151## Work effectively with Claude Code

152

153These tips help you get better results from Claude Code.

154

155### Ask Claude Code for help

156

157Claude Code can teach you how to use it. Ask questions like "how do I set up hooks?" or "what's the best way to structure my CLAUDE.md?" and Claude will explain.

158

159Built-in commands also guide you through setup:

160

161* `/init` walks you through creating a CLAUDE.md for your project

162* `/agents` helps you configure custom subagents

163* `/doctor` diagnoses common issues with your installation

164

165### It's a conversation

166

167Claude Code is conversational. You don't need perfect prompts. Start with what you want, then refine:

168

169```

170> Fix the login bug

171

172[Claude investigates, tries something]

173

174> That's not quite right. The issue is in the session handling.

175

176[Claude adjusts approach]

177```

178

179When the first attempt isn't right, you don't start over. You iterate.

180

181#### Interrupt and steer

182

183You can interrupt Claude at any point. If it's going down the wrong path, just type your correction and press Enter. Claude will stop what it's doing and adjust its approach based on your input. You don't have to wait for it to finish or start over.

184

185### Be specific upfront

186

187The more precise your initial prompt, the fewer corrections you'll need. Reference specific files, mention constraints, and point to example patterns.

188

189```

190> The checkout flow is broken for users with expired cards.

191> Check src/payments/ for the issue, especially token refresh.

192> Write a failing test first, then fix it.

193```

194

195Vague prompts like "fix the login bug" work, but you'll spend more time steering. Specific prompts like the above often succeed on the first attempt.

196

197### Give Claude something to verify against

198

199Claude performs better when it can check its own work. Include test cases, paste screenshots of expected UI, or define the output you want.

200

201```

202> Implement validateEmail. Test cases: 'user@example.com' → true,

203> 'invalid' → false, 'user@.com' → false. Run the tests after.

204```

205

206For visual work, paste a screenshot of the design and ask Claude to compare its implementation against it.

207

208### Explore before implementing

209

210For complex problems, separate research from coding. Use plan mode (`Shift+Tab` twice) to analyze the codebase first:

211

212```

213> Read src/auth/ and understand how we handle sessions.

214> Then create a plan for adding OAuth support.

215```

216

217Review the plan, refine it through conversation, then let Claude implement. This two-phase approach produces better results than jumping straight to code.

218

219### Delegate, don't dictate

220

221Think of delegating to a capable colleague. Give context and direction, then trust Claude to figure out the details:

222

223```

224> The checkout flow is broken for users with expired cards.

225> The relevant code is in src/payments/. Can you investigate and fix it?

226```

227

228You don't need to specify which files to read or what commands to run. Claude figures that out.

229

230## What's next

231

232<CardGroup cols={2}>

233 <Card title="Extend with features" icon="puzzle-piece" href="/en/features-overview">

234 Add Skills, MCP connections, and custom commands

235 </Card>

236

237 <Card title="Common workflows" icon="graduation-cap" href="/en/common-workflows">

238 Step-by-step guides for typical tasks

239 </Card>

240</CardGroup>

iam.md +0 −188 deleted

File DeletedView Diff

~~1# Identity and Access Management~~

~~3> Learn how to configure user authentication, authorization, and access controls for Claude Code in your organization.~~

~~5## Authentication methods~~

~~7Setting up Claude Code requires access to Anthropic models. For teams, you can set up Claude Code access in one of four ways:~~

~~9* Claude API via the Claude Console~~

~~10* Amazon Bedrock~~

~~11* Microsoft Foundry~~

~~12* Google Vertex AI~~

~~14### Claude API authentication~~

~~16**To set up Claude Code access for your team via Claude API:**~~

~~181. Use your existing Claude Console account or create a new Claude Console account~~

~~192. You can add users through either method below:~~

~~20 * Bulk invite users from within the Console (Console -> Settings -> Members -> Invite)~~

~~21 * [Set up SSO](https://support.claude.com/en/articles/10280258-setting-up-single-sign-on-on-the-api-console)~~

~~223. When inviting users, they need one of the following roles:~~

~~23 * "Claude Code" role means users can only create Claude Code API keys~~

~~24 * "Developer" role means users can create any kind of API key~~

~~254. Each invited user needs to complete these steps:~~

~~26 * Accept the Console invite~~

~~27 * [Check system requirements](/en/setup#system-requirements)~~

~~28 * [Install Claude Code](/en/setup#installation)~~

~~29 * Login with Console account credentials~~

~~31### Cloud provider authentication~~

~~33**To set up Claude Code access for your team via Bedrock, Vertex, or Azure:**~~

~~351. Follow the [Bedrock docs](/en/amazon-bedrock), [Vertex docs](/en/google-vertex-ai), or [Microsoft Foundry docs](/en/microsoft-foundry)~~

~~362. Distribute the environment variables and instructions for generating cloud credentials to your users. Read more about how to [manage configuration here](/en/settings).~~

~~373. Users can [install Claude Code](/en/setup#installation)~~

~~39## Access control and permissions~~

41We support fine-grained permissions so that you're able to specify exactly what the agent is allowed to do (e.g. run tests, run linter) and what it is not allowed to do (e.g. update cloud infrastructure). These permission settings can be checked into version control and distributed to all developers in your organization, as well as customized by individual developers.

~~43### Permission system~~

~~45Claude Code uses a tiered permission system to balance power and safety:~~

~~48| :---------------- | :------------------- | :---------------- | :-------------------------------------------- |~~

~~49| Read-only | File reads, LS, Grep | No | N/A |~~

~~50| Bash Commands | Shell execution | Yes | Permanently per project directory and command |~~

~~51| File Modification | Edit/write files | Yes | Until session end |~~

~~53### Configuring permissions~~

~~55You can view & manage Claude Code's tool permissions with `/permissions`. This UI lists all permission rules and the settings.json file they are sourced from.~~

~~57* **Allow** rules will allow Claude Code to use the specified tool without further manual approval.~~

~~58* **Ask** rules will ask the user for confirmation whenever Claude Code tries to use the specified tool. Ask rules take precedence over allow rules.~~

~~59* **Deny** rules will prevent Claude Code from using the specified tool. Deny rules take precedence over allow and ask rules.~~

~~60* **Additional directories** extend Claude's file access to directories beyond the initial working directory.~~

~~61* **Default mode** controls Claude's permission behavior when encountering new requests.~~

~~63Permission rules use the format: `Tool` or `Tool(optional-specifier)`~~

~~65A rule that is just the tool name matches any use of that tool. For example, adding `Bash` to the list of allow rules would allow Claude Code to use the Bash tool without requiring user approval.~~

~~67#### Permission modes~~

~~69Claude Code supports several permission modes that can be set as the `defaultMode` in [settings files](/en/settings#settings-files):~~

~~71| Mode | Description |~~

~~72| :------------------ | :------------------------------------------------------------------------------------------------------------------------ |~~

~~73| `default` | Standard behavior - prompts for permission on first use of each tool |~~

~~74| `acceptEdits` | Automatically accepts file edit permissions for the session |~~

~~75| `plan` | Plan Mode - Claude can analyze but not modify files or execute commands |~~

~~76| `dontAsk` | Auto-denies tools unless pre-approved via `/permissions` or [`permissions.allow`](/en/settings#permission-settings) rules |~~

~~77| `bypassPermissions` | Skips all permission prompts (requires safe environment - see warning below) |~~

~~79#### Working directories~~

~~81By default, Claude has access to files in the directory where it was launched. You can extend this access:~~

~~83* **During startup**: Use `--add-dir <path>` CLI argument~~

~~84* **During session**: Use `/add-dir` slash command~~

~~85* **Persistent configuration**: Add to `additionalDirectories` in [settings files](/en/settings#settings-files)~~

87Files in additional directories follow the same permission rules as the original working directory - they become readable without prompts, and file editing permissions follow the current permission mode.

~~89#### Tool-specific permission rules~~

~~91Some tools support more fine-grained permission controls:~~

~~93**Bash**~~

~~95* `Bash(npm run build)` Matches the exact Bash command `npm run build`~~

~~96* `Bash(npm run test:*)` Matches Bash commands starting with `npm run test`~~

~~97* `Bash(curl http://site.com/:*)` Matches curl commands that start with exactly `curl http://site.com/`~~

~~99<Tip>~~

100 Claude Code is aware of shell operators (like `&&`) so a prefix match rule like `Bash(safe-cmd:*)` won't give it permission to run the command `safe-cmd && other-cmd`

101</Tip>

~~102~~

103<Warning>

104 Important limitations of Bash permission patterns:

~~105~~

106 1. This tool uses **prefix matches**, not regex or glob patterns

107 2. The wildcard `:*` only works at the end of a pattern to match any continuation

108 3. Patterns like `Bash(curl http://github.com/:*)` can be bypassed in many ways:

109 * Options before URL: `curl -X GET http://github.com/...` won't match

110 * Different protocol: `curl https://github.com/...` won't match

111 * Redirects: `curl -L http://bit.ly/xyz` (redirects to github)

112 * Variables: `URL=http://github.com && curl $URL` won't match

113 * Extra spaces: `curl http://github.com` won't match

~~114~~

115 For more reliable URL filtering, consider:

~~116~~

117 * Using the WebFetch tool with `WebFetch(domain:github.com)` permission

118 * Instructing Claude Code about your allowed curl patterns via CLAUDE.md

119 * Using hooks for custom permission validation

120</Warning>

~~121~~

122**Read & Edit**

~~123~~

124`Edit` rules apply to all built-in tools that edit files. Claude will make a best-effort attempt to apply `Read` rules to all built-in tools that read files like Grep, Glob, and LS.

~~125~~

126Read & Edit rules both follow the [gitignore](https://git-scm.com/docs/gitignore) specification with four distinct pattern types:

~~127~~

129| ------------------ | -------------------------------------- | -------------------------------- | ---------------------------------- |

130| `//path` | **Absolute** path from filesystem root | `Read(//Users/alice/secrets/**)` | `/Users/alice/secrets/**` |

131| `~/path` | Path from **home** directory | `Read(~/Documents/*.pdf)` | `/Users/alice/Documents/*.pdf` |

132| `/path` | Path **relative to settings file** | `Edit(/src/**/*.ts)` | `<settings file path>/src/**/*.ts` |

133| `path` or `./path` | Path **relative to current directory** | `Read(*.env)` | `<cwd>/*.env` |

~~134~~

135<Warning>

136 A pattern like `/Users/alice/file` is NOT an absolute path - it's relative to your settings file! Use `//Users/alice/file` for absolute paths.

137</Warning>

~~138~~

139* `Edit(/docs/**)` - Edits in `<project>/docs/` (NOT `/docs/`!)

140* `Read(~/.zshrc)` - Reads your home directory's `.zshrc`

141* `Edit(//tmp/scratch.txt)` - Edits the absolute path `/tmp/scratch.txt`

142* `Read(src/**)` - Reads from `<current-directory>/src/`

~~143~~

144**WebFetch**

~~145~~

146* `WebFetch(domain:example.com)` Matches fetch requests to example.com

~~147~~

148**MCP**

~~149~~

150* `mcp__puppeteer` Matches any tool provided by the `puppeteer` server (name configured in Claude Code)

151* `mcp__puppeteer__*` Wildcard syntax that also matches all tools from the `puppeteer` server

152* `mcp__puppeteer__puppeteer_navigate` Matches the `puppeteer_navigate` tool provided by the `puppeteer` server

~~153~~

154### Additional permission control with hooks

~~155~~

156[Claude Code hooks](/en/hooks-guide) provide a way to register custom shell commands to perform permission evaluation at runtime. When Claude Code makes a tool call, PreToolUse hooks run before the permission system runs, and the hook output can determine whether to approve or deny the tool call in place of the permission system.

~~157~~

158### Enterprise managed settings

~~159~~

160For enterprise deployments of Claude Code, administrators can configure and distribute settings to their organization through the [Claude.ai admin console](https://claude.ai/admin-settings/claude-code). These settings are fetched automatically when users authenticate and cannot be overridden locally. This feature is available to Claude for Enterprise customers. If you don't see this option in your admin console, contact your Anthropic account team to have the feature enabled.

~~161~~

162For organizations that prefer file-based policy distribution, Claude Code also supports `managed-settings.json` files that can be deployed to [system directories](/en/settings#settings-files). These policy files follow the same format as regular settings files and cannot be overridden by user or project settings.

~~163~~

164### Settings precedence

~~165~~

166When multiple settings sources exist, they are applied in the following order (highest to lowest precedence):

~~167~~

1681. Managed settings (via Claude.ai admin console)

1692. File-based managed settings (`managed-settings.json`)

1703. Command line arguments

1714. Local project settings (`.claude/settings.local.json`)

1725. Shared project settings (`.claude/settings.json`)

1736. User settings (`~/.claude/settings.json`)

~~174~~

175This hierarchy ensures that organizational policies are always enforced while still allowing flexibility at the project and user levels where appropriate.

~~176~~

177## Credential management

~~178~~

179Claude Code securely manages your authentication credentials:

~~180~~

181* **Storage location**: On macOS, API keys, OAuth tokens, and other credentials are stored in the encrypted macOS Keychain.

182* **Supported authentication types**: Claude.ai credentials, Claude API credentials, Azure Auth, Bedrock Auth, and Vertex Auth.

183* **Custom credential scripts**: The [`apiKeyHelper`](/en/settings#available-settings) setting can be configured to run a shell script that returns an API key.

184* **Refresh intervals**: By default, `apiKeyHelper` is called after 5 minutes or on HTTP 401 response. Set `CLAUDE_CODE_API_KEY_HELPER_TTL_MS` environment variable for custom refresh intervals.

~~185~~

~~186~~

~~187~~

188> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

interactive-mode.md +171 −17

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Interactive mode5# Interactive mode

2 6

3> Complete reference for keyboard shortcuts, input modes, and interactive features in Claude Code sessions.7> Complete reference for keyboard shortcuts, input modes, and interactive features in Claude Code sessions.

5## Keyboard shortcuts9## Keyboard shortcuts

6 10

7<Note>11<Note>

8 Keyboard shortcuts may vary by platform and terminal. Press `?` to see available shortcuts for your environment. For example, Option key combinations on macOS may require configuring your terminal to use Option as a meta/escape key.12 Keyboard shortcuts may vary by platform and terminal. Press `?` to see available shortcuts for your environment.

14 **macOS users**: Option/Alt key shortcuts (`Alt+B`, `Alt+F`, `Alt+Y`, `Alt+M`, `Alt+P`) require configuring Option as Meta in your terminal:

16 * **iTerm2**: Settings → Profiles → Keys → Set Left/Right Option key to "Esc+"

17 * **Terminal.app**: Settings → Profiles → Keyboard → Check "Use Option as Meta Key"

18 * **VS Code**: Settings → Profiles → Keys → Set Left/Right Option key to "Esc+"

20 See [Terminal configuration](/en/terminal-config) for details.

9</Note>21</Note>

10 22

11### General controls23### General controls

12 24

~~14| :-------------------------------------------- | :--------------------------------- | :---------------------------------------------------------- |~~26| :------------------------------------------------ | :--------------------------------- | :--------------------------------------------------------------------------------------------------------------------------------------------------- |

29| `Ctrl+G` | Open in default text editor | Edit your prompt or custom response in your default text editor |

34| `Ctrl+B` | Background running tasks | Backgrounds bash commands and agents. Tmux users press twice |

35| `Ctrl+T` | Toggle task list | Show or hide the [task list](#task-list) in the terminal status area |

36| `Left/Right arrows` | Cycle through dialog tabs | Navigate between tabs in permission dialogs and menus |

~~23| `Shift+Tab` or `Alt+M` (some configurations) | Toggle permission modes | Switch between Auto-Accept Mode, Plan Mode, and normal mode |~~39| `Shift+Tab` or `Alt+M` (some configurations) | Toggle permission modes | Switch between Auto-Accept Mode, Plan Mode, and normal mode. When an [agent team](/en/agent-teams) is active, the cycle also includes Delegate Mode. |

41| `Option+T` (macOS) or `Alt+T` (Windows/Linux) | Toggle extended thinking | Enable or disable extended thinking mode. Run `/terminal-setup` first to enable this shortcut |

43### Text editing

45| Shortcut | Description | Context |

46| :----------------------- | :--------------------------- | :------------------------------------------------------------------------------------------------------------ |

47| `Ctrl+K` | Delete to end of line | Stores deleted text for pasting |

48| `Ctrl+U` | Delete entire line | Stores deleted text for pasting |

49| `Ctrl+Y` | Paste deleted text | Paste text deleted with `Ctrl+K` or `Ctrl+U` |

50| `Alt+Y` (after `Ctrl+Y`) | Cycle paste history | After pasting, cycle through previously deleted text. Requires [Option as Meta](#keyboard-shortcuts) on macOS |

51| `Alt+B` | Move cursor back one word | Word navigation. Requires [Option as Meta](#keyboard-shortcuts) on macOS |

52| `Alt+F` | Move cursor forward one word | Word navigation. Requires [Option as Meta](#keyboard-shortcuts) on macOS |

54### Theme and display

56| Shortcut | Description | Context |

57| :------- | :----------------------------------------- | :----------------------------------------------------------------------------------------------------------- |

58| `Ctrl+T` | Toggle syntax highlighting for code blocks | Only works inside the `/theme` picker menu. Controls whether code in Claude's responses uses syntax coloring |

60<Note>

61 Syntax highlighting is only available in the native build of Claude Code.

62</Note>

25 63

26### Multiline input64### Multiline input

27 65

~~29| :--------------- | :------------- | :-------------------------------- |~~67| :--------------- | :------------- | :------------------------------------------------------ |

35 73

36<Tip>74<Tip>

~~37 Configure your preferred line break behavior in terminal settings. Run `/terminal-setup` to install Shift+Enter binding for iTerm2 and VS Code terminals.~~75 Shift+Enter works without configuration in iTerm2, WezTerm, Ghostty, and Kitty. For other terminals (VS Code, Alacritty, Zed, Warp), run `/terminal-setup` to install the binding.

38</Tip>76</Tip>

39 77

40### Quick commands78### Quick commands

41 79

~~43| :----------- | :---------------- | :------------------------------------------------------------ |~~81| :----------- | :---------------- | :------------------------------------------------------------------- |

46| `@` | File path mention | Trigger file path autocomplete |84| `@` | File path mention | Trigger file path autocomplete |

47 85

86## Built-in commands

88Built-in commands are shortcuts for common actions. The table below covers commonly used commands but not all available options. Type `/` in Claude Code to see the full list, or type `/` followed by any letters to filter.

90To create your own commands you can invoke with `/`, see [skills](/en/skills).

92| Command | Purpose |

93| :------------------------ | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

94| `/clear` | Clear conversation history |

95| `/compact [instructions]` | Compact conversation with optional focus instructions |

96| `/config` | Open the Settings interface (Config tab) |

97| `/context` | Visualize current context usage as a colored grid |

98| `/cost` | Show token usage statistics. See [cost tracking guide](/en/costs#using-the-cost-command) for subscription-specific details. |

99| `/debug [description]` | Troubleshoot the current session by reading the session debug log. Optionally describe the issue |

100| `/doctor` | Checks the health of your Claude Code installation |

101| `/exit` | Exit the REPL |

102| `/export [filename]` | Export the current conversation to a file or clipboard |

103| `/help` | Get usage help |

104| `/init` | Initialize project with `CLAUDE.md` guide |

105| `/mcp` | Manage MCP server connections and OAuth authentication |

106| `/memory` | Edit `CLAUDE.md` memory files |

107| `/model` | Select or change the AI model. With Opus 4.6, use left/right arrows to [adjust effort level](/en/model-config#adjust-effort-level). The change takes effect immediately without waiting for the current response to finish |

108| `/permissions` | View or update [permissions](/en/permissions#manage-permissions) |

109| `/plan` | Enter plan mode directly from the prompt |

110| `/rename <name>` | Rename the current session for easier identification |

111| `/resume [session]` | Resume a conversation by ID or name, or open the session picker |

112| `/rewind` | Rewind the conversation and/or code, or summarize from a selected message |

113| `/stats` | Visualize daily usage, session history, streaks, and model preferences |

114| `/status` | Open the Settings interface (Status tab) showing version, model, account, and connectivity |

115| `/statusline` | Set up Claude Code's status line UI |

116| `/copy` | Copy the last assistant response to clipboard |

117| `/tasks` | List and manage background tasks |

118| `/teleport` | Resume a remote session from claude.ai (subscribers only) |

119| `/theme` | Change the color theme |

120| `/todos` | List current TODO items |

121| `/usage` | For subscription plans only: show plan usage limits and rate limit status |

122

123### MCP prompts

124

125MCP servers can expose prompts that appear as commands. These use the format `/mcp__<server>__<prompt>` and are dynamically discovered from connected servers. See [MCP prompts](/en/mcp#use-mcp-prompts-as-commands) for details.

126

48## Vim editor mode127## Vim editor mode

49 128

50Enable vim-style editing with `/vim` command or configure permanently via `/config`.129Enable vim-style editing with `/vim` command or configure permanently via `/config`.

64### Navigation (NORMAL mode)143### Navigation (NORMAL mode)

65 144

66| Command | Action |145| Command | Action |

~~67| :-------------- | :------------------------ |~~146| :-------------- | :-------------------------------------------------- |

69| `w` | Next word |148| `w` | Next word |

70| `e` | End of word |149| `e` | End of word |

74| `^` | First non-blank character |153| `^` | First non-blank character |

75| `gg` | Beginning of input |154| `gg` | Beginning of input |

76| `G` | End of input |155| `G` | End of input |

156| `f{char}` | Jump to next occurrence of character |

157| `F{char}` | Jump to previous occurrence of character |

158| `t{char}` | Jump to just before next occurrence of character |

159| `T{char}` | Jump to just after previous occurrence of character |

160| `;` | Repeat last f/F/t/T motion |

161| `,` | Repeat last f/F/t/T motion in reverse |

162

163<Note>

164 In vim normal mode, if the cursor is at the beginning or end of input and cannot move further, the arrow keys navigate command history instead.

165</Note>

77 166

78### Editing (NORMAL mode)167### Editing (NORMAL mode)

79 168

86| `cc` | Change line |175| `cc` | Change line |

87| `C` | Change to end of line |176| `C` | Change to end of line |

178| `yy`/`Y` | Yank (copy) line |

179| `yw`/`ye`/`yb` | Yank word/to end/back |

180| `p` | Paste after cursor |

181| `P` | Paste before cursor |

182| `>>` | Indent line |

183| `<<` | Dedent line |

184| `J` | Join lines |

89| `.` | Repeat last change |185| `.` | Repeat last change |

90 186

187### Text objects (NORMAL mode)

188

189Text objects work with operators like `d`, `c`, and `y`:

190

191| Command | Action |

192| :-------- | :--------------------------------------- |

193| `iw`/`aw` | Inner/around word |

194| `iW`/`aW` | Inner/around WORD (whitespace-delimited) |

195| `i"`/`a"` | Inner/around double quotes |

196| `i'`/`a'` | Inner/around single quotes |

197| `i(`/`a(` | Inner/around parentheses |

198| `i[`/`a[` | Inner/around brackets |

199| `i{`/`a{` | Inner/around braces |

200

91## Command history201## Command history

92 202

93Claude Code maintains command history for the current session:203Claude Code maintains command history for the current session:

128 238

129**Key features:**239**Key features:**

130 240

131* Output is buffered and Claude can retrieve it using the BashOutput tool241* Output is buffered and Claude can retrieve it using the TaskOutput tool

132* Background tasks have unique IDs for tracking and output retrieval242* Background tasks have unique IDs for tracking and output retrieval

133* Background tasks are automatically cleaned up when Claude Code exits243* Background tasks are automatically cleaned up when Claude Code exits

134 244

245To disable all background task functionality, set the `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to `1`. See [Environment variables](/en/settings#environment-variables) for details.

246

135**Common backgrounded commands:**247**Common backgrounded commands:**

136 248

137* Build tools (webpack, vite, make)249* Build tools (webpack, vite, make)

156* Shows real-time progress and output268* Shows real-time progress and output

157* Supports the same `Ctrl+B` backgrounding for long-running commands269* Supports the same `Ctrl+B` backgrounding for long-running commands

158* Does not require Claude to interpret or approve the command270* Does not require Claude to interpret or approve the command

271* Supports history-based autocomplete: type a partial command and press **Tab** to complete from previous `!` commands in the current project

159 272

160This is useful for quick shell operations while maintaining conversation context.273This is useful for quick shell operations while maintaining conversation context.

161 274

275## Prompt suggestions

276

277When you first open a session, a grayed-out example command appears in the prompt input to help you get started. Claude Code picks this from your project's git history, so it reflects files you've been working on recently.

278

279After Claude responds, suggestions continue to appear based on your conversation history, such as a follow-up step from a multi-part request or a natural continuation of your workflow.

280

281* Press **Tab** to accept the suggestion, or press **Enter** to accept and submit

282* Start typing to dismiss it

283

284The suggestion runs as a background request that reuses the parent conversation's prompt cache, so the additional cost is minimal. Claude Code skips suggestion generation when the cache is cold to avoid unnecessary cost.

285

286Suggestions are automatically skipped after the first turn of a conversation, in non-interactive mode, and in plan mode.

287

288To disable prompt suggestions entirely, set the environment variable or toggle the setting in `/config`:

289

290```bash theme={null}

291export CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false

292```

293

294## Task list

295

296When working on complex, multi-step work, Claude creates a task list to track progress. Tasks appear in the status area of your terminal with indicators showing what's pending, in progress, or complete.

297

298* Press `Ctrl+T` to toggle the task list view. The display shows up to 10 tasks at a time

299* To see all tasks or clear them, ask Claude directly: "show me all tasks" or "clear all tasks"

300* Tasks persist across context compactions, helping Claude stay organized on larger projects

301* To share a task list across sessions, set `CLAUDE_CODE_TASK_LIST_ID` to use a named directory in `~/.claude/tasks/`: `CLAUDE_CODE_TASK_LIST_ID=my-project claude`

302* To revert to the previous TODO list, set `CLAUDE_CODE_ENABLE_TASKS=false`.

303

304## PR review status

305

306When working on a branch with an open pull request, Claude Code displays a clickable PR link in the footer (for example, "PR #446"). The link has a colored underline indicating the review state:

307

308* Green: approved

309* Yellow: pending review

310* Red: changes requested

311* Gray: draft

312* Purple: merged

313

314`Cmd+click` (Mac) or `Ctrl+click` (Windows/Linux) the link to open the pull request in your browser. The status updates automatically every 60 seconds.

315

316<Note>

317 PR status requires the `gh` CLI to be installed and authenticated (`gh auth login`).

318</Note>

319

162## See also320## See also

163 321

164* [Slash commands](/en/slash-commands) - Interactive session commands322* [Skills](/en/skills) - Custom prompts and workflows

165* [Checkpointing](/en/checkpointing) - Rewind Claude's edits and restore previous states323* [Checkpointing](/en/checkpointing) - Rewind Claude's edits and restore previous states

166* [CLI reference](/en/cli-reference) - Command-line flags and options324* [CLI reference](/en/cli-reference) - Command-line flags and options

167* [Settings](/en/settings) - Configuration options325* [Settings](/en/settings) - Configuration options

168* [Memory management](/en/memory) - Managing CLAUDE.md files326* [Memory management](/en/memory) - Managing CLAUDE.md files

~~169~~

~~170~~

~~171~~

172> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

jetbrains.md +4 −4

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# JetBrains IDEs5# JetBrains IDEs

2 6

3> Use Claude Code with JetBrains IDEs including IntelliJ, PyCharm, WebStorm, and more7> Use Claude Code with JetBrains IDEs including IntelliJ, PyCharm, WebStorm, and more

146* Being aware of which files Claude Code has access to modify150* Being aware of which files Claude Code has access to modify

147 151

148For additional help, see our [troubleshooting guide](/en/troubleshooting).152For additional help, see our [troubleshooting guide](/en/troubleshooting).

~~149~~

~~150~~

~~151~~

152> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

keybindings.md +381 −0 added

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# Customize keyboard shortcuts

7> Customize keyboard shortcuts in Claude Code with a keybindings configuration file.

9Claude Code supports customizable keyboard shortcuts. Run `/keybindings` to create or open your configuration file at `~/.claude/keybindings.json`.

11## Configuration file

13The keybindings configuration file is an object with a `bindings` array. Each block specifies a context and a map of keystrokes to actions.

15<Note>Changes to the keybindings file are automatically detected and applied without restarting Claude Code.</Note>

17| Field | Description |

18| :--------- | :------------------------------------------------- |

19| `$schema` | Optional JSON Schema URL for editor autocompletion |

20| `$docs` | Optional documentation URL |

21| `bindings` | Array of binding blocks by context |

23This example binds `Ctrl+E` to open an external editor in the chat context, and unbinds `Ctrl+U`:

25```json theme={null}

26{

27 "$schema": "https://www.schemastore.org/claude-code-keybindings.json",

28 "$docs": "https://code.claude.com/docs/en/keybindings",

29 "bindings": [

30 {

31 "context": "Chat",

32 "bindings": {

33 "ctrl+e": "chat:externalEditor",

34 "ctrl+u": null

35 }

36 }

37 ]

38}

39```

41## Contexts

43Each binding block specifies a **context** where the bindings apply:

45| Context | Description |

46| :---------------- | :----------------------------------------------- |

47| `Global` | Applies everywhere in the app |

48| `Chat` | Main chat input area |

49| `Autocomplete` | Autocomplete menu is open |

50| `Settings` | Settings menu (escape-only dismiss) |

51| `Confirmation` | Permission and confirmation dialogs |

52| `Tabs` | Tab navigation components |

53| `Help` | Help menu is visible |

54| `Transcript` | Transcript viewer |

55| `HistorySearch` | History search mode (Ctrl+R) |

56| `Task` | Background task is running |

57| `ThemePicker` | Theme picker dialog |

58| `Attachments` | Image/attachment bar navigation |

59| `Footer` | Footer indicator navigation (tasks, teams, diff) |

60| `MessageSelector` | Rewind and summarize dialog message selection |

61| `DiffDialog` | Diff viewer navigation |

62| `ModelPicker` | Model picker effort level |

63| `Select` | Generic select/list components |

64| `Plugin` | Plugin dialog (browse, discover, manage) |

66## Available actions

68Actions follow a `namespace:action` format, such as `chat:submit` to send a message or `app:toggleTodos` to show the task list. Each context has specific actions available.

70### App actions

72Actions available in the `Global` context:

74| Action | Default | Description |

75| :--------------------- | :------ | :-------------------------- |

76| `app:interrupt` | Ctrl+C | Cancel current operation |

77| `app:exit` | Ctrl+D | Exit Claude Code |

78| `app:toggleTodos` | Ctrl+T | Toggle task list visibility |

79| `app:toggleTranscript` | Ctrl+O | Toggle verbose transcript |

81### History actions

83Actions for navigating command history:

85| Action | Default | Description |

86| :----------------- | :------ | :-------------------- |

87| `history:search` | Ctrl+R | Open history search |

88| `history:previous` | Up | Previous history item |

89| `history:next` | Down | Next history item |

91### Chat actions

93Actions available in the `Chat` context:

95| Action | Default | Description |

96| :-------------------- | :------------------------ | :----------------------- |

97| `chat:cancel` | Escape | Cancel current input |

98| `chat:cycleMode` | Shift+Tab\* | Cycle permission modes |

99| `chat:modelPicker` | Cmd+P / Meta+P | Open model picker |

100| `chat:thinkingToggle` | Cmd+T / Meta+T | Toggle extended thinking |

101| `chat:submit` | Enter | Submit message |

102| `chat:undo` | Ctrl+\_ | Undo last action |

103| `chat:externalEditor` | Ctrl+G | Open in external editor |

104| `chat:stash` | Ctrl+S | Stash current prompt |

105| `chat:imagePaste` | Ctrl+V (Alt+V on Windows) | Paste image |

106

107\*On Windows without VT mode (Node \<24.2.0/\<22.17.0, Bun \<1.2.23), defaults to Meta+M.

108

109### Autocomplete actions

110

111Actions available in the `Autocomplete` context:

112

113| Action | Default | Description |

114| :---------------------- | :------ | :------------------ |

115| `autocomplete:accept` | Tab | Accept suggestion |

116| `autocomplete:dismiss` | Escape | Dismiss menu |

117| `autocomplete:previous` | Up | Previous suggestion |

118| `autocomplete:next` | Down | Next suggestion |

119

120### Confirmation actions

121

122Actions available in the `Confirmation` context:

123

124| Action | Default | Description |

125| :-------------------------- | :-------- | :---------------------------- |

126| `confirm:yes` | Y, Enter | Confirm action |

127| `confirm:no` | N, Escape | Decline action |

128| `confirm:previous` | Up | Previous option |

129| `confirm:next` | Down | Next option |

130| `confirm:nextField` | Tab | Next field |

131| `confirm:previousField` | (unbound) | Previous field |

132| `confirm:cycleMode` | Shift+Tab | Cycle permission modes |

133| `confirm:toggleExplanation` | Ctrl+E | Toggle permission explanation |

134

135### Permission actions

136

137Actions available in the `Confirmation` context for permission dialogs:

138

139| Action | Default | Description |

140| :----------------------- | :------ | :--------------------------- |

141| `permission:toggleDebug` | Ctrl+D | Toggle permission debug info |

142

143### Transcript actions

144

145Actions available in the `Transcript` context:

146

147| Action | Default | Description |

148| :------------------------- | :------------- | :---------------------- |

149| `transcript:toggleShowAll` | Ctrl+E | Toggle show all content |

150| `transcript:exit` | Ctrl+C, Escape | Exit transcript view |

151

152### History search actions

153

154Actions available in the `HistorySearch` context:

155

156| Action | Default | Description |

157| :---------------------- | :---------- | :----------------------- |

158| `historySearch:next` | Ctrl+R | Next match |

159| `historySearch:accept` | Escape, Tab | Accept selection |

160| `historySearch:cancel` | Ctrl+C | Cancel search |

161| `historySearch:execute` | Enter | Execute selected command |

162

163### Task actions

164

165Actions available in the `Task` context:

166

167| Action | Default | Description |

168| :---------------- | :------ | :---------------------- |

169| `task:background` | Ctrl+B | Background current task |

170

171### Theme actions

172

173Actions available in the `ThemePicker` context:

174

175| Action | Default | Description |

176| :------------------------------- | :------ | :------------------------- |

177| `theme:toggleSyntaxHighlighting` | Ctrl+T | Toggle syntax highlighting |

178

179### Help actions

180

181Actions available in the `Help` context:

182

183| Action | Default | Description |

184| :------------- | :------ | :-------------- |

185| `help:dismiss` | Escape | Close help menu |

186

187### Tabs actions

188

189Actions available in the `Tabs` context:

190

191| Action | Default | Description |

192| :-------------- | :-------------- | :----------- |

193| `tabs:next` | Tab, Right | Next tab |

194| `tabs:previous` | Shift+Tab, Left | Previous tab |

195

196### Attachments actions

197

198Actions available in the `Attachments` context:

199

200| Action | Default | Description |

201| :--------------------- | :---------------- | :------------------------- |

202| `attachments:next` | Right | Next attachment |

203| `attachments:previous` | Left | Previous attachment |

204| `attachments:remove` | Backspace, Delete | Remove selected attachment |

205| `attachments:exit` | Down, Escape | Exit attachment bar |

206

207### Footer actions

208

209Actions available in the `Footer` context:

210

211| Action | Default | Description |

212| :---------------------- | :------ | :------------------------ |

213| `footer:next` | Right | Next footer item |

214| `footer:previous` | Left | Previous footer item |

215| `footer:openSelected` | Enter | Open selected footer item |

216| `footer:clearSelection` | Escape | Clear footer selection |

217

218### Message selector actions

219

220Actions available in the `MessageSelector` context:

221

222| Action | Default | Description |

223| :----------------------- | :---------------------------------------- | :---------------- |

224| `messageSelector:up` | Up, K | Move up in list |

225| `messageSelector:down` | Down, J | Move down in list |

226| `messageSelector:top` | Ctrl+Up, Shift+Up, Meta+Up, Shift+K | Jump to top |

227| `messageSelector:bottom` | Ctrl+Down, Shift+Down, Meta+Down, Shift+J | Jump to bottom |

228| `messageSelector:select` | Enter | Select message |

229

230### Diff actions

231

232Actions available in the `DiffDialog` context:

233

234| Action | Default | Description |

235| :-------------------- | :----------------- | :--------------------- |

236| `diff:dismiss` | Escape | Close diff viewer |

237| `diff:previousSource` | Left | Previous diff source |

238| `diff:nextSource` | Right | Next diff source |

239| `diff:previousFile` | Up | Previous file in diff |

240| `diff:nextFile` | Down | Next file in diff |

241| `diff:viewDetails` | Enter | View diff details |

242| `diff:back` | (context-specific) | Go back in diff viewer |

243

244### Model picker actions

245

246Actions available in the `ModelPicker` context:

247

248| Action | Default | Description |

249| :--------------------------- | :------ | :-------------------- |

250| `modelPicker:decreaseEffort` | Left | Decrease effort level |

251| `modelPicker:increaseEffort` | Right | Increase effort level |

252

253### Select actions

254

255Actions available in the `Select` context:

256

257| Action | Default | Description |

258| :---------------- | :-------------- | :--------------- |

259| `select:next` | Down, J, Ctrl+N | Next option |

260| `select:previous` | Up, K, Ctrl+P | Previous option |

261| `select:accept` | Enter | Accept selection |

262| `select:cancel` | Escape | Cancel selection |

263

264### Plugin actions

265

266Actions available in the `Plugin` context:

267

268| Action | Default | Description |

269| :--------------- | :------ | :----------------------- |

270| `plugin:toggle` | Space | Toggle plugin selection |

271| `plugin:install` | I | Install selected plugins |

272

273### Settings actions

274

275Actions available in the `Settings` context:

276

277| Action | Default | Description |

278| :---------------- | :------ | :---------------------------------- |

279| `settings:search` | / | Enter search mode |

280| `settings:retry` | R | Retry loading usage data (on error) |

281

282## Keystroke syntax

283

284### Modifiers

285

286Use modifier keys with the `+` separator:

287

288* `ctrl` or `control` - Control key

289* `alt`, `opt`, or `option` - Alt/Option key

290* `shift` - Shift key

291* `meta`, `cmd`, or `command` - Meta/Command key

292

293For example:

294

295```

296ctrl+k Single key with modifier

297shift+tab Shift + Tab

298meta+p Command/Meta + P

299ctrl+shift+c Multiple modifiers

300```

301

302### Uppercase letters

303

304A standalone uppercase letter implies Shift. For example, `K` is equivalent to `shift+k`. This is useful for vim-style bindings where uppercase and lowercase keys have different meanings.

305

306Uppercase letters with modifiers (e.g., `ctrl+K`) are treated as stylistic and do **not** imply Shift — `ctrl+K` is the same as `ctrl+k`.

307

308### Chords

309

310Chords are sequences of keystrokes separated by spaces:

311

312```

313ctrl+k ctrl+s Press Ctrl+K, release, then Ctrl+S

314```

315

316### Special keys

317

318* `escape` or `esc` - Escape key

319* `enter` or `return` - Enter key

320* `tab` - Tab key

321* `space` - Space bar

322* `up`, `down`, `left`, `right` - Arrow keys

323* `backspace`, `delete` - Delete keys

324

325## Unbind default shortcuts

326

327Set an action to `null` to unbind a default shortcut:

328

329```json theme={null}

330{

331 "bindings": [

332 {

333 "context": "Chat",

334 "bindings": {

335 "ctrl+s": null

336 }

337 }

338 ]

339}

340```

341

342## Reserved shortcuts

343

344These shortcuts cannot be rebound:

345

346| Shortcut | Reason |

347| :------- | :------------------------- |

348| Ctrl+C | Hardcoded interrupt/cancel |

349| Ctrl+D | Hardcoded exit |

350

351## Terminal conflicts

352

353Some shortcuts may conflict with terminal multiplexers:

354

355| Shortcut | Conflict |

356| :------- | :-------------------------------- |

357| Ctrl+B | tmux prefix (press twice to send) |

358| Ctrl+A | GNU screen prefix |

359| Ctrl+Z | Unix process suspend (SIGTSTP) |

360

361## Vim mode interaction

362

363When vim mode is enabled (`/vim`), keybindings and vim mode operate independently:

364

365* **Vim mode** handles input at the text input level (cursor movement, modes, motions)

366* **Keybindings** handle actions at the component level (toggle todos, submit, etc.)

367* The Escape key in vim mode switches INSERT to NORMAL mode; it does not trigger `chat:cancel`

368* Most Ctrl+key shortcuts pass through vim mode to the keybinding system

369* In vim NORMAL mode, `?` shows the help menu (vim behavior)

370

371## Validation

372

373Claude Code validates your keybindings and shows warnings for:

374

375* Parse errors (invalid JSON or structure)

376* Invalid context names

377* Reserved shortcut conflicts

378* Terminal multiplexer conflicts

379* Duplicate bindings in the same context

380

381Run `/doctor` to see any keybinding warnings.

legal-and-compliance.md +4 −4

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Legal and compliance5# Legal and compliance

2 6

3> Legal agreements, compliance certifications, and security information for Claude Code.7> Legal agreements, compliance certifications, and security information for Claude Code.

34***38***

35 39

~~40> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt~~

llm-gateway.md +4 −4

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# LLM gateway configuration5# LLM gateway configuration

2 6

3> Learn how to configure Claude Code to work with LLM gateway solutions. Covers gateway requirements, authentication configuration, model selection, and provider-specific endpoint setup.7> Learn how to configure Claude Code to work with LLM gateway solutions. Covers gateway requirements, authentication configuration, model selection, and provider-specific endpoint setup.

168* [Claude Code settings](/en/settings)172* [Claude Code settings](/en/settings)

169* [Enterprise network configuration](/en/network-config)173* [Enterprise network configuration](/en/network-config)

170* [Third-party integrations overview](/en/third-party-integrations)174* [Third-party integrations overview](/en/third-party-integrations)

~~171~~

~~172~~

~~173~~

174> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

mcp.md +139 −11

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Connect Claude Code to tools via MCP5# Connect Claude Code to tools via MCP

2 6

3> Learn how to connect Claude Code to your tools with the Model Context Protocol.7> Learn how to connect Claude Code to your tools with the Model Context Protocol.

115/mcp119/mcp

116```120```

117 121

122### Dynamic tool updates

123

124Claude Code supports MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring you to disconnect and reconnect. When an MCP server sends a `list_changed` notification, Claude Code automatically refreshes the available capabilities from that server.

125

118<Tip>126<Tip>

119 Tips:127 Tips:

120 128

212 220

213Local-scoped servers represent the default configuration level and are stored in `~/.claude.json` under your project's path. These servers remain private to you and are only accessible when working within the current project directory. This scope is ideal for personal development servers, experimental configurations, or servers containing sensitive credentials that shouldn't be shared.221Local-scoped servers represent the default configuration level and are stored in `~/.claude.json` under your project's path. These servers remain private to you and are only accessible when working within the current project directory. This scope is ideal for personal development servers, experimental configurations, or servers containing sensitive credentials that shouldn't be shared.

214 222

223<Note>

224 The term "local scope" for MCP servers differs from general local settings. MCP local-scoped servers are stored in `~/.claude.json` (your home directory), while general local settings use `.claude/settings.local.json` (in the project directory). See [Settings](/en/settings#settings-files) for details on settings file locations.

225</Note>

226

215```bash theme={null}227```bash theme={null}

216# Add a local-scoped server (default)228# Add a local-scoped server (default)

217claude mcp add --transport http stripe https://mcp.stripe.com229claude mcp add --transport http stripe https://mcp.stripe.com

267 279

268 * **User and local scope**: `~/.claude.json` (in the `mcpServers` field or under project paths)280 * **User and local scope**: `~/.claude.json` (in the `mcpServers` field or under project paths)

269 * **Project scope**: `.mcp.json` in your project root (checked into source control)281 * **Project scope**: `.mcp.json` in your project root (checked into source control)

270 * **Enterprise managed**: `managed-mcp.json` in system directories (see [Enterprise MCP configuration](#enterprise-mcp-configuration))282 * **Managed**: `managed-mcp.json` in system directories (see [Managed MCP configuration](#managed-mcp-configuration))

271</Note>283</Note>

272 284

273### Scope hierarchy and precedence285### Scope hierarchy and precedence

401 * OAuth authentication works with HTTP servers413 * OAuth authentication works with HTTP servers

402</Tip>414</Tip>

403 415

416### Use pre-configured OAuth credentials

417

418Some MCP servers don't support automatic OAuth setup. If you see an error like "Incompatible auth server: does not support dynamic client registration," the server requires pre-configured credentials. Register an OAuth app through the server's developer portal first, then provide the credentials when adding the server.

419

420<Steps>

421 <Step title="Register an OAuth app with the server">

422 Create an app through the server's developer portal and note your client ID and client secret.

423

424 Many servers also require a redirect URI. If so, choose a port and register a redirect URI in the format `http://localhost:PORT/callback`. Use that same port with `--callback-port` in the next step.

425 </Step>

426

427 <Step title="Add the server with your credentials">

428 Choose one of the following methods. The port used for `--callback-port` can be any available port. It just needs to match the redirect URI you registered in the previous step.

429

430 <Tabs>

431 <Tab title="claude mcp add">

432 Use `--client-id` to pass your app's client ID. The `--client-secret` flag prompts for the secret with masked input:

433

434 ```bash theme={null}

435 claude mcp add --transport http \

436 --client-id your-client-id --client-secret --callback-port 8080 \

437 my-server https://mcp.example.com/mcp

438 ```

439 </Tab>

440

441 <Tab title="claude mcp add-json">

442 Include the `oauth` object in the JSON config and pass `--client-secret` as a separate flag:

443

444 ```bash theme={null}

445 claude mcp add-json my-server \

446 '{"type":"http","url":"https://mcp.example.com/mcp","oauth":{"clientId":"your-client-id","callbackPort":8080}}' \

447 --client-secret

448 ```

449 </Tab>

450

451 <Tab title="CI / env var">

452 Set the secret via environment variable to skip the interactive prompt:

453

454 ```bash theme={null}

455 MCP_CLIENT_SECRET=your-secret claude mcp add --transport http \

456 --client-id your-client-id --client-secret --callback-port 8080 \

457 my-server https://mcp.example.com/mcp

458 ```

459 </Tab>

460 </Tabs>

461 </Step>

462

463 <Step title="Authenticate in Claude Code">

464 Run `/mcp` in Claude Code and follow the browser login flow.

465 </Step>

466</Steps>

467

468<Tip>

469 Tips:

470

471 * The client secret is stored securely in your system keychain (macOS) or a credentials file, not in your config

472 * If the server uses a public OAuth client with no secret, use only `--client-id` without `--client-secret`

473 * These flags only apply to HTTP and SSE transports. They have no effect on stdio servers

474 * Use `claude mcp get <name>` to verify that OAuth credentials are configured for a server

475</Tip>

476

404## Add MCP servers from JSON configuration477## Add MCP servers from JSON configuration

405 478

406If you have a JSON configuration for an MCP server, you can add it directly:479If you have a JSON configuration for an MCP server, you can add it directly:

416 489

417 # Example: Adding a stdio server with JSON configuration490 # Example: Adding a stdio server with JSON configuration

418 claude mcp add-json local-weather '{"type":"stdio","command":"/path/to/weather-cli","args":["--api-key","abc123"],"env":{"CACHE_DIR":"/tmp"}}'491 claude mcp add-json local-weather '{"type":"stdio","command":"/path/to/weather-cli","args":["--api-key","abc123"],"env":{"CACHE_DIR":"/tmp"}}'

492

493 # Example: Adding an HTTP server with pre-configured OAuth credentials

494 claude mcp add-json my-server '{"type":"http","url":"https://mcp.example.com/mcp","oauth":{"clientId":"your-client-id","callbackPort":8080}}' --client-secret

419 ```495 ```

420 </Step>496 </Step>

421 497

593 * Resources can contain any type of content that the MCP server provides (text, JSON, structured data, etc.)669 * Resources can contain any type of content that the MCP server provides (text, JSON, structured data, etc.)

594</Tip>670</Tip>

595 671

596## Use MCP prompts as slash commands672## Scale with MCP Tool Search

673

674When you have many MCP servers configured, tool definitions can consume a significant portion of your context window. MCP Tool Search solves this by dynamically loading tools on-demand instead of preloading all of them.

675

676### How it works

677

678Claude Code automatically enables Tool Search when your MCP tool descriptions would consume more than 10% of the context window. You can [adjust this threshold](#configure-tool-search) or disable tool search entirely. When triggered:

679

6801. MCP tools are deferred rather than loaded into context upfront

6812. Claude uses a search tool to discover relevant MCP tools when needed

6823. Only the tools Claude actually needs are loaded into context

6834. MCP tools continue to work exactly as before from your perspective

684

685### For MCP server authors

597 686

598MCP servers can expose prompts that become available as slash commands in Claude Code.687If you're building an MCP server, the server instructions field becomes more useful with Tool Search enabled. Server instructions help Claude understand when to search for your tools, similar to how [skills](/en/skills) work.

688

689Add clear, descriptive server instructions that explain:

690

691* What category of tasks your tools handle

692* When Claude should search for your tools

693* Key capabilities your server provides

694

695### Configure tool search

696

697Tool search runs in auto mode by default, meaning it activates only when your MCP tool definitions exceed the context threshold. If you have few tools, they load normally without tool search. This feature requires models that support `tool_reference` blocks: Sonnet 4 and later, or Opus 4 and later. Haiku models do not support tool search.

698

699Control tool search behavior with the `ENABLE_TOOL_SEARCH` environment variable:

700

701| Value | Behavior |

702| :--------- | :--------------------------------------------------------------------------------- |

703| `auto` | Activates when MCP tools exceed 10% of context (default) |

704| `auto:<N>` | Activates at custom threshold, where `<N>` is a percentage (e.g., `auto:5` for 5%) |

705| `true` | Always enabled |

706| `false` | Disabled, all MCP tools loaded upfront |

707

708```bash theme={null}

709# Use a custom 5% threshold

710ENABLE_TOOL_SEARCH=auto:5 claude

711

712# Disable tool search entirely

713ENABLE_TOOL_SEARCH=false claude

714```

715

716Or set the value in your [settings.json `env` field](/en/settings#available-settings).

717

718You can also disable the MCPSearch tool specifically using the `disallowedTools` setting:

719

720```json theme={null}

721{

722 "permissions": {

723 "deny": ["MCPSearch"]

724 }

725}

726```

727

728## Use MCP prompts as commands

729

730MCP servers can expose prompts that become available as commands in Claude Code.

599 731

600### Execute MCP prompts732### Execute MCP prompts

601 733

632 * Server and prompt names are normalized (spaces become underscores)764 * Server and prompt names are normalized (spaces become underscores)

633</Tip>765</Tip>

634 766

635## Enterprise MCP configuration767## Managed MCP configuration

636 768

637For organizations that need centralized control over MCP servers, Claude Code supports two enterprise configuration options:769For organizations that need centralized control over MCP servers, Claude Code supports two configuration options:

638 770

6391. **Exclusive control with `managed-mcp.json`**: Deploy a fixed set of MCP servers that users cannot modify or extend7711. **Exclusive control with `managed-mcp.json`**: Deploy a fixed set of MCP servers that users cannot modify or extend

6402. **Policy-based control with allowlists/denylists**: Allow users to add their own servers, but restrict which ones are permitted7722. **Policy-based control with allowlists/denylists**: Allow users to add their own servers, but restrict which ones are permitted

852 984

853#### Important notes985#### Important notes

854 986

855* **Option 1 and Option 2 can be combined**: If `managed-mcp.json` exists, it has exclusive control and users cannot add servers. Allowlists/denylists still apply to the enterprise servers themselves.987* **Option 1 and Option 2 can be combined**: If `managed-mcp.json` exists, it has exclusive control and users cannot add servers. Allowlists/denylists still apply to the managed servers themselves.

856* **Denylist takes absolute precedence**: If a server matches a denylist entry (by name, command, or URL), it will be blocked even if it's on the allowlist988* **Denylist takes absolute precedence**: If a server matches a denylist entry (by name, command, or URL), it will be blocked even if it's on the allowlist

857* Name-based, command-based, and URL-based restrictions work together: a server passes if it matches **either** a name entry, a command entry, or a URL pattern (unless blocked by denylist)989* Name-based, command-based, and URL-based restrictions work together: a server passes if it matches **either** a name entry, a command entry, or a URL pattern (unless blocked by denylist)

858 990

859<Note>991<Note>

860 **When using `managed-mcp.json`**: Users cannot add MCP servers through `claude mcp add` or configuration files. The `allowedMcpServers` and `deniedMcpServers` settings still apply to filter which enterprise servers are actually loaded.992 **When using `managed-mcp.json`**: Users cannot add MCP servers through `claude mcp add` or configuration files. The `allowedMcpServers` and `deniedMcpServers` settings still apply to filter which managed servers are actually loaded.

861</Note>993</Note>

~~862~~

~~863~~

~~864~~

865> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

memory.md +101 −19

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Manage Claude's memory5# Manage Claude's memory

2 6

3> Learn how to manage Claude Code's memory across sessions with different memory locations and best practices.7> Learn how to manage Claude Code's memory across sessions with different memory locations and best practices.

4 8

~~5Claude Code can remember your preferences across sessions, like style guidelines and common commands in your workflow.~~9Claude Code has two kinds of memory that persist across sessions:

11* **Auto memory**: Claude automatically saves useful context like project patterns, key commands, and your preferences. This persists across sessions.

12* **CLAUDE.md files**: Markdown files you write and maintain with instructions, rules, and preferences for Claude to follow.

14Both are loaded into Claude's context at the start of every session, though auto memory loads only the first 200 lines of its main file.

6 15

7## Determine memory type16## Determine memory type

8 17

~~9Claude Code offers four memory locations in a hierarchical structure, each serving a different purpose:~~18Claude Code offers several memory locations in a hierarchical structure, each serving a different purpose:

10 19

12| -------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------- | -------------------------------------------------------------------- | ------------------------------- |21| -------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------- | -------------------------------------------------------------------- | ------------------------------- |

13| **Enterprise policy** | • macOS: `/Library/Application Support/ClaudeCode/CLAUDE.md`<br />• Linux: `/etc/claude-code/CLAUDE.md`<br />• Windows: `C:\Program Files\ClaudeCode\CLAUDE.md` | Organization-wide instructions managed by IT/DevOps | Company coding standards, security policies, compliance requirements | All users in organization |22| **Managed policy** | • macOS: `/Library/Application Support/ClaudeCode/CLAUDE.md`<br />• Linux: `/etc/claude-code/CLAUDE.md`<br />• Windows: `C:\Program Files\ClaudeCode\CLAUDE.md` | Organization-wide instructions managed by IT/DevOps | Company coding standards, security policies, compliance requirements | All users in organization |

15| **Project rules** | `./.claude/rules/*.md` | Modular, topic-specific project instructions | Language-specific guidelines, testing conventions, API standards | Team members via source control |24| **Project rules** | `./.claude/rules/*.md` | Modular, topic-specific project instructions | Language-specific guidelines, testing conventions, API standards | Team members via source control |

18 28

19All memory files are automatically loaded into Claude Code's context when launched. Files higher in the hierarchy take precedence and are loaded first, providing a foundation that more specific memories build upon.29CLAUDE.md files in the directory hierarchy above the working directory are loaded in full at launch. CLAUDE.md files in child directories load on demand when Claude reads files in those directories. Auto memory loads only the first 200 lines of `MEMORY.md`. More specific instructions take precedence over broader ones.

20 30

21<Note>31<Note>

22 CLAUDE.local.md files are automatically added to .gitignore, making them ideal for private project-specific preferences that shouldn't be checked into version control.32 CLAUDE.local.md files are automatically added to .gitignore, making them ideal for private project-specific preferences that shouldn't be checked into version control.

23</Note>33</Note>

24 34

35## Auto memory

37Auto memory is a persistent directory where Claude records learnings, patterns, and insights as it works. Unlike CLAUDE.md files that contain instructions you write for Claude, auto memory contains notes Claude writes for itself based on what it discovers during sessions.

39<Note>

40 Auto memory is being rolled out gradually. If you aren't seeing auto memory, you can opt in by setting `CLAUDE_CODE_DISABLE_AUTO_MEMORY=0` in your environment.

41</Note>

43### What Claude remembers

45As Claude works, it may save things like:

47* Project patterns: build commands, test conventions, code style preferences

48* Debugging insights: solutions to tricky problems, common error causes

49* Architecture notes: key files, module relationships, important abstractions

50* Your preferences: communication style, workflow habits, tool choices

52### Where auto memory is stored

54Each project gets its own memory directory at `~/.claude/projects/<project>/memory/`. The `<project>` path is derived from the git repository root, so all subdirectories within the same repo share one auto memory directory. Git worktrees get separate memory directories. Outside a git repo, the working directory is used instead.

56The directory contains a `MEMORY.md` entrypoint and optional topic files:

58```text theme={null}

59~/.claude/projects/<project>/memory/

60├── MEMORY.md # Concise index, loaded into every session

61├── debugging.md # Detailed notes on debugging patterns

62├── api-conventions.md # API design decisions

63└── ... # Any other topic files Claude creates

64```

66`MEMORY.md` acts as an index of the memory directory. Claude reads and writes files in this directory throughout your session, using `MEMORY.md` to keep track of what's stored where.

68### How it works

70* The first 200 lines of `MEMORY.md` are loaded into Claude's system prompt at the start of every session. Content beyond 200 lines is not loaded automatically, and Claude is instructed to keep it concise by moving detailed notes into separate topic files.

71* Topic files like `debugging.md` or `patterns.md` are not loaded at startup. Claude reads them on demand using its standard file tools when it needs the information.

72* Claude reads and writes memory files during your session, so you'll see memory updates happen as you work.

74### Manage auto memory

76Auto memory files are markdown files you can edit at any time. Use `/memory` to open the file selector, which includes your auto memory entrypoint alongside your CLAUDE.md files.

78To ask Claude to save something specific, tell it directly: "remember that we use pnpm, not npm" or "save to memory that the API tests require a local Redis instance".

80When neither variable is set, auto memory follows the gradual rollout. The variable name uses double-negative logic: `DISABLE=0` means "don't disable" and forces auto memory on.

82```bash theme={null}

83export CLAUDE_CODE_DISABLE_AUTO_MEMORY=1 # Force off

84export CLAUDE_CODE_DISABLE_AUTO_MEMORY=0 # Force on

85```

25## CLAUDE.md imports87## CLAUDE.md imports

26 88

27CLAUDE.md files can import additional files using `@path/to/import` syntax. The following example imports 3 files:89CLAUDE.md files can import additional files using `@path/to/import` syntax. The following example imports 3 files:

33- git workflow @docs/git-instructions.md95- git workflow @docs/git-instructions.md

34```96```

35 97

36Both relative and absolute paths are allowed. In particular, importing files in user's home dir is a convenient way for your team members to provide individual instructions that are not checked into the repository. Imports are an alternative to CLAUDE.local.md that work better across multiple git worktrees.98Both relative and absolute paths are allowed. Relative paths resolve relative to the file containing the import, not the working directory. For private per-project preferences that shouldn't be checked into version control, prefer `CLAUDE.local.md`: it is automatically loaded and added to `.gitignore`.

100If you work across multiple git worktrees, `CLAUDE.local.md` only exists in one. Use a home-directory import instead so all worktrees share the same personal instructions:

37 101

38```102```

39# Individual Preferences103# Individual Preferences

40- @~/.claude/my-project-instructions.md104- @~/.claude/my-project-instructions.md

41```105```

42 106

107<Warning>

108 The first time Claude Code encounters external imports in a project, it shows an approval dialog listing the specific files. Approve to load them; decline to skip them. This is a one-time decision per project: once declined, the dialog does not resurface and the imports remain disabled.

109</Warning>

110

43To avoid potential collisions, imports are not evaluated inside markdown code spans and code blocks.111To avoid potential collisions, imports are not evaluated inside markdown code spans and code blocks.

44 112

45```113```

54 122

55Claude will also discover CLAUDE.md nested in subtrees under your current working directory. Instead of loading them at launch, they are only included when Claude reads files in those subtrees.123Claude will also discover CLAUDE.md nested in subtrees under your current working directory. Instead of loading them at launch, they are only included when Claude reads files in those subtrees.

56 124

125### Load memory from additional directories

126

127The `--add-dir` flag gives Claude access to additional directories outside your main working directory. By default, CLAUDE.md files from these directories are not loaded.

128

129To also load memory files (CLAUDE.md, .claude/CLAUDE.md, and .claude/rules/\*.md) from additional directories, set the `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD` environment variable:

130

131```bash theme={null}

132CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1 claude --add-dir ../shared-config

133```

134

57## Directly edit memories with `/memory`135## Directly edit memories with `/memory`

58 136

~~59Use the `/memory` slash command during a session to open any memory file in your system editor for more extensive additions or organization.~~137Use the `/memory` command during a session to open any memory file in your system editor for more extensive additions or organization.

60 138

61## Set up project memory139## Set up project memory

62 140

103 181

104```markdown theme={null}182```markdown theme={null}

105---183---

106paths: src/api/**/*.ts184paths:

185 - "src/api/**/*.ts"

107---186---

108 187

109# API Development Rules188# API Development Rules

126| `*.md` | Markdown files in the project root |205| `*.md` | Markdown files in the project root |

127| `src/components/*.tsx` | React components in a specific directory |206| `src/components/*.tsx` | React components in a specific directory |

128 207

129You can use braces to match multiple patterns efficiently:208You can specify multiple patterns:

130 209

131```markdown theme={null}210```markdown theme={null}

132---211---

133paths: src/**/*.{ts,tsx}212paths:

213 - "src/**/*.ts"

214 - "lib/**/*.ts"

215 - "tests/**/*.test.ts"

134---216---

~~135~~

136# TypeScript/React Rules

137```217```

138 218

139This expands to match both `src/**/*.ts` and `src/**/*.tsx`. You can also combine multiple patterns with commas:219Brace expansion is supported for matching multiple extensions or directories:

140 220

141```markdown theme={null}221```markdown theme={null}

142---222---

143paths: {src,lib}/**/*.ts, tests/**/*.test.ts223paths:

224 - "src/**/*.{ts,tsx}"

225 - "{src,lib}/**/*.ts"

144---226---

227

228# TypeScript/React Rules

145```229```

146 230

231This expands `src/**/*.{ts,tsx}` to match both `.ts` and `.tsx` files.

232

147### Subdirectories233### Subdirectories

148 234

149Rules can be organized into subdirectories for better structure:235Rules can be organized into subdirectories for better structure:

198 284

199## Organization-level memory management285## Organization-level memory management

200 286

201Enterprise organizations can deploy centrally managed CLAUDE.md files that apply to all users.287Organizations can deploy centrally managed CLAUDE.md files that apply to all users.

202 288

203To set up organization-level memory management:289To set up organization-level memory management:

204 290

2051. Create the enterprise memory file at the **Enterprise policy** location shown in the [memory types table above](#determine-memory-type).2911. Create the managed memory file at the **Managed policy** location shown in the [memory types table above](#determine-memory-type).

206 292

2072. Deploy via your configuration management system (MDM, Group Policy, Ansible, etc.) to ensure consistent distribution across all developer machines.2932. Deploy via your configuration management system (MDM, Group Policy, Ansible, etc.) to ensure consistent distribution across all developer machines.

208 294

211* **Be specific**: "Use 2-space indentation" is better than "Format code properly".297* **Be specific**: "Use 2-space indentation" is better than "Format code properly".

212* **Use structure to organize**: Format each individual memory as a bullet point and group related memories under descriptive markdown headings.298* **Use structure to organize**: Format each individual memory as a bullet point and group related memories under descriptive markdown headings.

213* **Review periodically**: Update memories as your project evolves to ensure Claude is always using the most up to date information and context.299* **Review periodically**: Update memories as your project evolves to ensure Claude is always using the most up to date information and context.

~~214~~

~~215~~

~~216~~

217> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

microsoft-foundry.md +6 −6

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code on Microsoft Foundry5# Claude Code on Microsoft Foundry

2 6

3> Learn about configuring Claude Code through Microsoft Foundry, including setup, configuration, and troubleshooting.7> Learn about configuring Claude Code through Microsoft Foundry, including setup, configuration, and troubleshooting.

64# Azure resource name (replace {resource} with your resource name)68# Azure resource name (replace {resource} with your resource name)

65export ANTHROPIC_FOUNDRY_RESOURCE={resource}69export ANTHROPIC_FOUNDRY_RESOURCE={resource}

66# Or provide the full base URL:70# Or provide the full base URL:

~~67# export ANTHROPIC_FOUNDRY_BASE_URL=https://{resource}.services.ai.azure.com~~71# export ANTHROPIC_FOUNDRY_BASE_URL=https://{resource}.services.ai.azure.com/anthropic

68 72

69# Set models to your resource's deployment names73# Set models to your resource's deployment names

70export ANTHROPIC_DEFAULT_SONNET_MODEL='claude-sonnet-4-5'74export ANTHROPIC_DEFAULT_SONNET_MODEL='claude-sonnet-4-5'

71export ANTHROPIC_DEFAULT_HAIKU_MODEL='claude-haiku-4-5'75export ANTHROPIC_DEFAULT_HAIKU_MODEL='claude-haiku-4-5'

~~72export ANTHROPIC_DEFAULT_OPUS_MODEL='claude-opus-4-1'~~76export ANTHROPIC_DEFAULT_OPUS_MODEL='claude-opus-4-6'

73```77```

74 78

75For more details on model configuration options, see [Model configuration](/en/model-config).79For more details on model configuration options, see [Model configuration](/en/model-config).

105* [Microsoft Foundry documentation](https://learn.microsoft.com/en-us/azure/ai-foundry/what-is-azure-ai-foundry)109* [Microsoft Foundry documentation](https://learn.microsoft.com/en-us/azure/ai-foundry/what-is-azure-ai-foundry)

106* [Microsoft Foundry models](https://ai.azure.com/explore/models)110* [Microsoft Foundry models](https://ai.azure.com/explore/models)

107* [Microsoft Foundry pricing](https://azure.microsoft.com/en-us/pricing/details/ai-foundry/)111* [Microsoft Foundry pricing](https://azure.microsoft.com/en-us/pricing/details/ai-foundry/)

~~108~~

~~109~~

~~110~~

111> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

model-config.md +44 −18

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Model configuration5# Model configuration

2 6

3> Learn about the Claude Code model configuration, including model aliases like `opusplan`7> Learn about the Claude Code model configuration, including model aliases like `opusplan`

8 12

9* A **model alias**13* A **model alias**

10* A **model name**14* A **model name**

~~11 * Anthropic API: A full **[model name](https://docs.claude.com/en/docs/about-claude/models/overview#model-names)**~~15 * Anthropic API: A full **[model name](https://platform.claude.com/docs/en/about-claude/models/overview)**

12 * Bedrock: an inference profile ARN16 * Bedrock: an inference profile ARN

13 * Foundry: a deployment name17 * Foundry: a deployment name

14 * Vertex: a version name18 * Vertex: a version name

19remembering exact version numbers:23remembering exact version numbers:

20 24

21| Model alias | Behavior |25| Model alias | Behavior |

~~22| ---------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- |~~26| ---------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

23| **`default`** | Recommended model setting, depending on your account type |27| **`default`** | Recommended model setting, depending on your account type |

24| **`sonnet`** | Uses the latest Sonnet model (currently Sonnet 4.5) for daily coding tasks |28| **`sonnet`** | Uses the latest Sonnet model (currently Sonnet 4.5) for daily coding tasks |

26| **`haiku`** | Uses the fast and efficient Haiku model for simple tasks |30| **`haiku`** | Uses the fast and efficient Haiku model for simple tasks |

~~27| **`sonnet[1m]`** | Uses Sonnet with a [1 million token context window](https://docs.claude.com/en/docs/build-with-claude/context-windows#1m-token-context-window) window for long sessions |~~31| **`sonnet[1m]`** | Uses Sonnet with a [1 million token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows#1m-token-context-window) for long sessions |

28| **`opusplan`** | Special mode that uses `opus` during plan mode, then switches to `sonnet` for execution |32| **`opusplan`** | Special mode that uses `opus` during plan mode, then switches to `sonnet` for execution |

29 33

34Aliases always point to the latest version. To pin to a specific version, use the full model name (for example, `claude-opus-4-6`) or set the corresponding environment variable like `ANTHROPIC_DEFAULT_OPUS_MODEL`.

30### Setting your model36### Setting your model

31 37

32You can configure your model in several ways, listed in order of priority:38You can configure your model in several ways, listed in order of priority:

62 68

63### `default` model setting69### `default` model setting

64 70

~~65The behavior of `default` depends on your account type.~~71The behavior of `default` depends on your account type:

66 72

~~67For certain Max users, Claude Code will automatically fall back to Sonnet if you~~73* **Max and Teams**: defaults to Opus 4.6

~~68hit a usage threshold with Opus.~~74* **Pro**: defaults to Opus 4.6 in Claude Code

75* **Enterprise**: Opus 4.6 is available but not the default

77Claude Code may automatically fall back to Sonnet if you hit a usage threshold with Opus.

69 78

70### `opusplan` model setting79### `opusplan` model setting

71 80

79This gives you the best of both worlds: Opus's superior reasoning for planning,88This gives you the best of both worlds: Opus's superior reasoning for planning,

80and Sonnet's efficiency for execution.89and Sonnet's efficiency for execution.

81 90

91### Adjust effort level

93[Effort levels](https://platform.claude.com/docs/en/build-with-claude/effort) control Opus 4.6's adaptive reasoning, which dynamically allocates thinking based on task complexity. Lower effort is faster and cheaper for straightforward tasks, while higher effort provides deeper reasoning for complex problems.

95Three levels are available: **low**, **medium**, and **high** (default).

97**Setting effort:**

99* **In `/model`**: use left/right arrow keys to adjust the effort slider when selecting a model

100* **Environment variable**: set `CLAUDE_CODE_EFFORT_LEVEL=low|medium|high`

101* **Settings**: set `effortLevel` in your settings file

102

103Effort is currently supported on Opus 4.6. The effort slider appears in `/model` when a supported model is selected.

104

82### Extended context with \[1m]105### Extended context with \[1m]

83 106

~~84For Console/API users, the `[1m]` suffix can be added to full model names to~~107The `[1m]` suffix enables a [1 million token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows#1m-token-context-window) for long sessions.

~~85enable a~~108

~~86[1 million token context window](https://docs.claude.com/en/docs/build-with-claude/context-windows#1m-token-context-window).~~109<Note>

110 For Opus 4.6, the 1M context window is available for API and Claude Code pay-as-you-go users. Pro, Max, Teams, and Enterprise subscription users do not have access to Opus 4.6 1M context at launch.

111</Note>

112

113You can use the `[1m]` suffix with model aliases or full model names:

87 114

88```bash theme={null}115```bash theme={null}

~~89# Example of using a full model name with the [1m] suffix~~116# Use the sonnet[1m] alias

~~90/model anthropic.claude-sonnet-4-5-20250929-v1:0[1m]~~117/model sonnet[1m]

118

119# Or append [1m] to a full model name

120/model claude-sonnet-4-5-20250929[1m]

91```121```

92 122

93Note: Extended context models have123Note: Extended context models have

~~94[different pricing](https://docs.claude.com/en/docs/about-claude/pricing#long-context-pricing).~~124[different pricing](https://platform.claude.com/docs/en/about-claude/pricing#long-context-pricing).

95 125

96## Checking your current model126## Checking your current model

97 127

117 147

118### Prompt caching configuration148### Prompt caching configuration

119 149

120Claude Code automatically uses [prompt caching](https://docs.claude.com/en/docs/build-with-claude/prompt-caching) to optimize performance and reduce costs. You can disable prompt caching globally or for specific model tiers:150Claude Code automatically uses [prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching) to optimize performance and reduce costs. You can disable prompt caching globally or for specific model tiers:

121 151

122| Environment variable | Description |152| Environment variable | Description |

123| ------------------------------- | ---------------------------------------------------------------------------------------------- |153| ------------------------------- | ---------------------------------------------------------------------------------------------- |

128 158

129These environment variables give you fine-grained control over prompt caching behavior. The global `DISABLE_PROMPT_CACHING` setting takes precedence over the model-specific settings, allowing you to quickly disable all caching when needed. The per-model settings are useful for selective control, such as when debugging specific models or working with cloud providers that may have different caching implementations.159These environment variables give you fine-grained control over prompt caching behavior. The global `DISABLE_PROMPT_CACHING` setting takes precedence over the model-specific settings, allowing you to quickly disable all caching when needed. The per-model settings are useful for selective control, such as when debugging specific models or working with cloud providers that may have different caching implementations.

~~130~~

~~131~~

~~132~~

133> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

monitoring-usage.md +15 −6

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Monitoring5# Monitoring

2 6

3> Learn how to enable and configure OpenTelemetry for Claude Code.7> Learn how to enable and configure OpenTelemetry for Claude Code.

67### Common configuration variables71### Common configuration variables

68 72

~~70| ----------------------------------------------- | ------------------------------------------------------------------------- | ------------------------------------ |~~74| ----------------------------------------------- | ------------------------------------------------------------------------------------------ | ------------------------------------ |

71| `CLAUDE_CODE_ENABLE_TELEMETRY` | Enables telemetry collection (required) | `1` |75| `CLAUDE_CODE_ENABLE_TELEMETRY` | Enables telemetry collection (required) | `1` |

83| `OTEL_METRIC_EXPORT_INTERVAL` | Export interval in milliseconds (default: 60000) | `5000`, `60000` |87| `OTEL_METRIC_EXPORT_INTERVAL` | Export interval in milliseconds (default: 60000) | `5000`, `60000` |

84| `OTEL_LOGS_EXPORT_INTERVAL` | Logs export interval in milliseconds (default: 5000) | `1000`, `10000` |88| `OTEL_LOGS_EXPORT_INTERVAL` | Logs export interval in milliseconds (default: 5000) | `1000`, `10000` |

90| `OTEL_LOG_TOOL_DETAILS` | Enable logging of MCP server/tool names and skill names in tool events (default: disabled) | `1` to enable |

86| `CLAUDE_CODE_OTEL_HEADERS_HELPER_DEBOUNCE_MS` | Interval for refreshing dynamic headers (default: 1740000ms / 29 minutes) | `900000` |91| `CLAUDE_CODE_OTEL_HEADERS_HELPER_DEBOUNCE_MS` | Interval for refreshing dynamic headers (default: 1740000ms / 29 minutes) | `900000` |

87 92

88### Metrics cardinality control93### Metrics cardinality control

330* All [standard attributes](#standard-attributes)335* All [standard attributes](#standard-attributes)

331* `event.name`: `"user_prompt"`336* `event.name`: `"user_prompt"`

332* `event.timestamp`: ISO 8601 timestamp337* `event.timestamp`: ISO 8601 timestamp

338* `event.sequence`: monotonically increasing counter for ordering events within a session

333* `prompt_length`: Length of the prompt339* `prompt_length`: Length of the prompt

334* `prompt`: Prompt content (redacted by default, enable with `OTEL_LOG_USER_PROMPTS=1`)340* `prompt`: Prompt content (redacted by default, enable with `OTEL_LOG_USER_PROMPTS=1`)

335 341

344* All [standard attributes](#standard-attributes)350* All [standard attributes](#standard-attributes)

345* `event.name`: `"tool_result"`351* `event.name`: `"tool_result"`

346* `event.timestamp`: ISO 8601 timestamp352* `event.timestamp`: ISO 8601 timestamp

353* `event.sequence`: monotonically increasing counter for ordering events within a session

347* `tool_name`: Name of the tool354* `tool_name`: Name of the tool

348* `success`: `"true"` or `"false"`355* `success`: `"true"` or `"false"`

349* `duration_ms`: Execution time in milliseconds356* `duration_ms`: Execution time in milliseconds

352* `source`: Decision source - `"config"`, `"user_permanent"`, `"user_temporary"`, `"user_abort"`, or `"user_reject"`359* `source`: Decision source - `"config"`, `"user_permanent"`, `"user_temporary"`, `"user_abort"`, or `"user_reject"`

353* `tool_parameters`: JSON string containing tool-specific parameters (when available)360* `tool_parameters`: JSON string containing tool-specific parameters (when available)

354 * For Bash tool: includes `bash_command`, `full_command`, `timeout`, `description`, `sandbox`361 * For Bash tool: includes `bash_command`, `full_command`, `timeout`, `description`, `sandbox`

362 * For MCP tools (when `OTEL_LOG_TOOL_DETAILS=1`): includes `mcp_server_name`, `mcp_tool_name`

363 * For Skill tool (when `OTEL_LOG_TOOL_DETAILS=1`): includes `skill_name`

355 364

356#### API request event365#### API request event

357 366

364* All [standard attributes](#standard-attributes)373* All [standard attributes](#standard-attributes)

365* `event.name`: `"api_request"`374* `event.name`: `"api_request"`

366* `event.timestamp`: ISO 8601 timestamp375* `event.timestamp`: ISO 8601 timestamp

376* `event.sequence`: monotonically increasing counter for ordering events within a session

367* `model`: Model used (for example, "claude-sonnet-4-5-20250929")377* `model`: Model used (for example, "claude-sonnet-4-5-20250929")

368* `cost_usd`: Estimated cost in USD378* `cost_usd`: Estimated cost in USD

369* `duration_ms`: Request duration in milliseconds379* `duration_ms`: Request duration in milliseconds

383* All [standard attributes](#standard-attributes)393* All [standard attributes](#standard-attributes)

384* `event.name`: `"api_error"`394* `event.name`: `"api_error"`

385* `event.timestamp`: ISO 8601 timestamp395* `event.timestamp`: ISO 8601 timestamp

396* `event.sequence`: monotonically increasing counter for ordering events within a session

386* `model`: Model used (for example, "claude-sonnet-4-5-20250929")397* `model`: Model used (for example, "claude-sonnet-4-5-20250929")

387* `error`: Error message398* `error`: Error message

388* `status_code`: HTTP status code (if applicable)399* `status_code`: HTTP status code (if applicable)

400* All [standard attributes](#standard-attributes)411* All [standard attributes](#standard-attributes)

401* `event.name`: `"tool_decision"`412* `event.name`: `"tool_decision"`

402* `event.timestamp`: ISO 8601 timestamp413* `event.timestamp`: ISO 8601 timestamp

414* `event.sequence`: monotonically increasing counter for ordering events within a session

403* `tool_name`: Name of the tool (for example, "Read", "Edit", "Write", "NotebookEdit")415* `tool_name`: Name of the tool (for example, "Read", "Edit", "Write", "NotebookEdit")

404* `decision`: Either `"accept"` or `"reject"`416* `decision`: Either `"accept"` or `"reject"`

405* `source`: Decision source - `"config"`, `"user_permanent"`, `"user_temporary"`, `"user_abort"`, or `"user_reject"`417* `source`: Decision source - `"config"`, `"user_permanent"`, `"user_temporary"`, `"user_abort"`, or `"user_reject"`

489 501

490* Telemetry is opt-in and requires explicit configuration502* Telemetry is opt-in and requires explicit configuration

491* Sensitive information like API keys or file contents are never included in metrics or events503* Sensitive information like API keys or file contents are never included in metrics or events

492* User prompt content is redacted by default - only prompt length is recorded. To enable user prompt logging, set `OTEL_LOG_USER_PROMPTS=1`504* User prompt content is redacted by default, only prompt length is recorded. To enable user prompt logging, set `OTEL_LOG_USER_PROMPTS=1`

505* MCP server/tool names and skill names are not logged by default because they can reveal user-specific configurations. To enable, set `OTEL_LOG_TOOL_DETAILS=1`

493 506

494## Monitoring Claude Code on Amazon Bedrock507## Monitoring Claude Code on Amazon Bedrock

495 508

496For detailed Claude Code usage monitoring guidance for Amazon Bedrock, see [Claude Code Monitoring Implementation (Bedrock)](https://github.com/aws-solutions-library-samples/guidance-for-claude-code-with-amazon-bedrock/blob/main/assets/docs/MONITORING.md).509For detailed Claude Code usage monitoring guidance for Amazon Bedrock, see [Claude Code Monitoring Implementation (Bedrock)](https://github.com/aws-solutions-library-samples/guidance-for-claude-code-with-amazon-bedrock/blob/main/assets/docs/MONITORING.md).

~~497~~

~~498~~

~~499~~

500> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

network-config.md +4 −4

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Enterprise network configuration5# Enterprise network configuration

2 6

3> Configure Claude Code for enterprise environments with proxy servers, custom Certificate Authorities (CA), and mutual Transport Layer Security (mTLS) authentication.7> Configure Claude Code for enterprise environments with proxy servers, custom Certificate Authorities (CA), and mutual Transport Layer Security (mTLS) authentication.

88* [Claude Code settings](/en/settings)92* [Claude Code settings](/en/settings)

89* [Environment variables reference](/en/settings#environment-variables)93* [Environment variables reference](/en/settings#environment-variables)

90* [Troubleshooting guide](/en/troubleshooting)94* [Troubleshooting guide](/en/troubleshooting)

~~94> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt~~

output-styles.md +6 −7

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Output styles5# Output styles

2 6

3> Adapt Claude Code for uses beyond software engineering7> Adapt Claude Code for uses beyond software engineering

103settings like the model to use, the tools they have available, and some context107settings like the model to use, the tools they have available, and some context

104about when to use the agent.108about when to use the agent.

105 109

106### Output Styles vs. [Custom Slash Commands](/en/slash-commands)110### Output Styles vs. [Skills](/en/skills)

~~107~~

108You can think of output styles as "stored system prompts" and custom slash

109commands as "stored prompts".

~~110~~

~~111~~

112 111

113> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt112Output styles modify how Claude responds (formatting, tone, structure) and are always active once selected. Skills are task-specific prompts that you invoke with `/skill-name` or that Claude loads automatically when relevant. Use output styles for consistent formatting preferences; use skills for reusable workflows and tasks.

overview.md +152 −73

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code overview5# Claude Code overview

2 6

~~3> Learn about Claude Code, Anthropic's agentic coding tool that lives in your terminal and helps you turn ideas into code faster than ever before.~~7> Claude Code is an agentic coding tool that reads your codebase, edits files, runs commands, and integrates with your development tools. Available in your terminal, IDE, desktop app, and browser.

4 8

~~5## Get started in 30 seconds~~9Claude Code is an agentic coding tool that reads your codebase, edits files, and runs commands. It works in your terminal, IDE, browser, and as a desktop app.

6 10

~~7Prerequisites:~~11## Get started

8 12

~~9* A [Claude.ai](https://claude.ai) (recommended) or [Claude Console](https://console.anthropic.com/) account~~13Choose your environment to get started. Most surfaces require a [Claude subscription](https://claude.com/pricing) or [Anthropic Console](https://console.anthropic.com/) account. The Terminal CLI and VS Code also support [third-party providers](/en/third-party-integrations).

10 14

~~11**Install Claude Code:**~~15<Tabs>

16 <Tab title="Terminal">

17 The full-featured CLI for working with Claude Code directly in your terminal. Edit files, run commands, and manage your entire project from the command line.

12 18

~~13To install Claude Code, use one of the following methods:~~19 To install Claude Code, use one of the following methods:

14 20

~~15<Tabs>~~21 <Tabs>

16 <Tab title="Native Install (Recommended)">22 <Tab title="Native Install (Recommended)">

17 **macOS, Linux, WSL:**23 **macOS, Linux, WSL:**

18 24

31 ```batch theme={null}37 ```batch theme={null}

32 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd38 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

33 ```39 ```

41 <Info>

42 Native installations automatically update in the background to keep you on the latest version.

43 </Info>

34 </Tab>44 </Tab>

35 45

36 <Tab title="Homebrew">46 <Tab title="Homebrew">

37 ```sh theme={null}47 ```sh theme={null}

38 brew install --cask claude-code48 brew install --cask claude-code

39 ```49 ```

51 <Info>

52 Homebrew installations do not auto-update. Run `brew upgrade claude-code` periodically to get the latest features and security fixes.

53 </Info>

54 </Tab>

56 <Tab title="WinGet">

57 ```powershell theme={null}

58 winget install Anthropic.ClaudeCode

59 ```

61 <Info>

62 WinGet installations do not auto-update. Run `winget upgrade Anthropic.ClaudeCode` periodically to get the latest features and security fixes.

63 </Info>

40 </Tab>64 </Tab>

65 </Tabs>

41 66

~~42 <Tab title="NPM">~~67 Then start Claude Code in any project:

~~43 If you have [Node.js 18 or newer installed](https://nodejs.org/en/download/):~~

44 68

~~45 ```sh theme={null}~~69 ```bash theme={null}

~~46 npm install -g @anthropic-ai/claude-code~~70 cd your-project

71 claude

47 ```72 ```

74 You'll be prompted to log in on first use. That's it! [Continue with the Quickstart →](/en/quickstart)

76 <Tip>

77 See [advanced setup](/en/setup) for installation options, manual updates, or uninstallation instructions. Visit [troubleshooting](/en/troubleshooting) if you hit issues.

78 </Tip>

79 </Tab>

81 <Tab title="VS Code">

82 The VS Code extension provides inline diffs, @-mentions, plan review, and conversation history directly in your editor.

84 * [Install for VS Code](vscode:extension/anthropic.claude-code)

85 * [Install for Cursor](cursor:extension/anthropic.claude-code)

87 Or search for "Claude Code" in the Extensions view (`Cmd+Shift+X` on Mac, `Ctrl+Shift+X` on Windows/Linux). After installing, open the Command Palette (`Cmd+Shift+P` / `Ctrl+Shift+P`), type "Claude Code", and select **Open in New Tab**.

89 [Get started with VS Code →](/en/vs-code#get-started)

90 </Tab>

92 <Tab title="Desktop app">

93 A standalone app for running Claude Code outside your IDE or terminal. Review diffs visually, run multiple sessions side by side, and kick off cloud sessions.

95 Download and install:

97 * [macOS](https://claude.ai/api/desktop/darwin/universal/dmg/latest/redirect?utm_source=claude_code\&utm_medium=docs) (Intel and Apple Silicon)

98 * [Windows](https://claude.ai/api/desktop/win32/x64/exe/latest/redirect?utm_source=claude_code\&utm_medium=docs) (x64)

100 After installing, launch Claude, sign in, and click the **Code** tab to start coding.

101

102 [Learn more about the desktop app →](/en/desktop#installation-and-setup)

103 </Tab>

104

105 <Tab title="Web">

106 Run Claude Code in your browser with no local setup. Kick off long-running tasks and check back when they're done, work on repos you don't have locally, or run multiple tasks in parallel. Available on desktop browsers and the Claude iOS app.

107

108 Start coding at [claude.ai/code](https://claude.ai/code).

109

110 [Get started on the web →](/en/claude-code-on-the-web#getting-started)

111 </Tab>

112

113 <Tab title="JetBrains">

114 A plugin for IntelliJ IDEA, PyCharm, WebStorm, and other JetBrains IDEs with interactive diff viewing and selection context sharing.

115

116 Install the [Claude Code plugin](https://plugins.jetbrains.com/plugin/27310-claude-code-beta-) from the JetBrains Marketplace and restart your IDE.

117

118 [Get started with JetBrains →](/en/jetbrains)

48 </Tab>119 </Tab>

49</Tabs>120</Tabs>

50 121

~~51**Start using Claude Code:**~~122## What you can do

123

124Here are some of the ways you can use Claude Code:

52 125

~~53```bash theme={null}~~126<AccordionGroup>

~~54cd your-project~~127 <Accordion title="Automate the work you keep putting off" icon="wand-magic-sparkles">

~~55claude~~128 Claude Code handles the tedious tasks that eat up your day: writing tests for untested code, fixing lint errors across a project, resolving merge conflicts, updating dependencies, and writing release notes.

~~56```~~

57 129

~~58You'll be prompted to log in on first use. That's it! [Continue with Quickstart (5 minutes) →](/en/quickstart)~~130 ```bash theme={null}

131 claude "write tests for the auth module, run them, and fix any failures"

132 ```

133 </Accordion>

59 134

~~60<Tip>~~135 <Accordion title="Build features and fix bugs" icon="hammer">

61 Claude Code automatically keeps itself up to date. See [advanced setup](/en/setup) for installation options, manual updates, or uninstallation instructions. Visit [troubleshooting](/en/troubleshooting) if you hit issues.136 Describe what you want in plain language. Claude Code plans the approach, writes the code across multiple files, and verifies it works.

~~62</Tip>~~

63 137

~~64## What Claude Code does for you~~138 For bugs, paste an error message or describe the symptom. Claude Code traces the issue through your codebase, identifies the root cause, and implements a fix. See [common workflows](/en/common-workflows) for more examples.

139 </Accordion>

65 140

~~66* **Build features from descriptions**: Tell Claude what you want to build in plain English. It will make a plan, write the code, and ensure it works.~~141 <Accordion title="Create commits and pull requests" icon="code-branch">

~~67* **Debug and fix issues**: Describe a bug or paste an error message. Claude Code will analyze your codebase, identify the problem, and implement a fix.~~142 Claude Code works directly with git. It stages changes, writes commit messages, creates branches, and opens pull requests.

68* **Navigate any codebase**: Ask anything about your team's codebase, and get a thoughtful answer back. Claude Code maintains awareness of your entire project structure, can find up-to-date information from the web, and with [MCP](/en/mcp) can pull from external data sources like Google Drive, Figma, and Slack.

~~69* **Automate tedious tasks**: Fix fiddly lint issues, resolve merge conflicts, and write release notes. Do all this in a single command from your developer machines, or automatically in CI.~~

70 143

~~71## Why developers love Claude Code~~144 ```bash theme={null}

145 claude "commit my changes with a descriptive message"

146 ```

72 147

~~73* **Works in your terminal**: Not another chat window. Not another IDE. Claude Code meets you where you already work, with the tools you already love.~~148 In CI, you can automate code review and issue triage with [GitHub Actions](/en/github-actions) or [GitLab CI/CD](/en/gitlab-ci-cd).

74* **Takes action**: Claude Code can directly edit files, run commands, and create commits. Need more? [MCP](/en/mcp) lets Claude read your design docs in Google Drive, update your tickets in Jira, or use *your* custom developer tooling.149 </Accordion>

75* **Unix philosophy**: Claude Code is composable and scriptable. `tail -f app.log | claude -p "Slack me if you see any anomalies appear in this log stream"` *works*. Your CI can run `claude -p "If there are new text strings, translate them into French and raise a PR for @lang-fr-team to review"`.

~~76* **Enterprise-ready**: Use the Claude API, or host on AWS or GCP. Enterprise-grade [security](/en/security), [privacy](/en/data-usage), and [compliance](https://trust.anthropic.com/) is built-in.~~

77 150

~~78## Next steps~~151 <Accordion title="Connect your tools with MCP" icon="plug">

152 The [Model Context Protocol (MCP)](/en/mcp) is an open standard for connecting AI tools to external data sources. With MCP, Claude Code can read your design docs in Google Drive, update tickets in Jira, pull data from Slack, or use your own custom tooling.

153 </Accordion>

79 154

~~80<CardGroup>~~155 <Accordion title="Customize with instructions, skills, and hooks" icon="sliders">

~~81 <Card title="Quickstart" icon="rocket" href="/en/quickstart">~~156 [`CLAUDE.md`](/en/claude-md) is a markdown file you add to your project root that Claude Code reads at the start of every session. Use it to set coding standards, architecture decisions, preferred libraries, and review checklists.

~~82 See Claude Code in action with practical examples~~

~~83 </Card>~~

84 157

~~85 <Card title="Common workflows" icon="graduation-cap" href="/en/common-workflows">~~158 Create [custom slash commands](/en/skills) to package repeatable workflows your team can share, like `/review-pr` or `/deploy-staging`.

~~86 Step-by-step guides for common workflows~~

~~87 </Card>~~

88 159

~~89 <Card title="Troubleshooting" icon="wrench" href="/en/troubleshooting">~~160 [Hooks](/en/hooks) let you run shell commands before or after Claude Code actions, like auto-formatting after every file edit or running lint before a commit.

~~90 Solutions for common issues with Claude Code~~161 </Accordion>

~~91 </Card>~~

92 162

~~93 <Card title="IDE setup" icon="laptop" href="/en/vs-code">~~163 <Accordion title="Run agent teams and build custom agents" icon="users">

~~94 Add Claude Code to your IDE~~164 Spawn [multiple Claude Code agents](/en/sub-agents) that work on different parts of a task simultaneously. A lead agent coordinates the work, assigns subtasks, and merges results.

~~95 </Card>~~

~~96</CardGroup>~~

97 165

~~98## Additional resources~~166 For fully custom workflows, the [Agent SDK](https://platform.claude.com/docs/en/agent-sdk/overview) lets you build your own agents powered by Claude Code's tools and capabilities, with full control over orchestration, tool access, and permissions.

167 </Accordion>

99 168

100<CardGroup>169 <Accordion title="Pipe, script, and automate with the CLI" icon="terminal">

101 <Card title="About Claude Code" icon="sparkles" href="https://claude.com/product/claude-code">170 Claude Code is composable and follows the Unix philosophy. Pipe logs into it, run it in CI, or chain it with other tools:

102 Learn more about Claude Code on claude.com

103 </Card>

104 171

105 <Card title="Build with the Agent SDK" icon="code-branch" href="https://docs.claude.com/en/docs/agent-sdk/overview">172 ```bash theme={null}

106 Create custom AI agents with the Claude Agent SDK173 # Monitor logs and get alerted

107 </Card>174 tail -f app.log | claude -p "Slack me if you see any anomalies"

108 175

109 <Card title="Host on AWS or GCP" icon="cloud" href="/en/third-party-integrations">176 # Automate translations in CI

110 Configure Claude Code with Amazon Bedrock or Google Vertex AI177 claude -p "translate new strings into French and raise a PR for review"

111 </Card>

112 178

113 <Card title="Settings" icon="gear" href="/en/settings">179 # Bulk operations across files

114 Customize Claude Code for your workflow180 git diff main --name-only | claude -p "review these changed files for security issues"

115 </Card>181 ```

182

183 See the [CLI reference](/en/cli-reference) for the full set of commands and flags.

184 </Accordion>

185

186 <Accordion title="Work from anywhere" icon="globe">

187 Start a task on your laptop and pick it up on your phone. [Claude Code on the web](/en/claude-code-on-the-web) and the [Claude iOS app](https://apps.apple.com/app/claude-by-anthropic/id6473753684) run sessions on cloud infrastructure, so you can kick off work from anywhere without a local dev environment.

116 188

117 <Card title="Commands" icon="terminal" href="/en/cli-reference">189 You can also route coding tasks straight from team chat: mention `@Claude` in [Slack](/en/slack) with a bug report or feature request, and get a pull request back.

118 Learn about CLI commands and controls190 </Accordion>

119 </Card>191</AccordionGroup>

120 192

121 <Card title="Reference implementation" icon="code" href="https://github.com/anthropics/claude-code/tree/main/.devcontainer">193## Use Claude Code everywhere

122 Clone our development container reference implementation

123 </Card>

124 194

125 <Card title="Security" icon="shield" href="/en/security">195Each surface connects to the same underlying Claude Code engine, so your CLAUDE.md files, settings, and MCP servers work across all of them.

126 Discover Claude Code's safeguards and best practices for safe usage

127 </Card>

128 196

129 <Card title="Privacy and data usage" icon="lock" href="/en/data-usage">197Beyond the [Terminal](/en/quickstart), [VS Code](/en/vs-code), [JetBrains](/en/jetbrains), [Desktop](/en/desktop), and [Web](/en/claude-code-on-the-web) environments above, Claude Code integrates with CI/CD, chat, and browser workflows:

130 Understand how Claude Code handles your data

131 </Card>

132</CardGroup>

133 198

199| I want to... | Best option |

200| --------------------------------------------- | ------------------------------------------------------------------------------------------------------------------ |

201| Start a task locally, continue on mobile | [Web](/en/claude-code-on-the-web) or [Claude iOS app](https://apps.apple.com/app/claude-by-anthropic/id6473753684) |

202| Automate PR reviews and issue triage | [GitHub Actions](/en/github-actions) or [GitLab CI/CD](/en/gitlab-ci-cd) |

203| Route bug reports from Slack to pull requests | [Slack](/en/slack) |

204| Debug live web applications | [Chrome](/en/chrome) |

205| Build custom agents for your own workflows | [Agent SDK](https://platform.claude.com/docs/en/agent-sdk/overview) |

206

207## Next steps

134 208

209Once you've installed Claude Code, these guides help you go deeper.

135 210

136> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt211* [Quickstart](/en/quickstart): walk through your first real task, from exploring a codebase to committing a fix

212* Level up with [best practices](/en/best-practices) and [common workflows](/en/common-workflows)

213* [Settings](/en/settings): customize Claude Code for your workflow

214* [Troubleshooting](/en/troubleshooting): solutions for common issues

215* [code.claude.com](https://code.claude.com/): demos, pricing, and product details

permissions.md +258 −0 added

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# Configure permissions

7> Control what Claude Code can access and do with fine-grained permission rules, modes, and managed policies.

9Claude Code supports fine-grained permissions so that you can specify exactly what the agent is allowed to do and what it cannot. Permission settings can be checked into version control and distributed to all developers in your organization, as well as customized by individual developers.

11## Permission system

13Claude Code uses a tiered permission system to balance power and safety:

16| :---------------- | :--------------- | :---------------- | :-------------------------------------------- |

17| Read-only | File reads, Grep | No | N/A |

18| Bash commands | Shell execution | Yes | Permanently per project directory and command |

19| File modification | Edit/write files | Yes | Until session end |

21## Manage permissions

23You can view and manage Claude Code's tool permissions with `/permissions`. This UI lists all permission rules and the settings.json file they are sourced from.

25* **Allow** rules let Claude Code use the specified tool without manual approval.

26* **Ask** rules prompt for confirmation whenever Claude Code tries to use the specified tool.

27* **Deny** rules prevent Claude Code from using the specified tool.

29Rules are evaluated in order: **deny -> ask -> allow**. The first matching rule wins, so deny rules always take precedence.

31## Permission modes

33Claude Code supports several permission modes that control how tools are approved. Set the `defaultMode` in your [settings files](/en/settings#settings-files):

35| Mode | Description |

36| :------------------ | :----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

37| `default` | Standard behavior: prompts for permission on first use of each tool |

38| `acceptEdits` | Automatically accepts file edit permissions for the session |

39| `plan` | Plan Mode: Claude can analyze but not modify files or execute commands |

40| `delegate` | Coordination-only mode for agent team leads. Restricts the lead to team management tools, so all implementation work happens through teammates. Only available when an agent team is active. See [delegate mode](/en/agent-teams#delegate-mode) for details. |

41| `dontAsk` | Auto-denies tools unless pre-approved via `/permissions` or `permissions.allow` rules |

42| `bypassPermissions` | Skips all permission prompts (requires safe environment, see warning below) |

44<Warning>

45 `bypassPermissions` mode disables all permission checks. Only use this in isolated environments like containers or VMs where Claude Code cannot cause damage. Administrators can prevent this mode by setting `disableBypassPermissionsMode` to `"disable"` in [managed settings](#managed-settings).

46</Warning>

48## Permission rule syntax

50Permission rules follow the format `Tool` or `Tool(specifier)`.

52### Match all uses of a tool

54To match all uses of a tool, use just the tool name without parentheses:

56| Rule | Effect |

57| :--------- | :----------------------------- |

58| `Bash` | Matches all Bash commands |

59| `WebFetch` | Matches all web fetch requests |

60| `Read` | Matches all file reads |

62`Bash(*)` is equivalent to `Bash` and matches all Bash commands.

64### Use specifiers for fine-grained control

66Add a specifier in parentheses to match specific tool uses:

68| Rule | Effect |

69| :----------------------------- | :------------------------------------------------------- |

70| `Bash(npm run build)` | Matches the exact command `npm run build` |

71| `Read(./.env)` | Matches reading the `.env` file in the current directory |

72| `WebFetch(domain:example.com)` | Matches fetch requests to example.com |

74### Wildcard patterns

76Bash rules support glob patterns with `*`. Wildcards can appear at any position in the command. This configuration allows npm and git commit commands while blocking git push:

78```json theme={null}

79{

80 "permissions": {

81 "allow": [

82 "Bash(npm run *)",

83 "Bash(git commit *)",

84 "Bash(git * main)",

85 "Bash(* --version)",

86 "Bash(* --help *)"

87 ],

88 "deny": [

89 "Bash(git push *)"

90 ]

91 }

92}

93```

95The space before `*` matters: `Bash(ls *)` matches `ls -la` but not `lsof`, while `Bash(ls*)` matches both. The legacy `:*` suffix syntax is equivalent to ` *` but is deprecated.

97## Tool-specific permission rules

99### Bash

100

101Bash permission rules support wildcard matching with `*`. Wildcards can appear at any position in the command, including at the beginning, middle, or end:

102

103* `Bash(npm run build)` matches the exact Bash command `npm run build`

104* `Bash(npm run test *)` matches Bash commands starting with `npm run test`

105* `Bash(npm *)` matches any command starting with `npm `

106* `Bash(* install)` matches any command ending with ` install`

107* `Bash(git * main)` matches commands like `git checkout main`, `git merge main`

108

109When `*` appears at the end with a space before it (like `Bash(ls *)`), it enforces a word boundary, requiring the prefix to be followed by a space or end-of-string. For example, `Bash(ls *)` matches `ls -la` but not `lsof`. In contrast, `Bash(ls*)` without a space matches both `ls -la` and `lsof` because there's no word boundary constraint.

110

111<Tip>

112 Claude Code is aware of shell operators (like `&&`) so a prefix match rule like `Bash(safe-cmd *)` won't give it permission to run the command `safe-cmd && other-cmd`.

113</Tip>

114

115<Warning>

116 Bash permission patterns that try to constrain command arguments are fragile. For example, `Bash(curl http://github.com/ *)` intends to restrict curl to GitHub URLs, but won't match variations like:

117

118 * Options before URL: `curl -X GET http://github.com/...`

119 * Different protocol: `curl https://github.com/...`

120 * Redirects: `curl -L http://bit.ly/xyz` (redirects to github)

121 * Variables: `URL=http://github.com && curl $URL`

122 * Extra spaces: `curl http://github.com`

123

124 For more reliable URL filtering, consider:

125

126 * **Restrict Bash network tools**: use deny rules to block `curl`, `wget`, and similar commands, then use the WebFetch tool with `WebFetch(domain:github.com)` permission for allowed domains

127 * **Use PreToolUse hooks**: implement a hook that validates URLs in Bash commands and blocks disallowed domains

128 * Instructing Claude Code about your allowed curl patterns via CLAUDE.md

129

130 Note that using WebFetch alone does not prevent network access. If Bash is allowed, Claude can still use `curl`, `wget`, or other tools to reach any URL.

131</Warning>

132

133### Read and Edit

134

135`Edit` rules apply to all built-in tools that edit files. Claude makes a best-effort attempt to apply `Read` rules to all built-in tools that read files like Grep and Glob.

136

137Read and Edit rules both follow the [gitignore](https://git-scm.com/docs/gitignore) specification with four distinct pattern types:

138

140| ------------------ | -------------------------------------- | -------------------------------- | ---------------------------------- |

141| `//path` | **Absolute** path from filesystem root | `Read(//Users/alice/secrets/**)` | `/Users/alice/secrets/**` |

142| `~/path` | Path from **home** directory | `Read(~/Documents/*.pdf)` | `/Users/alice/Documents/*.pdf` |

143| `/path` | Path **relative to settings file** | `Edit(/src/**/*.ts)` | `<settings file path>/src/**/*.ts` |

144| `path` or `./path` | Path **relative to current directory** | `Read(*.env)` | `<cwd>/*.env` |

145

146<Warning>

147 A pattern like `/Users/alice/file` is NOT an absolute path. It's relative to your settings file. Use `//Users/alice/file` for absolute paths.

148</Warning>

149

150Examples:

151

152* `Edit(/docs/**)`: edits in `<project>/docs/` (NOT `/docs/`)

153* `Read(~/.zshrc)`: reads your home directory's `.zshrc`

154* `Edit(//tmp/scratch.txt)`: edits the absolute path `/tmp/scratch.txt`

155* `Read(src/**)`: reads from `<current-directory>/src/`

156

157<Note>

158 In gitignore patterns, `*` matches files in a single directory while `**` matches recursively across directories. To allow all file access, use just the tool name without parentheses: `Read`, `Edit`, or `Write`.

159</Note>

160

161### WebFetch

162

163* `WebFetch(domain:example.com)` matches fetch requests to example.com

164

165### MCP

166

167* `mcp__puppeteer` matches any tool provided by the `puppeteer` server (name configured in Claude Code)

168* `mcp__puppeteer__*` wildcard syntax that also matches all tools from the `puppeteer` server

169* `mcp__puppeteer__puppeteer_navigate` matches the `puppeteer_navigate` tool provided by the `puppeteer` server

170

171### Task (subagents)

172

173Use `Task(AgentName)` rules to control which [subagents](/en/sub-agents) Claude can use:

174

175* `Task(Explore)` matches the Explore subagent

176* `Task(Plan)` matches the Plan subagent

177* `Task(my-custom-agent)` matches a custom subagent named `my-custom-agent`

178

179Add these rules to the `deny` array in your settings or use the `--disallowedTools` CLI flag to disable specific agents. To disable the Explore agent:

180

181```json theme={null}

182{

183 "permissions": {

184 "deny": ["Task(Explore)"]

185 }

186}

187```

188

189## Extend permissions with hooks

190

191[Claude Code hooks](/en/hooks-guide) provide a way to register custom shell commands to perform permission evaluation at runtime. When Claude Code makes a tool call, PreToolUse hooks run before the permission system, and the hook output can determine whether to approve or deny the tool call in place of the permission system.

192

193## Working directories

194

195By default, Claude has access to files in the directory where it was launched. You can extend this access:

196

197* **During startup**: use `--add-dir <path>` CLI argument

198* **During session**: use `/add-dir` command

199* **Persistent configuration**: add to `additionalDirectories` in [settings files](/en/settings#settings-files)

200

201Files in additional directories follow the same permission rules as the original working directory: they become readable without prompts, and file editing permissions follow the current permission mode.

202

203## How permissions interact with sandboxing

204

205Permissions and [sandboxing](/en/sandboxing) are complementary security layers:

206

207* **Permissions** control which tools Claude Code can use and which files or domains it can access. They apply to all tools (Bash, Read, Edit, WebFetch, MCP, and others).

208* **Sandboxing** provides OS-level enforcement that restricts the Bash tool's filesystem and network access. It applies only to Bash commands and their child processes.

209

210Use both for defense-in-depth:

211

212* Permission deny rules block Claude from even attempting to access restricted resources

213* Sandbox restrictions prevent Bash commands from reaching resources outside defined boundaries, even if a prompt injection bypasses Claude's decision-making

214* Filesystem restrictions in the sandbox use Read and Edit deny rules, not separate sandbox configuration

215* Network restrictions combine WebFetch permission rules with the sandbox's `allowedDomains` list

216

217## Managed settings

218

219For organizations that need centralized control over Claude Code configuration, administrators can deploy `managed-settings.json` files to system directories. These policy files follow the same format as regular settings files and cannot be overridden by user or project settings. For organizations without device management infrastructure, [server-managed settings](/en/server-managed-settings) provide an alternative that delivers configurations from Anthropic's servers.

220

221**Managed settings file locations**:

222

223* **macOS**: `/Library/Application Support/ClaudeCode/managed-settings.json`

224* **Linux and WSL**: `/etc/claude-code/managed-settings.json`

225* **Windows**: `C:\Program Files\ClaudeCode\managed-settings.json`

226

227<Note>

228 These are system-wide paths (not user home directories like `~/Library/...`) that require administrator privileges. They are designed to be deployed by IT administrators.

229</Note>

230

231### Managed-only settings

232

233Some settings are only effective in managed settings:

234

235| Setting | Description |

236| :-------------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------- |

237| `disableBypassPermissionsMode` | Set to `"disable"` to prevent `bypassPermissions` mode and the `--dangerously-skip-permissions` flag |

238| `allowManagedPermissionRulesOnly` | When `true`, prevents user and project settings from defining `allow`, `ask`, or `deny` permission rules. Only rules in managed settings apply |

239| `allowManagedHooksOnly` | When `true`, prevents loading of user, project, and plugin hooks. Only managed hooks and SDK hooks are allowed |

240| `strictKnownMarketplaces` | Controls which plugin marketplaces users can add. See [managed marketplace restrictions](/en/plugin-marketplaces#managed-marketplace-restrictions) |

241

242## Settings precedence

243

244Permission rules follow the same [settings precedence](/en/settings#settings-precedence) as all other Claude Code settings: managed settings have the highest precedence, followed by command line arguments, local project, shared project, and user settings.

245

246If a permission is allowed in user settings but denied in project settings, the project setting takes precedence and the permission is blocked.

247

248## Example configurations

249

250This [repository](https://github.com/anthropics/claude-code/tree/main/examples/settings) includes starter settings configurations for common deployment scenarios. Use these as starting points and adjust them to fit your needs.

251

252## See also

253

254* [Settings](/en/settings): complete configuration reference including the permission settings table

255* [Sandboxing](/en/sandboxing): OS-level filesystem and network isolation for Bash commands

256* [Authentication](/en/authentication): set up user access to Claude Code

257* [Security](/en/security): security safeguards and best practices

258* [Hooks](/en/hooks-guide): automate workflows and extend permission evaluation

plugin-marketplaces.md +137 −18

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Create and distribute a plugin marketplace5# Create and distribute a plugin marketplace

2 6

3> Build and host plugin marketplaces to distribute Claude Code extensions across teams and communities.7> Build and host plugin marketplaces to distribute Claude Code extensions across teams and communities.

19 23

20## Walkthrough: create a local marketplace24## Walkthrough: create a local marketplace

21 25

22This example creates a marketplace with one plugin: a `/review` command for code reviews. You'll create the directory structure, add a slash command, create the plugin manifest and marketplace catalog, then install and test it.26This example creates a marketplace with one plugin: a `/review` skill for code reviews. You'll create the directory structure, add a skill, create the plugin manifest and marketplace catalog, then install and test it.

23 27

24<Steps>28<Steps>

25 <Step title="Create the directory structure">29 <Step title="Create the directory structure">

26 ```bash theme={null}30 ```bash theme={null}

27 mkdir -p my-marketplace/.claude-plugin31 mkdir -p my-marketplace/.claude-plugin

28 mkdir -p my-marketplace/plugins/review-plugin/.claude-plugin32 mkdir -p my-marketplace/plugins/review-plugin/.claude-plugin

~~29 mkdir -p my-marketplace/plugins/review-plugin/commands~~33 mkdir -p my-marketplace/plugins/review-plugin/skills/review

30 ```34 ```

31 </Step>35 </Step>

32 36

~~33 <Step title="Create the plugin command">~~37 <Step title="Create the skill">

~~34 Create a Markdown file that defines what the `/review` command does.~~38 Create a `SKILL.md` file that defines what the `/review` skill does.

40 ```markdown my-marketplace/plugins/review-plugin/skills/review/SKILL.md theme={null}

41 ---

42 description: Review code for bugs, security, and performance

43 disable-model-invocation: true

44 ---

35 45

~~36 ```markdown my-marketplace/plugins/review-plugin/commands/review.md theme={null}~~

37 Review the code I've selected or the recent changes for:46 Review the code I've selected or the recent changes for:

38 - Potential bugs or edge cases47 - Potential bugs or edge cases

39 - Security concerns48 - Security concerns

50 ```json my-marketplace/plugins/review-plugin/.claude-plugin/plugin.json theme={null}59 ```json my-marketplace/plugins/review-plugin/.claude-plugin/plugin.json theme={null}

51 {60 {

52 "name": "review-plugin",61 "name": "review-plugin",

~~53 "description": "Adds a /review command for quick code reviews",~~62 "description": "Adds a /review skill for quick code reviews",

54 "version": "1.0.0"63 "version": "1.0.0"

55 }64 }

56 ```65 ```

69 {78 {

70 "name": "review-plugin",79 "name": "review-plugin",

71 "source": "./plugins/review-plugin",80 "source": "./plugins/review-plugin",

~~72 "description": "Adds a /review command for quick code reviews"~~81 "description": "Adds a /review skill for quick code reviews"

73 }82 }

74 ]83 ]

75 }84 }

182**Standard metadata fields:**191**Standard metadata fields:**

183 192

185| :------------ | :------ | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |194| :------------ | :------ | :----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

195| `strict` | boolean | Controls whether plugins need their own `plugin.json` file. When `true` (default), the plugin source must contain a `plugin.json`, and any fields you add here in the marketplace entry get merged with it. When `false`, the plugin doesn't need its own `plugin.json`; the marketplace entry itself defines everything about the plugin. Use `false` when you want to define simple plugins entirely in your marketplace file. |204| `strict` | boolean | When true (default), marketplace component fields merge with plugin.json. When false, the marketplace entry defines the plugin entirely, and plugin.json must not also declare components. |

196 205

197**Component configuration fields:**206**Component configuration fields:**

198 207

217}226}

218```227```

219 228

229<Note>

230 Relative paths only work when users add your marketplace via Git (GitHub, GitLab, or git URL). If users add your marketplace via a direct URL to the `marketplace.json` file, relative paths will not resolve correctly. For URL-based distribution, use GitHub, npm, or git URL sources instead. See [Troubleshooting](#plugins-with-relative-paths-fail-in-url-based-marketplaces) for details.

231</Note>

232

220### GitHub repositories233### GitHub repositories

221 234

222```json theme={null}235```json theme={null}

229}242}

230```243```

231 244

245You can pin to a specific branch, tag, or commit:

246

247```json theme={null}

248{

249 "name": "github-plugin",

250 "source": {

251 "source": "github",

252 "repo": "owner/plugin-repo",

253 "ref": "v2.0.0",

254 "sha": "a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7f8a9b0"

255 }

256}

257```

258

259| Field | Type | Description |

260| :----- | :----- | :-------------------------------------------------------------------- |

261| `repo` | string | Required. GitHub repository in `owner/repo` format |

262| `ref` | string | Optional. Git branch or tag (defaults to repository default branch) |

263| `sha` | string | Optional. Full 40-character git commit SHA to pin to an exact version |

264

232### Git repositories265### Git repositories

233 266

234```json theme={null}267```json theme={null}

241}274}

242```275```

243 276

277You can pin to a specific branch, tag, or commit:

278

279```json theme={null}

280{

281 "name": "git-plugin",

282 "source": {

283 "source": "url",

284 "url": "https://gitlab.com/team/plugin.git",

285 "ref": "main",

286 "sha": "a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7f8a9b0"

287 }

288}

289```

290

291| Field | Type | Description |

292| :---- | :----- | :-------------------------------------------------------------------- |

293| `url` | string | Required. Full git repository URL (must end with `.git`) |

294| `ref` | string | Optional. Git branch or tag (defaults to repository default branch) |

295| `sha` | string | Optional. Full 40-character git commit SHA to pin to an exact version |

296

244### Advanced plugin entries297### Advanced plugin entries

245 298

246This example shows a plugin entry using many of the optional fields, including custom paths for commands, agents, hooks, and MCP servers:299This example shows a plugin entry using many of the optional fields, including custom paths for commands, agents, hooks, and MCP servers:

318/plugin marketplace add https://gitlab.com/company/plugins.git371/plugin marketplace add https://gitlab.com/company/plugins.git

319```372```

320 373

374### Private repositories

375

376Claude Code supports installing plugins from private repositories. For manual installation and updates, Claude Code uses your existing git credential helpers. If `git clone` works for a private repository in your terminal, it works in Claude Code too. Common credential helpers include `gh auth login` for GitHub, macOS Keychain, and `git-credential-store`.

377

378Background auto-updates run at startup without credential helpers, since interactive prompts would block Claude Code from starting. To enable auto-updates for private marketplaces, set the appropriate authentication token in your environment:

379

380| Provider | Environment variables | Notes |

381| :-------- | :--------------------------- | :---------------------------------------- |

382| GitHub | `GITHUB_TOKEN` or `GH_TOKEN` | Personal access token or GitHub App token |

383| GitLab | `GITLAB_TOKEN` or `GL_TOKEN` | Personal access token or project token |

384| Bitbucket | `BITBUCKET_TOKEN` | App password or repository access token |

385

386Set the token in your shell configuration (for example, `.bashrc`, `.zshrc`) or pass it when running Claude Code:

387

388```bash theme={null}

389export GITHUB_TOKEN=ghp_xxxxxxxxxxxxxxxxxxxx

390```

391

392<Note>

393 For CI/CD environments, configure the token as a secret environment variable. GitHub Actions automatically provides `GITHUB_TOKEN` for repositories in the same organization.

394</Note>

395

321### Test locally before distribution396### Test locally before distribution

322 397

323Test your marketplace locally before sharing:398Test your marketplace locally before sharing:

359 434

360For full configuration options, see [Plugin settings](/en/settings#plugin-settings).435For full configuration options, see [Plugin settings](/en/settings#plugin-settings).

361 436

362### Enterprise marketplace restrictions437### Managed marketplace restrictions

363 438

364For organizations requiring strict control over plugin sources, enterprise administrators can restrict which plugin marketplaces users are allowed to add using the [`strictKnownMarketplaces`](/en/settings#strictknownmarketplaces) setting in managed settings.439For organizations requiring strict control over plugin sources, administrators can restrict which plugin marketplaces users are allowed to add using the [`strictKnownMarketplaces`](/en/settings#strictknownmarketplaces) setting in managed settings.

365 440

366When `strictKnownMarketplaces` is configured in managed settings, the restriction behavior depends on the value:441When `strictKnownMarketplaces` is configured in managed settings, the restriction behavior depends on the value:

367 442

403}478}

404```479```

405 480

481Allow all marketplaces from an internal git server using regex pattern matching:

482

483```json theme={null}

484{

485 "strictKnownMarketplaces": [

486 {

487 "source": "hostPattern",

488 "hostPattern": "^github\\.example\\.com$"

489 }

490 ]

491}

492```

493

406#### How restrictions work494#### How restrictions work

407 495

408Restrictions are validated early in the plugin installation process, before any network requests or filesystem operations occur. This prevents unauthorized marketplace access attempts.496Restrictions are validated early in the plugin installation process, before any network requests or filesystem operations occur. This prevents unauthorized marketplace access attempts.

409 497

410The allowlist uses exact matching. For a marketplace to be allowed, all specified fields must match exactly:498The allowlist uses exact matching for most source types. For a marketplace to be allowed, all specified fields must match exactly:

411 499

412* For GitHub sources: `repo` is required, and `ref` or `path` must also match if specified in the allowlist500* For GitHub sources: `repo` is required, and `ref` or `path` must also match if specified in the allowlist

413* For URL sources: the full URL must match exactly501* For URL sources: the full URL must match exactly

502* For `hostPattern` sources: the marketplace host is matched against the regex pattern

414 503

415Because `strictKnownMarketplaces` is set in [managed settings](/en/settings#settings-file-locations), individual users and project configurations cannot override these restrictions.504Because `strictKnownMarketplaces` is set in [managed settings](/en/settings#settings-files), individual users and project configurations cannot override these restrictions.

416 505

417For complete configuration details including all supported source types and comparison with `extraKnownMarketplaces`, see the [strictKnownMarketplaces reference](/en/settings#strictknownmarketplaces).506For complete configuration details including all supported source types and comparison with `extraKnownMarketplaces`, see the [strictKnownMarketplaces reference](/en/settings#strictknownmarketplaces).

418 507

487* For GitHub sources, ensure repositories are public or you have access576* For GitHub sources, ensure repositories are public or you have access

488* Test plugin sources manually by cloning/downloading577* Test plugin sources manually by cloning/downloading

489 578

579### Private repository authentication fails

580

581**Symptoms**: Authentication errors when installing plugins from private repositories

582

583**Solutions**:

584

585For manual installation and updates:

586

587* Verify you're authenticated with your git provider (for example, run `gh auth status` for GitHub)

588* Check that your credential helper is configured correctly: `git config --global credential.helper`

589* Try cloning the repository manually to verify your credentials work

590

591For background auto-updates:

592

593* Set the appropriate token in your environment: `echo $GITHUB_TOKEN`

594* Check that the token has the required permissions (read access to the repository)

595* For GitHub, ensure the token has the `repo` scope for private repositories

596* For GitLab, ensure the token has at least `read_repository` scope

597* Verify the token hasn't expired

598

599### Plugins with relative paths fail in URL-based marketplaces

600

601**Symptoms**: Added a marketplace via URL (such as `https://example.com/marketplace.json`), but plugins with relative path sources like `"./plugins/my-plugin"` fail to install with "path not found" errors.

602

603**Cause**: URL-based marketplaces only download the `marketplace.json` file itself. They do not download plugin files from the server. Relative paths in the marketplace entry reference files on the remote server that were not downloaded.

604

605**Solutions**:

606

607* **Use external sources**: Change plugin entries to use GitHub, npm, or git URL sources instead of relative paths:

608 ```json theme={null}

609 { "name": "my-plugin", "source": { "source": "github", "repo": "owner/repo" } }

610 ```

611* **Use a Git-based marketplace**: Host your marketplace in a Git repository and add it with the git URL. Git-based marketplaces clone the entire repository, making relative paths work correctly.

612

490### Files not found after installation613### Files not found after installation

491 614

492**Symptoms**: Plugin installs but references to files fail, especially files outside the plugin directory615**Symptoms**: Plugin installs but references to files fail, especially files outside the plugin directory

503* [Plugins](/en/plugins) - Creating your own plugins626* [Plugins](/en/plugins) - Creating your own plugins

504* [Plugins reference](/en/plugins-reference) - Complete technical specifications and schemas627* [Plugins reference](/en/plugins-reference) - Complete technical specifications and schemas

505* [Plugin settings](/en/settings#plugin-settings) - Plugin configuration options628* [Plugin settings](/en/settings#plugin-settings) - Plugin configuration options

506* [strictKnownMarketplaces reference](/en/settings#strictknownmarketplaces) - Enterprise marketplace restrictions629* [strictKnownMarketplaces reference](/en/settings#strictknownmarketplaces) - Managed marketplace restrictions

~~507~~

~~508~~

~~509~~

510> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

plugins.md +37 −39

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Create plugins5# Create plugins

2 6

~~3> Create custom plugins to extend Claude Code with slash commands, agents, hooks, Skills, and MCP servers.~~7> Create custom plugins to extend Claude Code with skills, agents, hooks, and MCP servers.

4 8

5Plugins let you extend Claude Code with custom functionality that can be shared across projects and teams. This guide covers creating your own plugins with slash commands, agents, Skills, hooks, and MCP servers.9Plugins let you extend Claude Code with custom functionality that can be shared across projects and teams. This guide covers creating your own plugins with skills, agents, hooks, and MCP servers.

6 10

7Looking to install existing plugins? See [Discover and install plugins](/en/discover-plugins). For complete technical specifications, see [Plugins reference](/en/plugins-reference).11Looking to install existing plugins? See [Discover and install plugins](/en/discover-plugins). For complete technical specifications, see [Plugins reference](/en/plugins-reference).

8 12

9## When to use plugins vs standalone configuration13## When to use plugins vs standalone configuration

10 14

~~11Claude Code supports two ways to add custom slash commands, agents, and hooks:~~15Claude Code supports two ways to add custom skills, agents, and hooks:

12 16

14| :---------------------------------------------------------- | :------------------- | :---------------------------------------------------------------------------------------------- |18| :---------------------------------------------------------- | :------------------- | :---------------------------------------------------------------------------------------------- |

19 23

20* You're customizing Claude Code for a single project24* You're customizing Claude Code for a single project

21* The configuration is personal and doesn't need to be shared25* The configuration is personal and doesn't need to be shared

~~22* You're experimenting with slash commands or hooks before packaging them~~26* You're experimenting with skills or hooks before packaging them

~~23* You want short slash command names like `/hello` or `/review`~~27* You want short skill names like `/hello` or `/review`

24 28

25**Use plugins when**:29**Use plugins when**:

26 30

27* You want to share functionality with your team or community31* You want to share functionality with your team or community

~~28* You need the same slash commands/agents across multiple projects~~32* You need the same skills/agents across multiple projects

29* You want version control and easy updates for your extensions33* You want version control and easy updates for your extensions

30* You're distributing through a marketplace34* You're distributing through a marketplace

~~31* You're okay with namespaced slash commands like `/my-plugin:hello` (namespacing prevents conflicts between plugins)~~35* You're okay with namespaced skills like `/my-plugin:hello` (namespacing prevents conflicts between plugins)

32 36

33<Tip>37<Tip>

34 Start with standalone configuration in `.claude/` for quick iteration, then [convert to a plugin](#convert-existing-configurations-to-plugins) when you're ready to share.38 Start with standalone configuration in `.claude/` for quick iteration, then [convert to a plugin](#convert-existing-configurations-to-plugins) when you're ready to share.

36 40

37## Quickstart41## Quickstart

38 42

39This quickstart walks you through creating a plugin with a custom slash command. You'll create a manifest (the configuration file that defines your plugin), add a slash command, and test it locally using the `--plugin-dir` flag.43This quickstart walks you through creating a plugin with a custom skill. You'll create a manifest (the configuration file that defines your plugin), add a skill, and test it locally using the `--plugin-dir` flag.

40 44

41### Prerequisites45### Prerequisites

42 46

51 55

52<Steps>56<Steps>

53 <Step title="Create the plugin directory">57 <Step title="Create the plugin directory">

~~54 Every plugin lives in its own directory containing a manifest and your custom commands, agents, or hooks. Create one now:~~58 Every plugin lives in its own directory containing a manifest and your skills, agents, or hooks. Create one now:

55 59

56 ```bash theme={null}60 ```bash theme={null}

57 mkdir my-first-plugin61 mkdir my-first-plugin

81 ```85 ```

82 86

83 | Field | Purpose |87 | Field | Purpose |

~~84 | :------------ | :--------------------------------------------------------------------------------------------------------------------- |~~88 | :------------ | :----------------------------------------------------------------------------------------------------- |

86 | `description` | Shown in the plugin manager when browsing or installing plugins. |90 | `description` | Shown in the plugin manager when browsing or installing plugins. |

87 | `version` | Track releases using [semantic versioning](/en/plugins-reference#version-management). |91 | `version` | Track releases using [semantic versioning](/en/plugins-reference#version-management). |

88 | `author` | Optional. Helpful for attribution. |92 | `author` | Optional. Helpful for attribution. |

90 For additional fields like `homepage`, `repository`, and `license`, see the [full manifest schema](/en/plugins-reference#plugin-manifest-schema).94 For additional fields like `homepage`, `repository`, and `license`, see the [full manifest schema](/en/plugins-reference#plugin-manifest-schema).

91 </Step>95 </Step>

92 96

~~93 <Step title="Add a slash command">~~97 <Step title="Add a skill">

94 Slash commands are Markdown files in the `commands/` directory. The filename becomes the slash command name, prefixed with the plugin's namespace (`hello.md` in a plugin named `my-first-plugin` creates `/my-first-plugin:hello`). The Markdown content tells Claude how to respond when someone runs the slash command.98 Skills live in the `skills/` directory. Each skill is a folder containing a `SKILL.md` file. The folder name becomes the skill name, prefixed with the plugin's namespace (`hello/` in a plugin named `my-first-plugin` creates `/my-first-plugin:hello`).

95 99

~~96 Create a `commands` directory in your plugin folder:~~100 Create a skill directory in your plugin folder:

97 101

98 ```bash theme={null}102 ```bash theme={null}

~~99 mkdir my-first-plugin/commands~~103 mkdir -p my-first-plugin/skills/hello

100 ```104 ```

101 105

102 Then create `my-first-plugin/commands/hello.md` with this content:106 Then create `my-first-plugin/skills/hello/SKILL.md` with this content:

103 107

104 ```markdown my-first-plugin/commands/hello.md theme={null}108 ```markdown my-first-plugin/skills/hello/SKILL.md theme={null}

105 ---109 ---

106 description: Greet the user with a friendly message110 description: Greet the user with a friendly message

111 disable-model-invocation: true

107 ---112 ---

108 113

109 # Hello Command

~~110~~

111 Greet the user warmly and ask how you can help them today.114 Greet the user warmly and ask how you can help them today.

112 ```115 ```

113 </Step>116 </Step>

128 You'll see Claude respond with a greeting. Run `/help` to see your command listed under the plugin namespace.131 You'll see Claude respond with a greeting. Run `/help` to see your command listed under the plugin namespace.

129 132

130 <Note>133 <Note>

131 **Why namespacing?** Plugin slash commands are always namespaced (like `/greet:hello`) to prevent conflicts when multiple plugins have commands with the same name.134 **Why namespacing?** Plugin skills are always namespaced (like `/greet:hello`) to prevent conflicts when multiple plugins have skills with the same name.

132 135

133 To change the namespace prefix, update the `name` field in `plugin.json`.136 To change the namespace prefix, update the `name` field in `plugin.json`.

134 </Note>137 </Note>

135 </Step>138 </Step>

136 139

137 <Step title="Add slash command arguments">140 <Step title="Add skill arguments">

138 Make your slash command dynamic by accepting user input. The `$ARGUMENTS` placeholder captures any text the user provides after the slash command.141 Make your skill dynamic by accepting user input. The `$ARGUMENTS` placeholder captures any text the user provides after the skill name.

139 142

140 Update your `hello.md` file:143 Update your `hello.md` file:

141 144

155 /my-first-plugin:hello Alex158 /my-first-plugin:hello Alex

156 ```159 ```

157 160

158 Claude will greet you by name. For more argument options like `$1`, `$2` for individual parameters, see [Slash commands](/en/slash-commands).161 Claude will greet you by name. For more on passing arguments to skills, see [Skills](/en/skills#pass-arguments-to-skills).

159 </Step>162 </Step>

160</Steps>163</Steps>

161 164

162You've successfully created and tested a plugin with these key components:165You've successfully created and tested a plugin with these key components:

163 166

164* **Plugin manifest** (`.claude-plugin/plugin.json`): describes your plugin's metadata167* **Plugin manifest** (`.claude-plugin/plugin.json`): describes your plugin's metadata

165* **Commands directory** (`commands/`): contains your custom slash commands168* **Commands directory** (`commands/`): contains your custom skills

166* **Command arguments** (`$ARGUMENTS`): captures user input for dynamic behavior169* **Skill arguments** (`$ARGUMENTS`): captures user input for dynamic behavior

167 170

168<Tip>171<Tip>

169 The `--plugin-dir` flag is useful for development and testing. When you're ready to share your plugin with others, see [Create and distribute a plugin marketplace](/en/plugin-marketplaces).172 The `--plugin-dir` flag is useful for development and testing. When you're ready to share your plugin with others, see [Create and distribute a plugin marketplace](/en/plugin-marketplaces).

171 174

172## Plugin structure overview175## Plugin structure overview

173 176

174You've created a plugin with a slash command, but plugins can include much more: custom agents, Skills, hooks, MCP servers, and LSP servers.177You've created a plugin with a skill, but plugins can include much more: custom agents, hooks, MCP servers, and LSP servers.

175 178

176<Warning>179<Warning>

177 **Common mistake**: Don't put `commands/`, `agents/`, `skills/`, or `hooks/` inside the `.claude-plugin/` directory. Only `plugin.json` goes inside `.claude-plugin/`. All other directories must be at the plugin root level.180 **Common mistake**: Don't put `commands/`, `agents/`, `skills/`, or `hooks/` inside the `.claude-plugin/` directory. Only `plugin.json` goes inside `.claude-plugin/`. All other directories must be at the plugin root level.

178</Warning>181</Warning>

179 182

181| :---------------- | :---------- | :---------------------------------------------- |184| :---------------- | :---------- | :----------------------------------------------------------------------------- |

302 305

303## Convert existing configurations to plugins306## Convert existing configurations to plugins

304 307

305If you already have custom commands, Skills, or hooks in your `.claude/` directory, you can convert them into a plugin for easier sharing and distribution.308If you already have skills or hooks in your `.claude/` directory, you can convert them into a plugin for easier sharing and distribution.

306 309

307### Migration steps310### Migration steps

308 311

347 mkdir my-plugin/hooks350 mkdir my-plugin/hooks

348 ```351 ```

349 352

350 Create `my-plugin/hooks/hooks.json` with your hooks configuration. Copy the `hooks` object from your `.claude/settings.json` or `settings.local.json`—the format is the same:353 Create `my-plugin/hooks/hooks.json` with your hooks configuration. Copy the `hooks` object from your `.claude/settings.json` or `settings.local.json`, since the format is the same. The command receives hook input as JSON on stdin, so use `jq` to extract the file path:

351 354

352 ```json my-plugin/hooks/hooks.json theme={null}355 ```json my-plugin/hooks/hooks.json theme={null}

353 {356 {

355 "PostToolUse": [358 "PostToolUse": [

356 {359 {

357 "matcher": "Write|Edit",360 "matcher": "Write|Edit",

358 "hooks": [{ "type": "command", "command": "npm run lint:fix $FILE" }]361 "hooks": [{ "type": "command", "command": "jq -r '.tool_input.file_path' | xargs npm run lint:fix" }]

359 }362 }

360 ]363 ]

361 }364 }

401* [Create and distribute a marketplace](/en/plugin-marketplaces): package and share your plugins404* [Create and distribute a marketplace](/en/plugin-marketplaces): package and share your plugins

402* [Plugins reference](/en/plugins-reference): complete technical specifications405* [Plugins reference](/en/plugins-reference): complete technical specifications

403* Dive deeper into specific plugin components:406* Dive deeper into specific plugin components:

404 * [Slash commands](/en/slash-commands): command development details407 * [Skills](/en/skills): skill development details

405 * [Subagents](/en/sub-agents): agent configuration and capabilities408 * [Subagents](/en/sub-agents): agent configuration and capabilities

406 * [Agent Skills](/en/skills): extend Claude's capabilities

407 * [Hooks](/en/hooks): event handling and automation409 * [Hooks](/en/hooks): event handling and automation

408 * [MCP](/en/mcp): external tool integration410 * [MCP](/en/mcp): external tool integration

~~409~~

~~410~~

~~411~~

412> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

plugins-reference.md +76 −101

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Plugins reference5# Plugins reference

2 6

3> Complete technical reference for Claude Code plugin system, including schemas, CLI commands, and component specifications.7> Complete technical reference for Claude Code plugin system, including schemas, CLI commands, and component specifications.

10 14

11## Plugin components reference15## Plugin components reference

12 16

~~13This section documents the five types of components that plugins can provide.~~17This section documents the types of components that plugins can provide.

19### Skills

14 20

~~15### Commands~~21Plugins add skills to Claude Code, creating `/name` shortcuts that you or Claude can invoke.

16 22

~~17Plugins add custom slash commands that integrate seamlessly with Claude Code's command system.~~23**Location**: `skills/` or `commands/` directory in plugin root

18 24

~~19**Location**: `commands/` directory in plugin root~~25**File format**: Skills are directories with `SKILL.md`; commands are simple markdown files

27**Skill structure**:

29```

30skills/

31├── pdf-processor/

32│ ├── SKILL.md

33│ ├── reference.md (optional)

34│ └── scripts/ (optional)

35└── code-reviewer/

36 └── SKILL.md

37```

39**Integration behavior**:

20 40

~~21**File format**: Markdown files with frontmatter~~41* Skills and commands are automatically discovered when the plugin is installed

42* Claude can invoke them automatically based on task context

43* Skills can include supporting files alongside SKILL.md

22 44

~~23For complete details on plugin command structure, invocation patterns, and features, see [Plugin commands](/en/slash-commands#plugin-commands).~~45For complete details, see [Skills](/en/skills).

24 46

25### Agents47### Agents

26 48

34 56

35```markdown theme={null}57```markdown theme={null}

36---58---

~~37description: What this agent specializes in~~59name: agent-name

~~38capabilities: ["task1", "task2", "task3"]~~60description: What this agent specializes in and when Claude should invoke it

39---61---

40 62

~~41# Agent Name~~63Detailed system prompt for the agent describing its role, expertise, and behavior.

~~43Detailed description of the agent's role, expertise, and when Claude should invoke it.~~

~~45## Capabilities~~

~~46- Specific task the agent excels at~~

~~47- Another specialized capability~~

~~48- When to use this agent vs others~~

~~50## Context and examples~~

~~51Provide examples of when this agent should be used and what kinds of problems it solves.~~

52```64```

53 65

54**Integration points**:66**Integration points**:

58* Agents can be invoked manually by users70* Agents can be invoked manually by users

59* Plugin agents work alongside built-in Claude agents71* Plugin agents work alongside built-in Claude agents

60 72

~~61### Skills~~73For complete details, see [Subagents](/en/sub-agents).

~~63Plugins can provide Agent Skills that extend Claude's capabilities. Skills are model-invoked—Claude autonomously decides when to use them based on the task context.~~

~~65**Location**: `skills/` directory in plugin root~~

~~67**File format**: Directories containing `SKILL.md` files with frontmatter~~

~~69**Skill structure**:~~

~~71```~~

~~72skills/~~

~~73├── pdf-processor/~~

~~74│ ├── SKILL.md~~

~~75│ ├── reference.md (optional)~~

~~76│ └── scripts/ (optional)~~

~~77└── code-reviewer/~~

~~78 └── SKILL.md~~

~~79```~~

~~81**Integration behavior**:~~

~~83* Plugin Skills are automatically discovered when the plugin is installed~~

~~84* Claude autonomously invokes Skills based on matching task context~~

~~85* Skills can include supporting files alongside SKILL.md~~

~~87For SKILL.md format and complete Skill authoring guidance, see:~~

~~89* [Use Skills in Claude Code](/en/skills)~~

~~90* [Agent Skills overview](https://docs.claude.com/en/docs/agents-and-tools/agent-skills/overview#skill-structure)~~

91 74

92### Hooks75### Hooks

93 76

130* `SubagentStop`: When a subagent attempts to stop113* `SubagentStop`: When a subagent attempts to stop

131* `SessionStart`: At the beginning of sessions114* `SessionStart`: At the beginning of sessions

132* `SessionEnd`: At the end of sessions115* `SessionEnd`: At the end of sessions

116* `TeammateIdle`: When an agent team teammate is about to go idle

117* `TaskCompleted`: When a task is being marked as completed

133* `PreCompact`: Before conversation history is compacted118* `PreCompact`: Before conversation history is compacted

134 119

135**Hook types**:120**Hook types**:

177### LSP servers162### LSP servers

178 163

179<Tip>164<Tip>

180 Looking to use LSP plugins? Install them from the official marketplace—search for "lsp" in the `/plugin` Discover tab. This section documents how to create LSP plugins for languages not covered by the official marketplace.165 Looking to use LSP plugins? Install them from the official marketplace: search for "lsp" in the `/plugin` Discover tab. This section documents how to create LSP plugins for languages not covered by the official marketplace.

181</Tip>166</Tip>

182 167

183Plugins can provide [Language Server Protocol](https://microsoft.github.io/language-server-protocol/) (LSP) servers to give Claude real-time code intelligence while working on your codebase.168Plugins can provide [Language Server Protocol](https://microsoft.github.io/language-server-protocol/) (LSP) servers to give Claude real-time code intelligence while working on your codebase.

247| `loggingConfig` | Debug logging configuration (see below) |

~~248~~

249**Debug logging configuration:**

~~250~~

251The `loggingConfig` field enables verbose LSP logging when users pass `--enable-lsp-logging`. This helps debug language server issues without impacting normal operation.

~~252~~

253```json theme={null}

254"loggingConfig": {

255 "args": ["--log-level", "4"],

256 "env": {

257 "TSS_LOG": "-level verbose -file ${CLAUDE_PLUGIN_LSP_LOG_FILE}"

258 }

259}

260```

~~261~~

262| Field | Description |

263| :----- | :----------------------------------------------------------------- |

264| `args` | Additional command-line arguments appended when logging is enabled |

265| `env` | Additional environment variables merged when logging is enabled |

~~266~~

267The `${CLAUDE_PLUGIN_LSP_LOG_FILE}` variable expands to the log file path. Logs are written to `~/.claude/debug/`.

268 232

269<Warning>233<Warning>

270 **You must install the language server binary separately.** LSP plugins configure how Claude Code connects to a language server, but they don't include the server itself. If you see `Executable not found in $PATH` in the `/plugin` Errors tab, install the required binary for your language.234 **You must install the language server binary separately.** LSP plugins configure how Claude Code connects to a language server, but they don't include the server itself. If you see `Executable not found in $PATH` in the `/plugin` Errors tab, install the required binary for your language.

295 259

296Plugins use the same scope system as other Claude Code configurations. For installation instructions and scope flags, see [Install plugins](/en/discover-plugins#install-plugins). For a complete explanation of scopes, see [Configuration scopes](/en/settings#configuration-scopes).260Plugins use the same scope system as other Claude Code configurations. For installation instructions and scope flags, see [Install plugins](/en/discover-plugins#install-plugins). For a complete explanation of scopes, see [Configuration scopes](/en/settings#configuration-scopes).

297 261

299 263

300## Plugin manifest schema264## Plugin manifest schema

301 265

302The `plugin.json` file defines your plugin's metadata and configuration. This section documents all supported fields and options.266The `.claude-plugin/plugin.json` file defines your plugin's metadata and configuration. This section documents all supported fields and options.

267

268The manifest is optional. If omitted, Claude Code auto-discovers components in [default locations](#file-locations-reference) and derives the plugin name from the directory name. Use a manifest when you need to provide metadata or custom component paths.

303 269

304### Complete schema270### Complete schema

305 271

329 295

330### Required fields296### Required fields

331 297

298If you include a manifest, `name` is the only required field.

299

333| :----- | :----- | :---------------------------------------- | :------------------- |301| :----- | :----- | :---------------------------------------- | :------------------- |

335 303

304This name is used for namespacing components. For example, in the UI, the

305agent `agent-creator` for the plugin with name `plugin-dev` will appear as

306`plugin-dev:agent-creator`.

307

336### Metadata fields308### Metadata fields

337 309

339| :------------ | :----- | :---------------------------------- | :------------------------------------------------- |311| :------------ | :----- | :-------------------------------------------------------------------------------------------------------------------------- | :------------------------------------------------- |

342| `author` | object | Author information | `{"name": "Dev Team", "email": "dev@company.com"}` |314| `author` | object | Author information | `{"name": "Dev Team", "email": "dev@company.com"}` |

348### Component path fields320### Component path fields

349 321

351| :------------- | :------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------- | :------------------------------------- |323| :------------- | :-------------------- | :-------------------------------------------------------------------------------------------------------------------------------------------------------- | :------------------------------------- |

359 331

360### Path behavior rules332### Path behavior rules

361 333

410 382

411### How plugin caching works383### How plugin caching works

412 384

413When you install a plugin, Claude Code copies the plugin files to a cache directory:385Plugins are specified in one of two ways:

386

387* Through `claude --plugin-dir`, for the duration of a session.

388* Through a marketplace, installed to the local plugin cache.

389

390When you install a plugin, Claude Code locates its marketplace and the plugin's `source` field within that marketplace.

391

392The source can be one of five types:

414 393

415* **For marketplace plugins with relative paths**: The path specified in the `source` field is copied recursively. For example, if your marketplace entry specifies `"source": "./plugins/my-plugin"`, the entire `./plugins` directory is copied.394* Relative path: copied recursively to the plugin cache. For example, if your marketplace entry specifies `"source": "./plugins/my-plugin"`, the entire `./plugins/my-plugin` directory is copied.

416* **For plugins with `.claude-plugin/plugin.json`**: The implicit root directory (the directory containing `.claude-plugin/plugin.json`) is copied recursively.395* npm - copied to the plugin cache from npm

396* pip - copied to the plugin cache from pip

397* url - any https\:// URL ending in .git

398* github - any owner/repo shorthand

417 399

418### Path traversal limitations400### Path traversal limitations

419 401

465 447

466```448```

467enterprise-plugin/449enterprise-plugin/

468├── .claude-plugin/ # Metadata directory450├── .claude-plugin/ # Metadata directory (optional)

469│ └── plugin.json # Required: plugin manifest451│ └── plugin.json # plugin manifest

470├── commands/ # Default command location452├── commands/ # Default command location

471│ ├── status.md453│ ├── status.md

472│ └── logs.md454│ └── logs.md

500### File locations reference482### File locations reference

501 483

503| :-------------- | :--------------------------- | :------------------------------- |485| :-------------- | :--------------------------- | :---------------------------------------------------------- |

536 518

519Scope determines which settings file the installed plugin is added to. For example, --scope project writes to `enabledPlugins` in .claude/settings.json, making the plugin available to everyone who clones the project repository.

520

537**Examples:**521**Examples:**

538 522

539```bash theme={null}523```bash theme={null}

631 615

632### Debugging commands616### Debugging commands

633 617

634Use `claude --debug` to see plugin loading details:618Use `claude --debug` (or `/debug` within the TUI) to see plugin loading details:

~~635~~

636```bash theme={null}

637claude --debug

638```

639 619

640This shows:620This shows:

641 621

667 647

668* `Warning: No commands found in plugin my-plugin custom directory: ./cmds. Expected .md files or SKILL.md in subdirectories.`: command path exists but contains no valid command files648* `Warning: No commands found in plugin my-plugin custom directory: ./cmds. Expected .md files or SKILL.md in subdirectories.`: command path exists but contains no valid command files

669* `Plugin directory not found at path: ./plugins/my-plugin. Check that the marketplace entry has the correct path.`: the `source` path in marketplace.json points to a non-existent directory649* `Plugin directory not found at path: ./plugins/my-plugin. Check that the marketplace entry has the correct path.`: the `source` path in marketplace.json points to a non-existent directory

670* `Plugin my-plugin has conflicting manifests: both plugin.json and marketplace entry specify components.`: remove duplicate component definitions or set `strict: true` in marketplace entry650* `Plugin my-plugin has conflicting manifests: both plugin.json and marketplace entry specify components.`: remove duplicate component definitions or remove `strict: false` in marketplace entry

671 651

672### Hook troubleshooting652### Hook troubleshooting

673 653

756 736

757* [Plugins](/en/plugins) - Tutorials and practical usage737* [Plugins](/en/plugins) - Tutorials and practical usage

758* [Plugin marketplaces](/en/plugin-marketplaces) - Creating and managing marketplaces738* [Plugin marketplaces](/en/plugin-marketplaces) - Creating and managing marketplaces

759* [Slash commands](/en/slash-commands) - Command development details739* [Skills](/en/skills) - Skill development details

760* [Subagents](/en/sub-agents) - Agent configuration and capabilities740* [Subagents](/en/sub-agents) - Agent configuration and capabilities

761* [Agent Skills](/en/skills) - Extend Claude's capabilities

762* [Hooks](/en/hooks) - Event handling and automation741* [Hooks](/en/hooks) - Event handling and automation

763* [MCP](/en/mcp) - External tool integration742* [MCP](/en/mcp) - External tool integration

764* [Settings](/en/settings) - Configuration options for plugins743* [Settings](/en/settings) - Configuration options for plugins

~~765~~

~~766~~

~~767~~

768> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

quickstart.md +71 −72

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Quickstart5# Quickstart

2 6

3> Welcome to Claude Code!7> Welcome to Claude Code!

10 14

11* A terminal or command prompt open15* A terminal or command prompt open

12* A code project to work with16* A code project to work with

~~13* A [Claude.ai](https://claude.ai) (recommended) or [Claude Console](https://console.anthropic.com/) account~~17* A [Claude subscription](https://claude.com/pricing) (Pro, Max, Teams, or Enterprise), [Claude Console](https://console.anthropic.com/) account, or access through a [supported cloud provider](/en/third-party-integrations)

19<Note>

20 This guide covers the terminal CLI. Claude Code is also available on the [web](https://claude.ai/code), as a [desktop app](/en/desktop), in [VS Code](/en/vs-code) and [JetBrains IDEs](/en/jetbrains), in [Slack](/en/slack), and in CI/CD with [GitHub Actions](/en/github-actions) and [GitLab](/en/gitlab-ci-cd). See [all interfaces](/en/overview#use-claude-code-everywhere).

21</Note>

14 22

15## Step 1: Install Claude Code23## Step 1: Install Claude Code

16 24

35 ```batch theme={null}43 ```batch theme={null}

36 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd44 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

37 ```45 ```

47 <Info>

48 Native installations automatically update in the background to keep you on the latest version.

49 </Info>

38 </Tab>50 </Tab>

39 51

40 <Tab title="Homebrew">52 <Tab title="Homebrew">

41 ```sh theme={null}53 ```sh theme={null}

42 brew install --cask claude-code54 brew install --cask claude-code

43 ```55 ```

~~44 </Tab>~~

45 56

~~46 <Tab title="NPM">~~57 <Info>

~~47 If you have [Node.js 18 or newer installed](https://nodejs.org/en/download/):~~58 Homebrew installations do not auto-update. Run `brew upgrade claude-code` periodically to get the latest features and security fixes.

59 </Info>

60 </Tab>

48 61

~~49 ```sh theme={null}~~62 <Tab title="WinGet">

~~50 npm install -g @anthropic-ai/claude-code~~63 ```powershell theme={null}

64 winget install Anthropic.ClaudeCode

51 ```65 ```

67 <Info>

68 WinGet installations do not auto-update. Run `winget upgrade Anthropic.ClaudeCode` periodically to get the latest features and security fixes.

69 </Info>

52 </Tab>70 </Tab>

53</Tabs>71</Tabs>

54 72

66# Follow the prompts to log in with your account84# Follow the prompts to log in with your account

67```85```

68 86

~~69You can log in using either account type:~~87You can log in using any of these account types:

~~71* [Claude.ai](https://claude.ai) (subscription plans - recommended)~~

~~72* [Claude Console](https://console.anthropic.com/) (API access with pre-paid credits)~~

~~74Once logged in, your credentials are stored and you won't need to log in again.~~

75 88

~~76<Note>~~89* [Claude Pro, Max, Teams, or Enterprise](https://claude.com/pricing) (recommended)

77 When you first authenticate Claude Code with your Claude Console account, a workspace called "Claude Code" is automatically created for you. This workspace provides centralized cost tracking and management for all Claude Code usage in your organization.90* [Claude Console](https://console.anthropic.com/) (API access with pre-paid credits). On first login, a "Claude Code" workspace is automatically created in the Console for centralized cost tracking.

~~78</Note>~~91* [Amazon Bedrock, Google Vertex AI, or Microsoft Foundry](/en/third-party-integrations) (enterprise cloud providers)

79 92

~~80<Note>~~93Once logged in, your credentials are stored and you won't need to log in again. To switch accounts later, use the `/login` command.

~~81 You can have both account types under the same email address. If you need to log in again or switch accounts, use the `/login` command within Claude Code.~~

~~82</Note>~~

83 94

84## Step 3: Start your first session95## Step 3: Start your first session

85 96

93You'll see the Claude Code welcome screen with your session information, recent conversations, and latest updates. Type `/help` for available commands or `/resume` to continue a previous conversation.104You'll see the Claude Code welcome screen with your session information, recent conversations, and latest updates. Type `/help` for available commands or `/resume` to continue a previous conversation.

94 105

95<Tip>106<Tip>

~~96 After logging in (Step 2), your credentials are stored on your system. Learn more in [Credential Management](/en/iam#credential-management).~~107 After logging in (Step 2), your credentials are stored on your system. Learn more in [Credential Management](/en/authentication#credential-management).

97</Tip>108</Tip>

98 109

99## Step 4: Ask your first question110## Step 4: Ask your first question

101Let's start with understanding your codebase. Try one of these commands:112Let's start with understanding your codebase. Try one of these commands:

102 113

103```114```

104> what does this project do?115what does this project do?

105```116```

106 117

107Claude will analyze your files and provide a summary. You can also ask more specific questions:118Claude will analyze your files and provide a summary. You can also ask more specific questions:

108 119

109```120```

110> what technologies does this project use?121what technologies does this project use?

111```122```

112 123

113```124```

114> where is the main entry point?125where is the main entry point?

115```126```

116 127

117```128```

118> explain the folder structure129explain the folder structure

119```130```

120 131

121You can also ask Claude about its own capabilities:132You can also ask Claude about its own capabilities:

122 133

123```134```

124> what can Claude Code do?135what can Claude Code do?

125```136```

126 137

127```138```

128> how do I use slash commands in Claude Code?139how do I create custom skills in Claude Code?

129```140```

130 141

131```142```

132> can Claude Code work with Docker?143can Claude Code work with Docker?

133```144```

134 145

135<Note>146<Note>

141Now let's make Claude Code do some actual coding. Try a simple task:152Now let's make Claude Code do some actual coding. Try a simple task:

142 153

143```154```

144> add a hello world function to the main file155add a hello world function to the main file

145```156```

146 157

147Claude Code will:158Claude Code will:

160Claude Code makes Git operations conversational:171Claude Code makes Git operations conversational:

161 172

162```173```

163> what files have I changed?174what files have I changed?

164```175```

165 176

166```177```

167> commit my changes with a descriptive message178commit my changes with a descriptive message

168```179```

169 180

170You can also prompt for more complex Git operations:181You can also prompt for more complex Git operations:

171 182

172```183```

173> create a new branch called feature/quickstart184create a new branch called feature/quickstart

174```185```

175 186

176```187```

177> show me the last 5 commits188show me the last 5 commits

178```189```

179 190

180```191```

181> help me resolve merge conflicts192help me resolve merge conflicts

182```193```

183 194

184## Step 7: Fix a bug or add a feature195## Step 7: Fix a bug or add a feature

188Describe what you want in natural language:199Describe what you want in natural language:

189 200

190```201```

191> add input validation to the user registration form202add input validation to the user registration form

192```203```

193 204

194Or fix existing issues:205Or fix existing issues:

195 206

196```207```

197> there's a bug where users can submit empty forms - fix it208there's a bug where users can submit empty forms - fix it

198```209```

199 210

200Claude Code will:211Claude Code will:

211**Refactor code**222**Refactor code**

212 223

213```224```

214> refactor the authentication module to use async/await instead of callbacks225refactor the authentication module to use async/await instead of callbacks

215```226```

216 227

217**Write tests**228**Write tests**

218 229

219```230```

220> write unit tests for the calculator functions231write unit tests for the calculator functions

221```232```

222 233

223**Update documentation**234**Update documentation**

224 235

225```236```

226> update the README with installation instructions237update the README with installation instructions

227```238```

228 239

229**Code review**240**Code review**

230 241

231```242```

232> review my changes and suggest improvements243review my changes and suggest improvements

233```244```

234 245

235<Tip>246<Tip>

241Here are the most important commands for daily use:252Here are the most important commands for daily use:

242 253

244| ------------------- | --------------------------------- | ----------------------------------- |255| ------------------- | ------------------------------------------------------ | ----------------------------------- |

254 265

255See the [CLI reference](/en/cli-reference) for a complete list of commands.266See the [CLI reference](/en/cli-reference) for a complete list of commands.

256 267

257## Pro tips for beginners268## Pro tips for beginners

258 269

270For more, see [best practices](/en/best-practices) and [common workflows](/en/common-workflows).

271

259<AccordionGroup>272<AccordionGroup>

260 <Accordion title="Be specific with your requests">273 <Accordion title="Be specific with your requests">

261 Instead of: "fix the bug"274 Instead of: "fix the bug"

267 Break complex tasks into steps:280 Break complex tasks into steps:

268 281

269 ```282 ```

270 > 1. create a new database table for user profiles283 1. create a new database table for user profiles

271 ```284 2. create an API endpoint to get and update user profiles

~~272~~ 285 3. build a webpage that allows users to see and edit their information

273 ```

274 > 2. create an API endpoint to get and update user profiles

275 ```

~~276~~

277 ```

278 > 3. build a webpage that allows users to see and edit their information

279 ```286 ```

280 </Accordion>287 </Accordion>

281 288

283 Before making changes, let Claude understand your code:290 Before making changes, let Claude understand your code:

284 291

285 ```292 ```

286 > analyze the database schema293 analyze the database schema

287 ```294 ```

288 295

289 ```296 ```

290 > build a dashboard showing products that are most frequently returned by our UK customers297 build a dashboard showing products that are most frequently returned by our UK customers

291 ```298 ```

292 </Accordion>299 </Accordion>

293 300

295 * Press `?` to see all available keyboard shortcuts302 * Press `?` to see all available keyboard shortcuts

296 * Use Tab for command completion303 * Use Tab for command completion

297 * Press ↑ for command history304 * Press ↑ for command history

298 * Type `/` to see all slash commands305 * Type `/` to see all commands and skills

299 </Accordion>306 </Accordion>

300</AccordionGroup>307</AccordionGroup>

301 308

303 310

304Now that you've learned the basics, explore more advanced features:311Now that you've learned the basics, explore more advanced features:

305 312

306<CardGroup cols={3}>313<CardGroup cols={2}>

307 <Card title="Common workflows" icon="graduation-cap" href="/en/common-workflows">314 <Card title="How Claude Code works" icon="microchip" href="/en/how-claude-code-works">

308 Step-by-step guides for common tasks315 Understand the agentic loop, built-in tools, and how Claude Code interacts with your project

309 </Card>

~~310~~

311 <Card title="CLI reference" icon="terminal" href="/en/cli-reference">

312 Master all commands and options

313 </Card>316 </Card>

314 317

315 <Card title="Configuration" icon="gear" href="/en/settings">318 <Card title="Best practices" icon="star" href="/en/best-practices">

316 Customize Claude Code for your workflow319 Get better results with effective prompting and project setup

317 </Card>320 </Card>

318 321

319 <Card title="Claude Code on the web" icon="cloud" href="/en/claude-code-on-the-web">322 <Card title="Common workflows" icon="graduation-cap" href="/en/common-workflows">

320 Run tasks asynchronously in the cloud323 Step-by-step guides for common tasks

321 </Card>324 </Card>

322 325

323 <Card title="About Claude Code" icon="sparkles" href="https://claude.com/product/claude-code">326 <Card title="Extend Claude Code" icon="puzzle-piece" href="/en/features-overview">

324 Learn more on claude.com327 Customize with CLAUDE.md, skills, hooks, MCP, and more

325 </Card>328 </Card>

326</CardGroup>329</CardGroup>

327 330

330* **In Claude Code**: Type `/help` or ask "how do I..."333* **In Claude Code**: Type `/help` or ask "how do I..."

331* **Documentation**: You're here! Browse other guides334* **Documentation**: You're here! Browse other guides

332* **Community**: Join our [Discord](https://www.anthropic.com/discord) for tips and support335* **Community**: Join our [Discord](https://www.anthropic.com/discord) for tips and support

~~333~~

~~334~~

~~335~~

336> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

sandboxing.md +50 −12

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Sandboxing5# Sandboxing

2 6

3> Learn how Claude Code's sandboxed bash tool provides filesystem and network isolation for safer, more autonomous agent execution.7> Learn how Claude Code's sandboxed bash tool provides filesystem and network isolation for safer, more autonomous agent execution.

51 55

52The sandboxed bash tool leverages operating system security primitives:56The sandboxed bash tool leverages operating system security primitives:

53 57

~~54* **Linux**: Uses [bubblewrap](https://github.com/containers/bubblewrap) for isolation~~

55* **macOS**: Uses Seatbelt for sandbox enforcement58* **macOS**: Uses Seatbelt for sandbox enforcement

59* **Linux**: Uses [bubblewrap](https://github.com/containers/bubblewrap) for isolation

60* **WSL2**: Uses bubblewrap, same as Linux

62WSL1 is not supported because bubblewrap requires kernel features only available in WSL2.

56 63

57These OS-level restrictions ensure that all child processes spawned by Claude Code's commands inherit the same security boundaries.64These OS-level restrictions ensure that all child processes spawned by Claude Code's commands inherit the same security boundaries.

58 65

59## Getting started66## Getting started

60 67

68### Prerequisites

70On **macOS**, sandboxing works out of the box using the built-in Seatbelt framework.

72On **Linux and WSL2**, install the required packages first:

74<Tabs>

75 <Tab title="Ubuntu/Debian">

76 ```bash theme={null}

77 sudo apt-get install bubblewrap socat

78 ```

79 </Tab>

81 <Tab title="Fedora">

82 ```bash theme={null}

83 sudo dnf install bubblewrap socat

84 ```

85 </Tab>

86</Tabs>

61### Enable sandboxing88### Enable sandboxing

62 89

~~63You can enable sandboxing by running the `/sandbox` slash command:~~90You can enable sandboxing by running the `/sandbox` command:

64 91

65```92```

66> /sandbox93> /sandbox

67```94```

68 95

~~69This opens a menu where you can choose between sandbox modes.~~96This opens a menu where you can choose between sandbox modes. If required dependencies are missing (such as `bubblewrap` or `socat` on Linux), the menu displays installation instructions for your platform.

70 97

71### Sandbox modes98### Sandbox modes

72 99

110 137

111* Cannot modify critical config files such as `~/.bashrc`138* Cannot modify critical config files such as `~/.bashrc`

112* Cannot modify system-level files in `/bin/`139* Cannot modify system-level files in `/bin/`

113* Cannot read files that are denied in your [Claude permission settings](/en/iam#configuring-permissions)140* Cannot read files that are denied in your [Claude permission settings](/en/permissions#manage-permissions)

114 141

115**Network protection:**142**Network protection:**

116 143

157* Filesystem Permission Escalation: Overly broad filesystem write permissions can enable privilege escalation attacks. Allowing writes to directories containing executables in `$PATH`, system configuration directories, or user shell configuration files (`.bashrc`, `.zshrc`) can lead to code execution in different security contexts when other users or system processes access these files.184* Filesystem Permission Escalation: Overly broad filesystem write permissions can enable privilege escalation attacks. Allowing writes to directories containing executables in `$PATH`, system configuration directories, or user shell configuration files (`.bashrc`, `.zshrc`) can lead to code execution in different security contexts when other users or system processes access these files.

158* Linux Sandbox Strength: The Linux implementation provides strong filesystem and network isolation but includes an `enableWeakerNestedSandbox` mode that enables it to work inside of Docker environments without privileged namespaces. This option considerably weakens security and should only be used in cases where additional isolation is otherwise enforced.185* Linux Sandbox Strength: The Linux implementation provides strong filesystem and network isolation but includes an `enableWeakerNestedSandbox` mode that enables it to work inside of Docker environments without privileged namespaces. This option considerably weakens security and should only be used in cases where additional isolation is otherwise enforced.

159 186

187## How sandboxing relates to permissions

188

189Sandboxing and [permissions](/en/permissions) are complementary security layers that work together:

190

191* **Permissions** control which tools Claude Code can use and are evaluated before any tool runs. They apply to all tools: Bash, Read, Edit, WebFetch, MCP, and others.

192* **Sandboxing** provides OS-level enforcement that restricts what Bash commands can access at the filesystem and network level. It applies only to Bash commands and their child processes.

193

194Filesystem and network restrictions are configured through permission rules, not sandbox settings:

195

196* Use `Read` and `Edit` deny rules to block access to specific files or directories

197* Use `WebFetch` allow/deny rules to control domain access

198* Use sandbox `allowedDomains` to control which domains Bash commands can reach

199

200This [repository](https://github.com/anthropics/claude-code/tree/main/examples/settings) includes starter settings configurations for common deployment scenarios, including sandbox-specific examples. Use these as starting points and adjust them to fit your needs.

201

160## Advanced usage202## Advanced usage

161 203

162### Custom proxy configuration204### Custom proxy configuration

183 225

184The sandboxed bash tool works alongside:226The sandboxed bash tool works alongside:

185 227

186* **IAM policies**: Combine with [permission settings](/en/iam) for defense-in-depth228* **Permission rules**: Combine with [permission settings](/en/permissions) for defense-in-depth

187* **Development containers**: Use with [devcontainers](/en/devcontainer) for additional isolation229* **Development containers**: Use with [devcontainers](/en/devcontainer) for additional isolation

188* **Enterprise policies**: Enforce sandbox configurations through [managed settings](/en/settings#settings-precedence)230* **Enterprise policies**: Enforce sandbox configurations through [managed settings](/en/settings#settings-precedence)

189 231

209 251

210* **Performance overhead**: Minimal, but some filesystem operations may be slightly slower252* **Performance overhead**: Minimal, but some filesystem operations may be slightly slower

211* **Compatibility**: Some tools that require specific system access patterns may need configuration adjustments, or may even need to be run outside of the sandbox253* **Compatibility**: Some tools that require specific system access patterns may need configuration adjustments, or may even need to be run outside of the sandbox

212* **Platform support**: Currently supports Linux and macOS; Windows support planned254* **Platform support**: Supports macOS, Linux, and WSL2. WSL1 is not supported. Native Windows support is planned.

213 255

214## See also256## See also

215 257

216* [Security](/en/security) - Comprehensive security features and best practices258* [Security](/en/security) - Comprehensive security features and best practices

217* [IAM](/en/iam) - Permission configuration and access control259* [Permissions](/en/permissions) - Permission configuration and access control

218* [Settings](/en/settings) - Complete configuration reference260* [Settings](/en/settings) - Complete configuration reference

219* [CLI reference](/en/cli-reference) - Command-line options including `-sb`261* [CLI reference](/en/cli-reference) - Command-line options

~~220~~

~~221~~

~~222~~

223> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

security.md +10 −10

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Security5# Security

2 6

3> Learn about Claude Code's security safeguards and best practices for safe usage.7> Learn about Claude Code's security safeguards and best practices for safe usage.

14 18

15We designed Claude Code to be transparent and secure. For example, we require approval for bash commands before executing them, giving you direct control. This approach enables users and organizations to configure permissions directly.19We designed Claude Code to be transparent and secure. For example, we require approval for bash commands before executing them, giving you direct control. This approach enables users and organizations to configure permissions directly.

16 20

~~17For detailed permission configuration, see [Identity and Access Management](/en/iam).~~21For detailed permission configuration, see [Permissions](/en/permissions).

18 22

19### Built-in protections23### Built-in protections

20 24

38* **Permission system**: Sensitive operations require explicit approval42* **Permission system**: Sensitive operations require explicit approval

39* **Context-aware analysis**: Detects potentially harmful instructions by analyzing the full request43* **Context-aware analysis**: Detects potentially harmful instructions by analyzing the full request

40* **Input sanitization**: Prevents command injection by processing user inputs44* **Input sanitization**: Prevents command injection by processing user inputs

41* **Command blocklist**: Blocks risky commands that fetch arbitrary content from the web like `curl` and `wget` by default. When explicitly allowed, be aware of [permission pattern limitations](/en/iam#tool-specific-permission-rules)45* **Command blocklist**: Blocks risky commands that fetch arbitrary content from the web like `curl` and `wget` by default. When explicitly allowed, be aware of [permission pattern limitations](/en/permissions#tool-specific-permission-rules)

42 46

43### Privacy safeguards47### Privacy safeguards

44 48

59* **Command injection detection**: Suspicious bash commands require manual approval even if previously allowlisted63* **Command injection detection**: Suspicious bash commands require manual approval even if previously allowlisted

60* **Fail-closed matching**: Unmatched commands default to requiring manual approval64* **Fail-closed matching**: Unmatched commands default to requiring manual approval

61* **Natural language descriptions**: Complex bash commands include explanations for user understanding65* **Natural language descriptions**: Complex bash commands include explanations for user understanding

~~62* **Secure credential storage**: API keys and tokens are encrypted. See [Credential Management](/en/iam#credential-management)~~66* **Secure credential storage**: API keys and tokens are encrypted. See [Credential Management](/en/authentication#credential-management)

63 67

64<Warning>68<Warning>

65 **Windows WebDAV security risk**: When running Claude Code on Windows, we recommend against enabling WebDAV or allowing Claude Code to access paths such as `\\*` that may contain WebDAV subdirectories. [WebDAV has been deprecated by Microsoft](https://learn.microsoft.com/en-us/windows/whats-new/deprecated-features#:~:text=The%20Webclient%20$WebDAV$%20service%20is%20deprecated) due to security risks. Enabling WebDAV may allow Claude Code to trigger network requests to remote hosts, bypassing the permission system.69 **Windows WebDAV security risk**: When running Claude Code on Windows, we recommend against enabling WebDAV or allowing Claude Code to access paths such as `\\*` that may contain WebDAV subdirectories. [WebDAV has been deprecated by Microsoft](https://learn.microsoft.com/en-us/windows/whats-new/deprecated-features#:~:text=The%20Webclient%20$WebDAV$%20service%20is%20deprecated) due to security risks. Enabling WebDAV may allow Claude Code to trigger network requests to remote hosts, bypassing the permission system.

87 91

88## IDE security92## IDE security

89 93

~~90See [here](/en/vs-code#security) for more information on the security of running Claude Code in an IDE.~~94See [VS Code security and privacy](/en/vs-code#security-and-privacy) for more information on running Claude Code in an IDE.

91 95

92## Cloud execution security96## Cloud execution security

93 97

113 117

114### Team security118### Team security

115 119

116* Use [enterprise managed settings](/en/iam#enterprise-managed-settings) to enforce organizational standards120* Use [managed settings](/en/permissions#managed-settings) to enforce organizational standards

117* Share approved permission configurations through version control121* Share approved permission configurations through version control

118* Train team members on security best practices122* Train team members on security best practices

119* Monitor Claude Code usage through [OpenTelemetry metrics](/en/monitoring-usage)123* Monitor Claude Code usage through [OpenTelemetry metrics](/en/monitoring-usage)

130## Related resources134## Related resources

131 135

132* [Sandboxing](/en/sandboxing) - Filesystem and network isolation for bash commands136* [Sandboxing](/en/sandboxing) - Filesystem and network isolation for bash commands

133* [Identity and Access Management](/en/iam) - Configure permissions and access controls137* [Permissions](/en/permissions) - Configure permissions and access controls

134* [Monitoring usage](/en/monitoring-usage) - Track and audit Claude Code activity138* [Monitoring usage](/en/monitoring-usage) - Track and audit Claude Code activity

135* [Development containers](/en/devcontainer) - Secure, isolated environments139* [Development containers](/en/devcontainer) - Secure, isolated environments

136* [Anthropic Trust Center](https://trust.anthropic.com) - Security certifications and compliance140* [Anthropic Trust Center](https://trust.anthropic.com) - Security certifications and compliance

~~137~~

~~138~~

~~139~~

140> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

server-managed-settings.md +162 −0 added

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# Configure server-managed settings (public beta)

7> Centrally configure Claude Code for your organization through server-delivered settings, without requiring device management infrastructure.

9Server-managed settings allow administrators to centrally configure Claude Code through a web-based interface on Claude.ai. Claude Code clients automatically receive these settings when users authenticate with their organization credentials.

11This approach is designed for organizations that do not have device management infrastructure in place, or need to manage settings for users on unmanaged devices.

13<Note>

14 Server-managed settings are in public beta and available for [Claude for Teams](https://claude.com/pricing#team-&-enterprise) and [Claude for Enterprise](https://anthropic.com/contact-sales) customers. Features may evolve before general availability.

15</Note>

17## Requirements

19To use server-managed settings, you need:

21* Claude for Teams or Claude for Enterprise plan

22* Claude Code version 2.1.38 or later for Claude for Teams, or version 2.1.30 or later for Claude for Enterprise

23* Network access to `api.anthropic.com`

25## Choose between server-managed and endpoint-managed settings

27Claude Code supports two approaches for centralized configuration. Server-managed settings deliver configuration from Anthropic's servers. [Endpoint-managed settings](/en/permissions#managed-settings) deploy a `managed-settings.json` file to system directories via MDM (mobile device management).

29| Approach | Best for | Security model |

30| :---------------------------------------------------------------- | :------------------------------------------------------- | :----------------------------------------------------------------- |

31| **Server-managed settings** | Organizations without MDM, or users on unmanaged devices | Settings delivered from Anthropic's servers at authentication time |

32| **[Endpoint-managed settings](/en/permissions#managed-settings)** | Organizations with MDM or endpoint management | Settings deployed to protected system directories by IT |

34If your devices are enrolled in an MDM or endpoint management solution, endpoint-managed settings provide stronger security guarantees because the settings file can be protected from user modification at the OS level.

36## Configure server-managed settings

38<Steps>

39 <Step title="Open the admin console">

40 In [Claude.ai](https://claude.ai), navigate to **Admin Settings > Claude Code > Managed settings**.

41 </Step>

43 <Step title="Define your settings">

44 Add your configuration as JSON. All [settings available in `settings.json`](/en/settings#available-settings) are supported, including [managed-only settings](/en/permissions#managed-only-settings) like `disableBypassPermissionsMode`.

46 This example enforces a permission deny list and prevents users from bypassing permissions:

48 ```json theme={null}

49 {

50 "permissions": {

51 "deny": [

52 "Bash(curl *)",

53 "Read(./.env)",

54 "Read(./.env.*)",

55 "Read(./secrets/**)"

56 ]

57 },

58 "disableBypassPermissionsMode": "disable"

59 }

60 ```

61 </Step>

63 <Step title="Save and deploy">

64 Save your changes. Claude Code clients receive the updated settings on their next startup or hourly polling cycle.

65 </Step>

66</Steps>

68### Verify settings delivery

70To confirm that settings are being applied, ask a user to restart Claude Code. If the configuration includes settings that trigger the [security approval dialog](#security-approval-dialogs), the user sees a prompt describing the managed settings on startup. You can also verify that managed permission rules are active by having a user run `/permissions` to view their effective permission rules.

72### Access control

74The following roles can manage server-managed settings:

76* **Primary Owner**

77* **Owner**

79Restrict access to trusted personnel, as settings changes apply to all users in the organization.

81### Current limitations

83Server-managed settings have the following limitations during the beta period:

85* Settings apply uniformly to all users in the organization. Per-group configurations are not yet supported.

86* [MCP server configurations](/en/mcp#managed-mcp-configuration) cannot be distributed through server-managed settings.

88## Settings delivery

90### Settings precedence

92Server-managed settings and [endpoint-managed settings](/en/permissions#managed-settings) both occupy the highest tier in the Claude Code [settings hierarchy](/en/settings#settings-precedence), and user or project settings cannot override them. When both are present, server-managed settings take precedence and the local `managed-settings.json` file is not used.

94### Fetch and caching behavior

96Claude Code fetches settings from Anthropic's servers at startup and polls for updates hourly during active sessions.

98**First launch without cached settings:**

100* Claude Code fetches settings asynchronously

101* If the fetch fails, Claude Code continues without managed settings

102* There is a brief window before settings load where restrictions are not yet enforced

103

104**Subsequent launches with cached settings:**

105

106* Cached settings apply immediately at startup

107* Claude Code fetches fresh settings in the background

108* Cached settings persist through network failures

109

110Claude Code applies settings updates automatically without a restart, except for advanced settings like OpenTelemetry configuration, which require a full restart to take effect.

111

112### Security approval dialogs

113

114Certain settings that could pose security risks require explicit user approval before being applied:

115

116* **Shell command settings**: settings that execute shell commands

117* **Custom environment variables**: variables not in the known safe allowlist

118* **Hook configurations**: any hook definition

119

120When these settings are present, users see a security dialog explaining what is being configured. Users must approve to proceed. If a user rejects the settings, Claude Code exits.

121

122<Note>

123 In non-interactive mode with the `-p` flag, Claude Code skips security dialogs and applies settings without user approval.

124</Note>

125

126## Platform availability

127

128Server-managed settings require a direct connection to `api.anthropic.com` and are not available when using third-party model providers:

129

130* Amazon Bedrock

131* Google Vertex AI

132* Microsoft Foundry

133* Custom API endpoints via `ANTHROPIC_BASE_URL` or [LLM gateways](/en/llm-gateway)

134

135## Audit logging

136

137Audit log events for settings changes are available through the compliance API or audit log export. Contact your Anthropic account team for access.

138

139Audit events include the type of action performed, the account and device that performed the action, and references to the previous and new values.

140

141## Security considerations

142

143Server-managed settings provide centralized policy enforcement, but they operate as a client-side control. On unmanaged devices, users with admin or sudo access can modify the Claude Code binary, filesystem, or network configuration.

144

145| Scenario | Behavior |

146| :----------------------------------------------- | :-------------------------------------------------------------------------------------------------------------- |

147| User edits the cached settings file | Tampered file applies at startup, but correct settings restore on the next server fetch |

148| User deletes the cached settings file | First-launch behavior occurs: settings fetch asynchronously with a brief unenforced window |

149| API is unavailable | Cached settings apply if available, otherwise managed settings are not enforced until the next successful fetch |

150| User authenticates with a different organization | Settings are not delivered for accounts outside the managed organization |

151| User sets a non-default `ANTHROPIC_BASE_URL` | Server-managed settings are bypassed when using third-party API providers |

152

153For stronger enforcement guarantees, use [endpoint-managed settings](/en/permissions#managed-settings) on devices enrolled in an MDM solution.

154

155## See also

156

157Related pages for managing Claude Code configuration:

158

159* [Settings](/en/settings): complete configuration reference including all available settings

160* [Endpoint-managed settings](/en/permissions#managed-settings): file-based managed settings deployed by IT

161* [Authentication](/en/authentication): set up user access to Claude Code

162* [Security](/en/security): security safeguards and best practices

settings.md +228 −133

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code settings5# Claude Code settings

2 6

3> Configure Claude Code with global and project-level settings, and environment variables.7> Configure Claude Code with global and project-level settings, and environment variables.

11### Available scopes15### Available scopes

12 16

~~14| :------------- | :----------------------------------- | :----------------------------------- | :--------------------- |~~18| :---------- | :----------------------------------- | :----------------------------------- | :--------------------- |

16| **User** | `~/.claude/` directory | You, across all projects | No |20| **User** | `~/.claude/` directory | You, across all projects | No |

18| **Local** | `.claude/*.local.*` files | You, in this repository only | No (gitignored) |22| **Local** | `.claude/*.local.*` files | You, in this repository only | No (gitignored) |

19 23

20### When to use each scope24### When to use each scope

21 25

~~22**Enterprise scope** is for:~~26**Managed scope** is for:

23 27

24* Security policies that must be enforced organization-wide28* Security policies that must be enforced organization-wide

25* Compliance requirements that can't be overridden29* Compliance requirements that can't be overridden

47 51

48When the same setting is configured in multiple scopes, more specific scopes take precedence:52When the same setting is configured in multiple scopes, more specific scopes take precedence:

49 53

~~501. **Enterprise** (highest) - can't be overridden by anything~~541. **Managed** (highest) - can't be overridden by anything

512. **Command line arguments** - temporary session overrides552. **Command line arguments** - temporary session overrides

523. **Local** - overrides project and user settings563. **Local** - overrides project and user settings

534. **Project** - overrides user settings574. **Project** - overrides user settings

79* **Project settings** are saved in your project directory:83* **Project settings** are saved in your project directory:

80 * `.claude/settings.json` for settings that are checked into source control and shared with your team84 * `.claude/settings.json` for settings that are checked into source control and shared with your team

81 * `.claude/settings.local.json` for settings that are not checked in, useful for personal preferences and experimentation. Claude Code will configure git to ignore `.claude/settings.local.json` when it is created.85 * `.claude/settings.local.json` for settings that are not checked in, useful for personal preferences and experimentation. Claude Code will configure git to ignore `.claude/settings.local.json` when it is created.

82* **Managed settings** (Enterprise): Enterprise administrators can configure and distribute Claude Code settings to their organization through the [Claude.ai admin console](https://claude.ai/admin-settings/claude-code). These settings are fetched automatically when users authenticate, take precedence over user and project settings, and cannot be overridden locally. This feature is available to Claude for Enterprise customers. If you don't see this option in your admin console, contact your Anthropic account team to have the feature enabled.86* **Managed settings**: For organizations that need centralized control, Claude Code supports `managed-settings.json` and `managed-mcp.json` files that can be deployed to system directories:

~~84 For organizations that prefer file-based policy distribution, Claude Code also supports `managed-settings.json` and `managed-mcp.json` files that can be deployed to system directories:~~

85 87

86 * macOS: `/Library/Application Support/ClaudeCode/`88 * macOS: `/Library/Application Support/ClaudeCode/`

87 * Linux and WSL: `/etc/claude-code/`89 * Linux and WSL: `/etc/claude-code/`

91 These are system-wide paths (not user home directories like `~/Library/...`) that require administrator privileges. They are designed to be deployed by IT administrators.93 These are system-wide paths (not user home directories like `~/Library/...`) that require administrator privileges. They are designed to be deployed by IT administrators.

92 </Note>94 </Note>

93 95

~~94 See [Enterprise managed settings](/en/iam#enterprise-managed-settings) and [Enterprise MCP configuration](/en/mcp#enterprise-mcp-configuration) for details.~~96 See [Managed settings](/en/permissions#managed-settings) and [Managed MCP configuration](/en/mcp#managed-mcp-configuration) for details. For organizations without device management infrastructure, see [server-managed settings](/en/server-managed-settings).

95 97

96 <Note>98 <Note>

~~97 Enterprise deployments can also restrict **plugin marketplace additions** using~~99 Managed deployments can also restrict **plugin marketplace additions** using

~~98 `strictKnownMarketplaces`. For more information, see [Enterprise marketplace restrictions](/en/plugin-marketplaces#enterprise-marketplace-restrictions).~~100 `strictKnownMarketplaces`. For more information, see [Managed marketplace restrictions](/en/plugin-marketplaces#managed-marketplace-restrictions).

99 </Note>101 </Note>

100* **Other configuration** is stored in `~/.claude.json`. This file contains your preferences (theme, notification settings, editor mode), OAuth session, [MCP server](/en/mcp) configurations for user and local scopes, per-project state (allowed tools, trust settings), and various caches. Project-scoped MCP servers are stored separately in `.mcp.json`.102* **Other configuration** is stored in `~/.claude.json`. This file contains your preferences (theme, notification settings, editor mode), OAuth session, [MCP server](/en/mcp) configurations for user and local scopes, per-project state (allowed tools, trust settings), and various caches. Project-scoped MCP servers are stored separately in `.mcp.json`.

101 103

104<Note>

105 Claude Code automatically creates timestamped backups of configuration files and retains the five most recent backups to prevent data loss.

106</Note>

107

102```JSON Example settings.json theme={null}108```JSON Example settings.json theme={null}

103{109{

110 "$schema": "https://json.schemastore.org/claude-code-settings.json",

104 "permissions": {111 "permissions": {

105 "allow": [112 "allow": [

106 "Bash(npm run lint)",113 "Bash(npm run lint)",

107 "Bash(npm run test:*)",114 "Bash(npm run test *)",

108 "Read(~/.zshrc)"115 "Read(~/.zshrc)"

109 ],116 ],

110 "deny": [117 "deny": [

111 "Bash(curl:*)",118 "Bash(curl *)",

112 "Read(./.env)",119 "Read(./.env)",

113 "Read(./.env.*)",120 "Read(./.env.*)",

114 "Read(./secrets/**)"121 "Read(./secrets/**)"

126}133}

127```134```

128 135

136The `$schema` line in the example above points to the [official JSON schema](https://json.schemastore.org/claude-code-settings.json) for Claude Code settings. Adding it to your `settings.json` enables autocomplete and inline validation in VS Code, Cursor, and any other editor that supports JSON schema validation.

137

129### Available settings138### Available settings

130 139

131`settings.json` supports a number of options:140`settings.json` supports a number of options:

132 141

133| Key | Description | Example |142| Key | Description | Example |

134| :--------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | :---------------------------------------------------------------------- |143| :-------------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | :---------------------------------------------------------------------- |

135| `apiKeyHelper` | Custom script, to be executed in `/bin/sh`, to generate an auth value. This value will be sent as `X-Api-Key` and `Authorization: Bearer` headers for model requests | `/bin/generate_temp_api_key.sh` |144| `apiKeyHelper` | Custom script, to be executed in `/bin/sh`, to generate an auth value. This value will be sent as `X-Api-Key` and `Authorization: Bearer` headers for model requests | `/bin/generate_temp_api_key.sh` |

136| `cleanupPeriodDays` | Sessions inactive for longer than this period are deleted at startup. Setting to `0` immediately deletes all sessions. (default: 30 days) | `20` |145| `cleanupPeriodDays` | Sessions inactive for longer than this period are deleted at startup. Setting to `0` immediately deletes all sessions. (default: 30 days) | `20` |

137| `companyAnnouncements` | Announcement to display to users at startup. If multiple announcements are provided, they will be cycled through at random. | `["Welcome to Acme Corp! Review our code guidelines at docs.acme.com"]` |146| `companyAnnouncements` | Announcement to display to users at startup. If multiple announcements are provided, they will be cycled through at random. | `["Welcome to Acme Corp! Review our code guidelines at docs.acme.com"]` |

139| `attribution` | Customize attribution for git commits and pull requests. See [Attribution settings](#attribution-settings) | `{"commit": "🤖 Generated with Claude Code", "pr": ""}` |148| `attribution` | Customize attribution for git commits and pull requests. See [Attribution settings](#attribution-settings) | `{"commit": "🤖 Generated with Claude Code", "pr": ""}` |

140| `includeCoAuthoredBy` | **Deprecated**: Use `attribution` instead. Whether to include the `co-authored-by Claude` byline in git commits and pull requests (default: `true`) | `false` |149| `includeCoAuthoredBy` | **Deprecated**: Use `attribution` instead. Whether to include the `co-authored-by Claude` byline in git commits and pull requests (default: `true`) | `false` |

142| `hooks` | Configure custom commands to run before or after tool executions. See [hooks documentation](/en/hooks) | `{"PreToolUse": {"Bash": "echo 'Running command...'"}}` |151| `hooks` | Configure custom commands to run at lifecycle events. See [hooks documentation](/en/hooks) for format | See [hooks](/en/hooks) |

144| `allowManagedHooksOnly` | (Enterprise) Prevent loading of user, project, and plugin hooks. Only allows managed hooks and SDK hooks. See [Hook configuration](#hook-configuration) | `true` |153| `allowManagedHooksOnly` | (Managed settings only) Prevent loading of user, project, and plugin hooks. Only allows managed hooks and SDK hooks. See [Hook configuration](#hook-configuration) | `true` |

154| `allowManagedPermissionRulesOnly` | (Managed settings only) Prevent user and project settings from defining `allow`, `ask`, or `deny` permission rules. Only rules in managed settings apply. See [Managed-only settings](/en/permissions#managed-only-settings) | `true` |

146| `otelHeadersHelper` | Script to generate dynamic OpenTelemetry headers. Runs at startup and periodically (see [Dynamic headers](/en/monitoring-usage#dynamic-headers)) | `/bin/generate_otel_headers.sh` |156| `otelHeadersHelper` | Script to generate dynamic OpenTelemetry headers. Runs at startup and periodically (see [Dynamic headers](/en/monitoring-usage#dynamic-headers)) | `/bin/generate_otel_headers.sh` |

147| `statusLine` | Configure a custom status line to display context. See [`statusLine` documentation](/en/statusline) | `{"type": "command", "command": "~/.claude/statusline.sh"}` |157| `statusLine` | Configure a custom status line to display context. See [`statusLine` documentation](/en/statusline) | `{"type": "command", "command": "~/.claude/statusline.sh"}` |

148| `fileSuggestion` | Configure a custom script for `@` file autocomplete. See [File suggestion settings](#file-suggestion-settings) | `{"type": "command", "command": "~/.claude/file-suggestion.sh"}` |158| `fileSuggestion` | Configure a custom script for `@` file autocomplete. See [File suggestion settings](#file-suggestion-settings) | `{"type": "command", "command": "~/.claude/file-suggestion.sh"}` |

159| `respectGitignore` | Control whether the `@` file picker respects `.gitignore` patterns. When `true` (default), files matching `.gitignore` patterns are excluded from suggestions | `false` |

150| `forceLoginMethod` | Use `claudeai` to restrict login to Claude.ai accounts, `console` to restrict login to Claude Console (API usage billing) accounts | `claudeai` |161| `forceLoginMethod` | Use `claudeai` to restrict login to Claude.ai accounts, `console` to restrict login to Claude Console (API usage billing) accounts | `claudeai` |

151| `forceLoginOrgUUID` | Specify the UUID of an organization to automatically select it during login, bypassing the organization selection step. Requires `forceLoginMethod` to be set | `"xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"` |162| `forceLoginOrgUUID` | Specify the UUID of an organization to automatically select it during login, bypassing the organization selection step. Requires `forceLoginMethod` to be set | `"xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"` |

155| `allowedMcpServers` | When set in managed-settings.json, allowlist of MCP servers users can configure. Undefined = no restrictions, empty array = lockdown. Applies to all scopes. Denylist takes precedence. See [Enterprise MCP configuration](/en/mcp#enterprise-mcp-configuration) | `[{ "serverName": "github" }]` |166| `allowedMcpServers` | When set in managed-settings.json, allowlist of MCP servers users can configure. Undefined = no restrictions, empty array = lockdown. Applies to all scopes. Denylist takes precedence. See [Managed MCP configuration](/en/mcp#managed-mcp-configuration) | `[{ "serverName": "github" }]` |

156| `deniedMcpServers` | When set in managed-settings.json, denylist of MCP servers that are explicitly blocked. Applies to all scopes including enterprise servers. Denylist takes precedence over allowlist. See [Enterprise MCP configuration](/en/mcp#enterprise-mcp-configuration) | `[{ "serverName": "filesystem" }]` |167| `deniedMcpServers` | When set in managed-settings.json, denylist of MCP servers that are explicitly blocked. Applies to all scopes including managed servers. Denylist takes precedence over allowlist. See [Managed MCP configuration](/en/mcp#managed-mcp-configuration) | `[{ "serverName": "filesystem" }]` |

157| `strictKnownMarketplaces` | When set in managed-settings.json, allowlist of plugin marketplaces users can add. Undefined = no restrictions, empty array = lockdown. Applies to marketplace additions only. See [Enterprise marketplace restrictions](/en/plugin-marketplaces#enterprise-marketplace-restrictions) | `[{ "source": "github", "repo": "acme-corp/plugins" }]` |168| `strictKnownMarketplaces` | When set in managed-settings.json, allowlist of plugin marketplaces users can add. Undefined = no restrictions, empty array = lockdown. Applies to marketplace additions only. See [Managed marketplace restrictions](/en/plugin-marketplaces#managed-marketplace-restrictions) | `[{ "source": "github", "repo": "acme-corp/plugins" }]` |

158| `awsAuthRefresh` | Custom script that modifies the `.aws` directory (see [advanced credential configuration](/en/amazon-bedrock#advanced-credential-configuration)) | `aws sso login --profile myprofile` |169| `awsAuthRefresh` | Custom script that modifies the `.aws` directory (see [advanced credential configuration](/en/amazon-bedrock#advanced-credential-configuration)) | `aws sso login --profile myprofile` |

159| `awsCredentialExport` | Custom script that outputs JSON with AWS credentials (see [advanced credential configuration](/en/amazon-bedrock#advanced-credential-configuration)) | `/bin/generate_aws_grant.sh` |170| `awsCredentialExport` | Custom script that outputs JSON with AWS credentials (see [advanced credential configuration](/en/amazon-bedrock#advanced-credential-configuration)) | `/bin/generate_aws_grant.sh` |

160| `alwaysThinkingEnabled` | Enable [extended thinking](/en/common-workflows#use-extended-thinking) by default for all sessions. Typically configured via the `/config` command rather than editing directly | `true` |171| `alwaysThinkingEnabled` | Enable [extended thinking](/en/common-workflows#use-extended-thinking-thinking-mode) by default for all sessions. Typically configured via the `/config` command rather than editing directly | `true` |

172| `plansDirectory` | Customize where plan files are stored. Path is relative to project root. Default: `~/.claude/plans` | `"./plans"` |

173| `showTurnDuration` | Show turn duration messages after responses (e.g., "Cooked for 1m 6s"). Set to `false` to hide these messages | `true` |

174| `spinnerVerbs` | Customize the action verbs shown in the spinner and turn duration messages. Set `mode` to `"replace"` to use only your verbs, or `"append"` to add them to the defaults | `{"mode": "append", "verbs": ["Pondering", "Crafting"]}` |

175| `language` | Configure Claude's preferred response language (e.g., `"japanese"`, `"spanish"`, `"french"`). Claude will respond in this language by default | `"japanese"` |

176| `autoUpdatesChannel` | Release channel to follow for updates. Use `"stable"` for a version that is typically about one week old and skips versions with major regressions, or `"latest"` (default) for the most recent release | `"stable"` |

177| `spinnerTipsEnabled` | Show tips in the spinner while Claude is working. Set to `false` to disable tips (default: `true`) | `false` |

178| `terminalProgressBarEnabled` | Enable the terminal progress bar that shows progress in supported terminals like Windows Terminal and iTerm2 (default: `true`) | `false` |

179| `prefersReducedMotion` | Reduce or disable UI animations (spinners, shimmer, flash effects) for accessibility | `true` |

180| `teammateMode` | How [agent team](/en/agent-teams) teammates display: `auto` (picks split panes in tmux or iTerm2, in-process otherwise), `in-process`, or `tmux`. See [set up agent teams](/en/agent-teams#set-up-agent-teams) | `"in-process"` |

161 181

162### Permission settings182### Permission settings

163 183

165| :----------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | :--------------------------------------------------------------------- |185| :----------------------------- | :----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | :--------------------------------------------------------------------- |

166| `allow` | Array of [permission rules](/en/iam#configuring-permissions) to allow tool use. **Note:** Bash rules use prefix matching, not regex | `[ "Bash(git diff:*)" ]` |186| `allow` | Array of permission rules to allow tool use. See [Permission rule syntax](#permission-rule-syntax) below for pattern matching details | `[ "Bash(git diff *)" ]` |

167| `ask` | Array of [permission rules](/en/iam#configuring-permissions) to ask for confirmation upon tool use. | `[ "Bash(git push:*)" ]` |187| `ask` | Array of permission rules to ask for confirmation upon tool use. See [Permission rule syntax](#permission-rule-syntax) below | `[ "Bash(git push *)" ]` |

168| `deny` | Array of [permission rules](/en/iam#configuring-permissions) to deny tool use. Use this to also exclude sensitive files from Claude Code access. **Note:** Bash patterns are prefix matches and can be bypassed (see [Bash permission limitations](/en/iam#tool-specific-permission-rules)) | `[ "WebFetch", "Bash(curl:*)", "Read(./.env)", "Read(./secrets/**)" ]` |188| `deny` | Array of permission rules to deny tool use. Use this to exclude sensitive files from Claude Code access. See [Permission rule syntax](#permission-rule-syntax) and [Bash permission limitations](/en/permissions#tool-specific-permission-rules) | `[ "WebFetch", "Bash(curl *)", "Read(./.env)", "Read(./secrets/**)" ]` |

171| `disableBypassPermissionsMode` | Set to `"disable"` to prevent `bypassPermissions` mode from being activated. This disables the `--dangerously-skip-permissions` command-line flag. See [managed settings](/en/iam#enterprise-managed-settings) | `"disable"` |191| `disableBypassPermissionsMode` | Set to `"disable"` to prevent `bypassPermissions` mode from being activated. This disables the `--dangerously-skip-permissions` command-line flag. See [managed settings](/en/permissions#managed-settings) | `"disable"` |

192

193### Permission rule syntax

194

195Permission rules follow the format `Tool` or `Tool(specifier)`. Rules are evaluated in order: deny rules first, then ask, then allow. The first matching rule wins.

196

197Quick examples:

198

199| Rule | Effect |

200| :----------------------------- | :--------------------------------------- |

201| `Bash` | Matches all Bash commands |

202| `Bash(npm run *)` | Matches commands starting with `npm run` |

203| `Read(./.env)` | Matches reading the `.env` file |

204| `WebFetch(domain:example.com)` | Matches fetch requests to example.com |

205

206For the complete rule syntax reference, including wildcard behavior, tool-specific patterns for Read, Edit, WebFetch, MCP, and Task rules, and security limitations of Bash patterns, see [Permission rule syntax](/en/permissions#permission-rule-syntax).

172 207

173### Sandbox settings208### Sandbox settings

174 209

177**Filesystem and network restrictions** are configured via Read, Edit, and WebFetch permission rules, not via these sandbox settings.212**Filesystem and network restrictions** are configured via Read, Edit, and WebFetch permission rules, not via these sandbox settings.

178 213

180| :-------------------------- | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | :------------------------ |215| :---------------------------- | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | :------------------------------ |

184| `allowUnsandboxedCommands` | Allow commands to run outside the sandbox via the `dangerouslyDisableSandbox` parameter. When set to `false`, the `dangerouslyDisableSandbox` escape hatch is completely disabled and all commands must run sandboxed (or be in `excludedCommands`). Useful for enterprise policies that require strict sandboxing. Default: true | `false` |219| `allowUnsandboxedCommands` | Allow commands to run outside the sandbox via the `dangerouslyDisableSandbox` parameter. When set to `false`, the `dangerouslyDisableSandbox` escape hatch is completely disabled and all commands must run sandboxed (or be in `excludedCommands`). Useful for enterprise policies that require strict sandboxing. Default: true | `false` |

221| `network.allowAllUnixSockets` | Allow all Unix socket connections in sandbox. Default: false | `true` |

223| `network.allowedDomains` | Array of domains to allow for outbound network traffic. Supports wildcards (e.g., `*.example.com`). | `["github.com", "*.npmjs.org"]` |

187| `network.httpProxyPort` | HTTP proxy port used if you wish to bring your own proxy. If not specified, Claude will run its own proxy. | `8080` |224| `network.httpProxyPort` | HTTP proxy port used if you wish to bring your own proxy. If not specified, Claude will run its own proxy. | `8080` |

188| `network.socksProxyPort` | SOCKS5 proxy port used if you wish to bring your own proxy. If not specified, Claude will run its own proxy. | `8081` |225| `network.socksProxyPort` | SOCKS5 proxy port used if you wish to bring your own proxy. If not specified, Claude will run its own proxy. | `8081` |

190 227

191**Configuration example:**228**Configuration example:**

192 229

197 "autoAllowBashIfSandboxed": true,234 "autoAllowBashIfSandboxed": true,

198 "excludedCommands": ["docker"],235 "excludedCommands": ["docker"],

199 "network": {236 "network": {

237 "allowedDomains": ["github.com", "*.npmjs.org", "registry.yarnpkg.com"],

200 "allowUnixSockets": [238 "allowUnixSockets": [

201 "/var/run/docker.sock"239 "/var/run/docker.sock"

202 ],240 ],

297 335

298### Hook configuration336### Hook configuration

299 337

300**Enterprise-only setting**: Controls which hooks are allowed to run. This setting can only be configured in [managed settings](#settings-files) and provides enterprise administrators with strict control over hook execution.338**Managed settings only**: Controls which hooks are allowed to run. This setting can only be configured in [managed settings](#settings-files) and provides administrators with strict control over hook execution.

301 339

302**Behavior when `allowManagedHooksOnly` is `true`:**340**Behavior when `allowManagedHooksOnly` is `true`:**

303 341

316 354

317Settings apply in order of precedence. From highest to lowest:355Settings apply in order of precedence. From highest to lowest:

318 356

3191. **Managed settings** (Enterprise)3571. **Managed settings** ([`managed-settings.json`](/en/permissions#managed-settings) or [server-managed settings](/en/server-managed-settings))

320 * Remote settings configured via the [Claude.ai admin console](https://claude.ai/admin-settings/claude-code)358 * Policies deployed by IT/DevOps to system directories, or delivered from Anthropic's servers for Claude for Enterprise customers

321 * Fetched automatically when users authenticate

322 * Cannot be overridden

~~323~~

3242. **File-based managed settings** (`managed-settings.json`)

325 * Policies deployed by IT/DevOps to system directories

326 * Cannot be overridden by user or project settings359 * Cannot be overridden by user or project settings

327 * Ignored when remote managed settings are configured

328 360

3293. **Command line arguments**3612. **Command line arguments**

330 * Temporary overrides for a specific session362 * Temporary overrides for a specific session

331 363

3324. **Local project settings** (`.claude/settings.local.json`)3643. **Local project settings** (`.claude/settings.local.json`)

333 * Personal project-specific settings365 * Personal project-specific settings

334 366

3355. **Shared project settings** (`.claude/settings.json`)3674. **Shared project settings** (`.claude/settings.json`)

336 * Team-shared project settings in source control368 * Team-shared project settings in source control

337 369

3386. **User settings** (`~/.claude/settings.json`)3705. **User settings** (`~/.claude/settings.json`)

339 * Personal global settings371 * Personal global settings

340 372

341This hierarchy ensures that enterprise security policies are always enforced while still allowing teams and individuals to customize their experience.373This hierarchy ensures that organizational policies are always enforced while still allowing teams and individuals to customize their experience.

342 374

343For example, if your user settings allow `Bash(npm run:*)` but a project's shared settings deny it, the project setting takes precedence and the command is blocked.375For example, if your user settings allow `Bash(npm run *)` but a project's shared settings deny it, the project setting takes precedence and the command is blocked.

344 376

345### Key points about the configuration system377### Key points about the configuration system

346 378

347* **Memory files (`CLAUDE.md`)**: Contain instructions and context that Claude loads at startup379* **Memory files (`CLAUDE.md`)**: Contain instructions and context that Claude loads at startup

348* **Settings files (JSON)**: Configure permissions, environment variables, and tool behavior380* **Settings files (JSON)**: Configure permissions, environment variables, and tool behavior

349* **Slash commands**: Custom commands that can be invoked during a session with `/command-name`381* **Skills**: Custom prompts that can be invoked with `/skill-name` or loaded by Claude automatically

350* **MCP servers**: Extend Claude Code with additional tools and integrations382* **MCP servers**: Extend Claude Code with additional tools and integrations

351* **Precedence**: Higher-level configurations (Enterprise) override lower-level ones (User/Project)383* **Precedence**: Higher-level configurations (Managed) override lower-level ones (User/Project)

352* **Inheritance**: Settings are merged, with more specific settings adding to or overriding broader ones384* **Inheritance**: Settings are merged, with more specific settings adding to or overriding broader ones

353 385

354### System prompt386### System prompt

373}405}

374```406```

375 407

376This replaces the deprecated `ignorePatterns` configuration. Files matching these patterns will be completely invisible to Claude Code, preventing any accidental exposure of sensitive data.408This replaces the deprecated `ignorePatterns` configuration. Files matching these patterns are excluded from file discovery and search results, and read operations on these files are denied.

377 409

378## Subagent configuration410## Subagent configuration

379 411

386 418

387## Plugin configuration419## Plugin configuration

388 420

389Claude Code supports a plugin system that lets you extend functionality with custom commands, agents, hooks, and MCP servers. Plugins are distributed through marketplaces and can be configured at both user and repository levels.421Claude Code supports a plugin system that lets you extend functionality with skills, agents, hooks, and MCP servers. Plugins are distributed through marketplaces and can be configured at both user and repository levels.

390 422

391### Plugin settings423### Plugin settings

392 424

467* `github`: GitHub repository (uses `repo`)499* `github`: GitHub repository (uses `repo`)

468* `git`: Any git URL (uses `url`)500* `git`: Any git URL (uses `url`)

469* `directory`: Local filesystem path (uses `path`, for development only)501* `directory`: Local filesystem path (uses `path`, for development only)

502* `hostPattern`: regex pattern to match marketplace hosts (uses `hostPattern`)

470 503

471#### `strictKnownMarketplaces`504#### `strictKnownMarketplaces`

472 505

473**Enterprise-only setting**: Controls which plugin marketplaces users are allowed to add. This setting can only be configured in [`managed-settings.json`](/en/iam#enterprise-managed-settings) and provides enterprise administrators with strict control over marketplace sources.506**Managed settings only**: Controls which plugin marketplaces users are allowed to add. This setting can only be configured in [`managed-settings.json`](/en/permissions#managed-settings) and provides administrators with strict control over marketplace sources.

474 507

475**Managed settings file locations**:508**Managed settings file locations**:

476 509

480 513

481**Key characteristics**:514**Key characteristics**:

482 515

483* Only available in enterprise managed settings (`managed-settings.json`)516* Only available in managed settings (`managed-settings.json`)

484* Cannot be overridden by user or project settings (highest precedence)517* Cannot be overridden by user or project settings (highest precedence)

485* Enforced BEFORE network/filesystem operations (blocked sources never execute)518* Enforced BEFORE network/filesystem operations (blocked sources never execute)

486* Uses exact matching for source specifications (including `ref`, `path` for git sources)519* Uses exact matching for source specifications (including `ref`, `path` for git sources), except `hostPattern`, which uses regex matching

487 520

488**Allowlist behavior**:521**Allowlist behavior**:

489 522

493 526

494**All supported source types**:527**All supported source types**:

495 528

496The allowlist supports six marketplace source types. Each source must match exactly for a user's marketplace addition to be allowed.529The allowlist supports seven marketplace source types. Most sources use exact matching, while `hostPattern` uses regex matching against the marketplace host.

497 530

4981. **GitHub repositories**:5311. **GitHub repositories**:

499 532

524 557

525Fields: `url` (required), `headers` (optional: HTTP headers for authenticated access)558Fields: `url` (required), `headers` (optional: HTTP headers for authenticated access)

526 559

560<Note>

561 URL-based marketplaces only download the `marketplace.json` file. They do not download plugin files from the server. Plugins in URL-based marketplaces must use external sources (GitHub, npm, or git URLs) rather than relative paths. For plugins with relative paths, use a Git-based marketplace instead. See [Troubleshooting](/en/plugin-marketplaces#plugins-with-relative-paths-fail-in-url-based-marketplaces) for details.

562</Note>

563

5274. **NPM packages**:5644. **NPM packages**:

528 565

529```json theme={null}566```json theme={null}

551 588

552Fields: `path` (required: absolute path to directory containing `.claude-plugin/marketplace.json`)589Fields: `path` (required: absolute path to directory containing `.claude-plugin/marketplace.json`)

553 590

5917. **Host pattern matching**:

592

593```json theme={null}

594{ "source": "hostPattern", "hostPattern": "^github\\.example\\.com$" }

595{ "source": "hostPattern", "hostPattern": "^gitlab\\.internal\\.example\\.com$" }

596```

597

598Fields: `hostPattern` (required: regex pattern to match against the marketplace host)

599

600Use host pattern matching when you want to allow all marketplaces from a specific host without enumerating each repository individually. This is useful for organizations with internal GitHub Enterprise or GitLab servers where developers create their own marketplaces.

601

602Host extraction by source type:

603

604* `github`: always matches against `github.com`

605* `git`: extracts hostname from the URL (supports both HTTPS and SSH formats)

606* `url`: extracts hostname from the URL

607* `npm`, `file`, `directory`: not supported for host pattern matching

608

554**Configuration examples**:609**Configuration examples**:

555 610

556Example - Allow specific marketplaces only:611Example: allow specific marketplaces only:

557 612

558```json theme={null}613```json theme={null}

559{614{

587}642}

588```643```

589 644

645Example: allow all marketplaces from an internal git server:

646

647```json theme={null}

648{

649 "strictKnownMarketplaces": [

650 {

651 "source": "hostPattern",

652 "hostPattern": "^github\\.example\\.com$"

653 }

654 ]

655}

656```

657

590**Exact matching requirements**:658**Exact matching requirements**:

591 659

592Marketplace sources must match **exactly** for a user's addition to be allowed. For git-based sources (`github` and `git`), this includes all optional fields:660Marketplace sources must match **exactly** for a user's addition to be allowed. For git-based sources (`github` and `git`), this includes all optional fields:

611 679

613| --------------------- | ------------------------------------ | ------------------------------------ |681| --------------------- | ------------------------------------ | ------------------------------------ |

646**Important notes**:714**Important notes**:

647 715

648* Restrictions are checked BEFORE any network requests or filesystem operations716* Restrictions are checked BEFORE any network requests or filesystem operations

649* When blocked, users see clear error messages indicating the source is blocked by enterprise policy717* When blocked, users see clear error messages indicating the source is blocked by managed policy

650* The restriction applies only to adding NEW marketplaces; previously installed marketplaces remain accessible718* The restriction applies only to adding NEW marketplaces; previously installed marketplaces remain accessible

651* Enterprise managed settings have the highest precedence and cannot be overridden719* Managed settings have the highest precedence and cannot be overridden

652 720

653See [Enterprise marketplace restrictions](/en/plugin-marketplaces#enterprise-marketplace-restrictions) for user-facing documentation.721See [Managed marketplace restrictions](/en/plugin-marketplaces#managed-marketplace-restrictions) for user-facing documentation.

654 722

655### Managing plugins723### Managing plugins

656 724

672 All environment variables can also be configured in [`settings.json`](#available-settings). This is useful as a way to automatically set environment variables for each session, or to roll out a set of environment variables for your whole team or organization.740 All environment variables can also be configured in [`settings.json`](#available-settings). This is useful as a way to automatically set environment variables for each session, or to roll out a set of environment variables for your whole team or organization.

673</Note>741</Note>

674 742

676| :-------------------------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |744| :--------------------------------------------- | :---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --- |

677| `ANTHROPIC_API_KEY` | API key sent as `X-Api-Key` header, typically for the Claude SDK (for interactive usage, run `/login`) |745| `ANTHROPIC_API_KEY` | API key sent as `X-Api-Key` header, typically for the Claude SDK (for interactive usage, run `/login`) | |

678| `ANTHROPIC_AUTH_TOKEN` | Custom value for the `Authorization` header (the value you set here will be prefixed with `Bearer `) |746| `ANTHROPIC_AUTH_TOKEN` | Custom value for the `Authorization` header (the value you set here will be prefixed with `Bearer `) | |

683| `ANTHROPIC_FOUNDRY_API_KEY` | API key for Microsoft Foundry authentication (see [Microsoft Foundry](/en/microsoft-foundry)) |751| `ANTHROPIC_FOUNDRY_API_KEY` | API key for Microsoft Foundry authentication (see [Microsoft Foundry](/en/microsoft-foundry)) | |

684| `ANTHROPIC_MODEL` | Name of the model setting to use (see [Model Configuration](/en/model-config#environment-variables)) |752| `ANTHROPIC_FOUNDRY_BASE_URL` | Full base URL for the Foundry resource (for example, `https://my-resource.services.ai.azure.com/anthropic`). Alternative to `ANTHROPIC_FOUNDRY_RESOURCE` (see [Microsoft Foundry](/en/microsoft-foundry)) | |

685| `ANTHROPIC_SMALL_FAST_MODEL` | \[DEPRECATED] Name of [Haiku-class model for background tasks](/en/costs) |753| `ANTHROPIC_FOUNDRY_RESOURCE` | Foundry resource name (for example, `my-resource`). Required if `ANTHROPIC_FOUNDRY_BASE_URL` is not set (see [Microsoft Foundry](/en/microsoft-foundry)) | |

687| `AWS_BEARER_TOKEN_BEDROCK` | Bedrock API key for authentication (see [Bedrock API keys](https://aws.amazon.com/blogs/machine-learning/accelerate-ai-development-with-amazon-bedrock-api-keys/)) |755| `ANTHROPIC_SMALL_FAST_MODEL` | \[DEPRECATED] Name of [Haiku-class model for background tasks](/en/costs) | |

689| `BASH_MAX_OUTPUT_LENGTH` | Maximum number of characters in bash outputs before they are middle-truncated |757| `AWS_BEARER_TOKEN_BEDROCK` | Bedrock API key for authentication (see [Bedrock API keys](https://aws.amazon.com/blogs/machine-learning/accelerate-ai-development-with-amazon-bedrock-api-keys/)) | |

693| `CLAUDE_CODE_CLIENT_CERT` | Path to client certificate file for mTLS authentication |761| `CLAUDE_AUTOCOMPACT_PCT_OVERRIDE` | Set the percentage of context capacity (1-100) at which auto-compaction triggers. By default, auto-compaction triggers at approximately 95% capacity. Use lower values like `50` to compact earlier. Values above the default threshold have no effect. Applies to both main conversations and subagents. This percentage aligns with the `context_window.used_percentage` field available in [status line](/en/statusline) | |

695| `CLAUDE_CODE_CLIENT_KEY` | Path to client private key file for mTLS authentication |763| `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD` | Set to `1` to load CLAUDE.md files from directories specified with `--add-dir`. By default, additional directories do not load memory files | `1` |

696| `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS` | Set to `1` to disable Anthropic API-specific `anthropic-beta` headers. Use this if experiencing issues like "Unexpected value(s) for the `anthropic-beta` header" when using an LLM gateway with third-party providers |764| `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS` | Set to `1` to enable [agent teams](/en/agent-teams). Agent teams are experimental and disabled by default | |

697| `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` | Equivalent of setting `DISABLE_AUTOUPDATER`, `DISABLE_BUG_COMMAND`, `DISABLE_ERROR_REPORTING`, and `DISABLE_TELEMETRY` |765| `CLAUDE_CODE_API_KEY_HELPER_TTL_MS` | Interval in milliseconds at which credentials should be refreshed (when using `apiKeyHelper`) | |

701| `CLAUDE_CODE_OTEL_HEADERS_HELPER_DEBOUNCE_MS` | Interval for refreshing dynamic OpenTelemetry headers in milliseconds (default: 1740000 / 29 minutes). See [Dynamic headers](/en/monitoring-usage#dynamic-headers) |769| `CLAUDE_CODE_EFFORT_LEVEL` | Set the effort level for supported models. Values: `low`, `medium`, `high` (default). Lower effort is faster and cheaper, higher effort provides deeper reasoning. Currently supported with Opus 4.6 only. See [Adjust effort level](/en/model-config#adjust-effort-level) | |

702| `CLAUDE_CODE_SHELL_PREFIX` | Command prefix to wrap all bash commands (for example, for logging or auditing). Example: `/path/to/logger.sh` will execute `/path/to/logger.sh <command>` |770| `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS` | Set to `1` to disable Anthropic API-specific `anthropic-beta` headers. Use this if experiencing issues like "Unexpected value(s) for the `anthropic-beta` header" when using an LLM gateway with third-party providers | |

703| `CLAUDE_CODE_SKIP_BEDROCK_AUTH` | Skip AWS authentication for Bedrock (for example, when using an LLM gateway) |771| `CLAUDE_CODE_DISABLE_AUTO_MEMORY` | Set to `1` to disable [auto memory](/en/memory#auto-memory). Set to `0` to force auto memory on during the gradual rollout. When disabled, Claude does not create or load auto memory files | |

704| `CLAUDE_CODE_SKIP_FOUNDRY_AUTH` | Skip Azure authentication for Microsoft Foundry (for example, when using an LLM gateway) |772| `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` | Set to `1` to disable all background task functionality, including the `run_in_background` parameter on Bash and subagent tools, auto-backgrounding, and the Ctrl+B shortcut | |

705| `CLAUDE_CODE_SKIP_VERTEX_AUTH` | Skip Google authentication for Vertex (for example, when using an LLM gateway) |773| `CLAUDE_CODE_DISABLE_FEEDBACK_SURVEY` | Set to `1` to disable the "How is Claude doing?" session quality surveys. Also disabled when using third-party providers or when telemetry is disabled. See [Session quality surveys](/en/data-usage#session-quality-surveys) | |

706| `CLAUDE_CODE_SUBAGENT_MODEL` | See [Model configuration](/en/model-config) |774| `CLAUDE_CODE_EXIT_AFTER_STOP_DELAY` | Time in milliseconds to wait after the query loop becomes idle before automatically exiting. Useful for automated workflows and scripts using SDK mode | |

707| `CLAUDE_CODE_USE_BEDROCK` | Use [Bedrock](/en/amazon-bedrock) |775| `CLAUDE_CODE_PROXY_RESOLVES_HOSTS` | Set to `true` to allow the proxy to perform DNS resolution instead of the caller. Opt-in for environments where the proxy should handle hostname resolution | |

708| `CLAUDE_CODE_USE_FOUNDRY` | Use [Microsoft Foundry](/en/microsoft-foundry) |776| `CLAUDE_CODE_TASK_LIST_ID` | Share a task list across sessions. Set the same ID in multiple Claude Code instances to coordinate on a shared task list. See [Task list](/en/interactive-mode#task-list) | |

710| `CLAUDE_CONFIG_DIR` | Customize where Claude Code stores its configuration and data files |778| `CLAUDE_CODE_TMPDIR` | Override the temp directory used for internal temp files. Claude Code appends `/claude/` to this path. Default: `/tmp` on Unix/macOS, `os.tmpdir()` on Windows | |

713| `DISABLE_COST_WARNINGS` | Set to `1` to disable cost warning messages |781| `CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION` | Set to `false` to disable prompt suggestions (the "Prompt suggestions" toggle in `/config`). These are the grayed-out predictions that appear in your prompt input after Claude responds. See [Prompt suggestions](/en/interactive-mode#prompt-suggestions) | |

714| `DISABLE_ERROR_REPORTING` | Set to `1` to opt out of Sentry error reporting |782| `CLAUDE_CODE_ENABLE_TASKS` | Set to `false` to temporarily revert to the previous TODO list instead of the task tracking system. Default: `true`. See [Task list](/en/interactive-mode#task-list) | |

715| `DISABLE_NON_ESSENTIAL_MODEL_CALLS` | Set to `1` to disable model calls for non-critical paths like flavor text |783| `CLAUDE_CODE_ENABLE_TELEMETRY` | Set to `1` to enable OpenTelemetry data collection for metrics and logging. Required before configuring OTel exporters. See [Monitoring](/en/monitoring-usage) | |

716| `DISABLE_PROMPT_CACHING` | Set to `1` to disable prompt caching for all models (takes precedence over per-model settings) |784| `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` | Override the default token limit for file reads. Useful when you need to read larger files in full | |

717| `DISABLE_PROMPT_CACHING_HAIKU` | Set to `1` to disable prompt caching for Haiku models |785| `CLAUDE_CODE_HIDE_ACCOUNT_INFO` | Set to `1` to hide your email address and organization name from the Claude Code UI. Useful when streaming or recording | |

719| `DISABLE_PROMPT_CACHING_SONNET` | Set to `1` to disable prompt caching for Sonnet models |787| `CLAUDE_CODE_MAX_OUTPUT_TOKENS` | Set the maximum number of output tokens for most requests. Default: 32,000. Maximum: 64,000. Increasing this value reduces the effective context window available before [auto-compaction](/en/costs#reduce-token-usage) triggers. | |

720| `DISABLE_TELEMETRY` | Set to `1` to opt out of Statsig telemetry (note that Statsig events do not include user data like code, file paths, or bash commands) |788| `CLAUDE_CODE_OTEL_HEADERS_HELPER_DEBOUNCE_MS` | Interval for refreshing dynamic OpenTelemetry headers in milliseconds (default: 1740000 / 29 minutes). See [Dynamic headers](/en/monitoring-usage#dynamic-headers) | |

721| `HTTP_PROXY` | Specify HTTP proxy server for network connections |789| `CLAUDE_CODE_PLAN_MODE_REQUIRED` | Auto-set to `true` on [agent team](/en/agent-teams) teammates that require plan approval. Read-only: set by Claude Code when spawning teammates. See [require plan approval](/en/agent-teams#require-plan-approval-for-teammates) | |

723| `MAX_MCP_OUTPUT_TOKENS` | Maximum number of tokens allowed in MCP tool responses. Claude Code displays a warning when output exceeds 10,000 tokens (default: 25000) |791| `CLAUDE_CODE_SHELL_PREFIX` | Command prefix to wrap all bash commands (for example, for logging or auditing). Example: `/path/to/logger.sh` will execute `/path/to/logger.sh <command>` | |

724| `MAX_THINKING_TOKENS` | Enable [extended thinking](https://docs.claude.com/en/docs/build-with-claude/extended-thinking) and set the token budget for the thinking process. Extended thinking improves performance on complex reasoning and coding tasks but impacts [prompt caching efficiency](https://docs.claude.com/en/docs/build-with-claude/prompt-caching#caching-with-thinking-blocks). Disabled by default. |792| `CLAUDE_CODE_SKIP_BEDROCK_AUTH` | Skip AWS authentication for Bedrock (for example, when using an LLM gateway) | |

803| `DISABLE_ERROR_REPORTING` | Set to `1` to opt out of Sentry error reporting | |

804| `DISABLE_INSTALLATION_CHECKS` | Set to `1` to disable installation warnings. Use only when manually managing the installation location, as this can mask issues with standard installations | |

805| `DISABLE_NON_ESSENTIAL_MODEL_CALLS` | Set to `1` to disable model calls for non-critical paths like flavor text | |

806| `DISABLE_PROMPT_CACHING` | Set to `1` to disable prompt caching for all models (takes precedence over per-model settings) | |

807| `DISABLE_PROMPT_CACHING_HAIKU` | Set to `1` to disable prompt caching for Haiku models | |

808| `DISABLE_PROMPT_CACHING_OPUS` | Set to `1` to disable prompt caching for Opus models | |

809| `DISABLE_PROMPT_CACHING_SONNET` | Set to `1` to disable prompt caching for Sonnet models | |

810| `DISABLE_TELEMETRY` | Set to `1` to opt out of Statsig telemetry (note that Statsig events do not include user data like code, file paths, or bash commands) | |

811| `ENABLE_TOOL_SEARCH` | Controls [MCP tool search](/en/mcp#scale-with-mcp-tool-search). Values: `auto` (default, enables at 10% context), `auto:N` (custom threshold, e.g., `auto:5` for 5%), `true` (always on), `false` (disabled) | |

812| `FORCE_AUTOUPDATE_PLUGINS` | Set to `true` to force plugin auto-updates even when the main auto-updater is disabled via `DISABLE_AUTOUPDATER` | |

813| `HTTP_PROXY` | Specify HTTP proxy server for network connections | |

814| `HTTPS_PROXY` | Specify HTTPS proxy server for network connections | |

815| `IS_DEMO` | Set to `true` to enable demo mode: hides email and organization from the UI, skips onboarding, and hides internal commands. Useful for streaming or recording sessions | |

816| `MAX_MCP_OUTPUT_TOKENS` | Maximum number of tokens allowed in MCP tool responses. Claude Code displays a warning when output exceeds 10,000 tokens (default: 25000) | |

817| `MAX_THINKING_TOKENS` | Override the [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) token budget. Thinking is enabled at max budget (31,999 tokens) by default. Use this to limit the budget (for example, `MAX_THINKING_TOKENS=10000`) or disable thinking entirely (`MAX_THINKING_TOKENS=0`). For Opus 4.6, thinking depth is controlled by [effort level](/en/model-config#adjust-effort-level) instead, and this variable is ignored unless set to `0` to disable thinking. | |

818| `MCP_CLIENT_SECRET` | OAuth client secret for MCP servers that require [pre-configured credentials](/en/mcp#use-pre-configured-oauth-credentials). Avoids the interactive prompt when adding a server with `--client-secret` | |

819| `MCP_OAUTH_CALLBACK_PORT` | Fixed port for the OAuth redirect callback, as an alternative to `--callback-port` when adding an MCP server with [pre-configured credentials](/en/mcp#use-pre-configured-oauth-credentials) | |

820| `MCP_TIMEOUT` | Timeout in milliseconds for MCP server startup | |

821| `MCP_TOOL_TIMEOUT` | Timeout in milliseconds for MCP tool execution | |

822| `NO_PROXY` | List of domains and IPs to which requests will be directly issued, bypassing proxy | |

823| `SLASH_COMMAND_TOOL_CHAR_BUDGET` | Override the character budget for skill metadata shown to the [Skill tool](/en/skills#control-who-invokes-a-skill). The budget scales dynamically at 2% of the context window, with a fallback of 16,000 characters. Legacy name kept for backwards compatibility | |

824| `USE_BUILTIN_RIPGREP` | Set to `0` to use system-installed `rg` instead of `rg` included with Claude Code | |

825| `VERTEX_REGION_CLAUDE_3_5_HAIKU` | Override region for Claude 3.5 Haiku when using Vertex AI | |

826| `VERTEX_REGION_CLAUDE_3_7_SONNET` | Override region for Claude 3.7 Sonnet when using Vertex AI | |

827| `VERTEX_REGION_CLAUDE_4_0_OPUS` | Override region for Claude 4.0 Opus when using Vertex AI | |

828| `VERTEX_REGION_CLAUDE_4_0_SONNET` | Override region for Claude 4.0 Sonnet when using Vertex AI | |

829| `VERTEX_REGION_CLAUDE_4_1_OPUS` | Override region for Claude 4.1 Opus when using Vertex AI | |

735 830

736## Tools available to Claude831## Tools available to Claude

737 832

738Claude Code has access to a set of powerful tools that help it understand and modify your codebase:833Claude Code has access to a set of powerful tools that help it understand and modify your codebase:

739 834

741| :------------------ | :------------------------------------------------------------------------------------------------ | :------------------ |836| :------------------ | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | :------------------ |

742| **AskUserQuestion** | Asks the user multiple choice questions to gather information or clarify ambiguity | No |837| **AskUserQuestion** | Asks multiple-choice questions to gather requirements or clarify ambiguity | No |

743| **Bash** | Executes shell commands in your environment (see [Bash tool behavior](#bash-tool-behavior) below) | Yes |838| **Bash** | Executes shell commands in your environment (see [Bash tool behavior](#bash-tool-behavior) below) | Yes |

744| **BashOutput** | Retrieves output from a background bash shell | No |839| **TaskOutput** | Retrieves output from a background task (bash shell or subagent) | No |

745| **Edit** | Makes targeted edits to specific files | Yes |840| **Edit** | Makes targeted edits to specific files | Yes |

746| **ExitPlanMode** | Prompts the user to exit plan mode and start coding | Yes |841| **ExitPlanMode** | Prompts the user to exit plan mode and start coding | Yes |

747| **Glob** | Finds files based on pattern matching | No |842| **Glob** | Finds files based on pattern matching | No |

748| **Grep** | Searches for patterns in file contents | No |843| **Grep** | Searches for patterns in file contents | No |

749| **KillShell** | Kills a running background bash shell by its ID | No |844| **KillShell** | Kills a running background bash shell by its ID | No |

845| **MCPSearch** | Searches for and loads MCP tools when [tool search](/en/mcp#scale-with-mcp-tool-search) is enabled | No |

750| **NotebookEdit** | Modifies Jupyter notebook cells | Yes |846| **NotebookEdit** | Modifies Jupyter notebook cells | Yes |

751| **Read** | Reads the contents of files | No |847| **Read** | Reads the contents of files | No |

752| **Skill** | Executes a skill within the main conversation | Yes |848| **Skill** | Executes a [skill](/en/skills#control-who-invokes-a-skill) within the main conversation | Yes |

753| **SlashCommand** | Runs a [custom slash command](/en/slash-commands#slashcommand-tool) | Yes |

754| **Task** | Runs a sub-agent to handle complex, multi-step tasks | No |849| **Task** | Runs a sub-agent to handle complex, multi-step tasks | No |

755| **TodoWrite** | Creates and manages structured task lists | No |850| **TaskCreate** | Creates a new task in the task list | No |

851| **TaskGet** | Retrieves full details for a specific task | No |

852| **TaskList** | Lists all tasks with their current status | No |

853| **TaskUpdate** | Updates task status, dependencies, details, or deletes tasks | No |

756| **WebFetch** | Fetches content from a specified URL | Yes |854| **WebFetch** | Fetches content from a specified URL | Yes |

757| **WebSearch** | Performs web searches with domain filtering | Yes |855| **WebSearch** | Performs web searches with domain filtering | Yes |

758| **Write** | Creates or overwrites files | Yes |856| **Write** | Creates or overwrites files | Yes |

857| **LSP** | Code intelligence via language servers. Reports type errors and warnings automatically after file edits. Also supports navigation operations: jump to definitions, find references, get type info, list symbols, find implementations, trace call hierarchies. Requires a [code intelligence plugin](/en/discover-plugins#code-intelligence) and its language server binary | No |

759 858

760Permission rules can be configured using `/allowed-tools` or in [permission settings](/en/settings#available-settings). Also see [Tool-specific permission rules](/en/iam#tool-specific-permission-rules).859Permission rules can be configured using `/allowed-tools` or in [permission settings](/en/settings#available-settings). Also see [Tool-specific permission rules](/en/permissions#tool-specific-permission-rules).

761 860

762### Bash tool behavior861### Bash tool behavior

763 862

819 918

820The hook writes to `$CLAUDE_ENV_FILE`, which is then sourced before each Bash command. This is ideal for team-shared project configurations.919The hook writes to `$CLAUDE_ENV_FILE`, which is then sourced before each Bash command. This is ideal for team-shared project configurations.

821 920

822See [SessionStart hooks](/en/hooks#persisting-environment-variables) for more details on Option 3.921See [SessionStart hooks](/en/hooks#persist-environment-variables) for more details on Option 3.

823 922

824### Extending tools with hooks923### Extending tools with hooks

825 924

832 931

833## See also932## See also

834 933

835* [Identity and Access Management](/en/iam#configuring-permissions) - Learn about Claude Code's permission system934* [Permissions](/en/permissions): permission system, rule syntax, tool-specific patterns, and managed policies

836* [IAM and access control](/en/iam#enterprise-managed-settings) - Enterprise policy management935* [Authentication](/en/authentication): set up user access to Claude Code

837* [Troubleshooting](/en/troubleshooting#auto-updater-issues) - Solutions for common configuration issues936* [Troubleshooting](/en/troubleshooting): solutions for common configuration issues

~~838~~

~~839~~

~~840~~

841> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

setup.md +153 −92

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Set up Claude Code5# Set up Claude Code

2 6

3> Install, authenticate, and start using Claude Code on your development machine.7> Install, authenticate, and start using Claude Code on your development machine.

4 8

5## System requirements9## System requirements

6 10

~~7* **Operating Systems**: macOS 10.15+, Ubuntu 20.04+/Debian 10+, or Windows 10+ (with WSL 1, WSL 2, or Git for Windows)~~11* **Operating System**:

12 * macOS 13.0+

13 * Windows 10 1809+ or Windows Server 2019+ ([see setup notes](#platform-specific-setup))

14 * Ubuntu 20.04+

15 * Debian 10+

16 * Alpine Linux 3.19+ ([additional dependencies required](#platform-specific-setup))

8* **Hardware**: 4 GB+ RAM17* **Hardware**: 4 GB+ RAM

~~9* **Software**: [Node.js 18+](https://nodejs.org/en/download) (only required for npm installation)~~18* **Network**: Internet connection required (see [network configuration](/en/network-config#network-access-requirements))

~~10* **Network**: Internet connection required for authentication and AI processing~~19* **Shell**: Works best in Bash or Zsh

~~11* **Shell**: Works best in Bash, Zsh or Fish~~

12* **Location**: [Anthropic supported countries](https://www.anthropic.com/supported-countries)20* **Location**: [Anthropic supported countries](https://www.anthropic.com/supported-countries)

13 21

14### Additional dependencies22### Additional dependencies

15 23

16* **ripgrep**: Usually included with Claude Code. If search fails, see [search troubleshooting](/en/troubleshooting#search-and-discovery-issues).24* **ripgrep**: Usually included with Claude Code. If search fails, see [search troubleshooting](/en/troubleshooting#search-and-discovery-issues).

25* **[Node.js 18+](https://nodejs.org/en/download)**: Only required for [deprecated npm installation](#npm-installation-deprecated)

17 26

~~18## Standard installation~~27## Installation

19 28

20To install Claude Code, use one of the following methods:29To install Claude Code, use one of the following methods:

21 30

38 ```batch theme={null}47 ```batch theme={null}

39 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd48 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

40 ```49 ```

51 <Info>

52 Native installations automatically update in the background to keep you on the latest version.

53 </Info>

41 </Tab>54 </Tab>

42 55

43 <Tab title="Homebrew">56 <Tab title="Homebrew">

44 ```sh theme={null}57 ```sh theme={null}

45 brew install --cask claude-code58 brew install --cask claude-code

46 ```59 ```

~~47 </Tab>~~

48 60

~~49 <Tab title="NPM">~~61 <Info>

~~50 If you have [Node.js 18 or newer installed](https://nodejs.org/en/download/):~~62 Homebrew installations do not auto-update. Run `brew upgrade claude-code` periodically to get the latest features and security fixes.

63 </Info>

64 </Tab>

51 65

~~52 ```sh theme={null}~~66 <Tab title="WinGet">

~~53 npm install -g @anthropic-ai/claude-code~~67 ```powershell theme={null}

68 winget install Anthropic.ClaudeCode

54 ```69 ```

71 <Info>

72 WinGet installations do not auto-update. Run `winget upgrade Anthropic.ClaudeCode` periodically to get the latest features and security fixes.

73 </Info>

55 </Tab>74 </Tab>

56</Tabs>75</Tabs>

57 76

~~58<Note>~~

~~59 Some users may be automatically migrated to an improved installation method.~~

~~60</Note>~~

62After the installation process completes, navigate to your project and start Claude Code:77After the installation process completes, navigate to your project and start Claude Code:

63 78

64```bash theme={null}79```bash theme={null}

66claude81claude

67```82```

68 83

~~69Claude Code offers the following authentication options:~~84If you encounter any issues during installation, consult the [troubleshooting guide](/en/troubleshooting).

70 85

711. **Claude Console**: The default option. Connect through the Claude Console and complete the OAuth process. Requires active billing in the [Anthropic console](https://console.anthropic.com). A "Claude Code" workspace is automatically created for usage tracking and cost management. You can't create API keys for the Claude Code workspace; it's dedicated exclusively for Claude Code usage.86<Tip>

722. **Claude App (with Pro or Max plan)**: Subscribe to Claude's [Pro or Max plan](https://claude.com/pricing) for a unified subscription that includes both Claude Code and the web interface. Get more value at the same price point while managing your account in one place. Log in with your Claude.ai account. During launch, choose the option that matches your subscription type.87 Run `claude doctor` after installation to check your installation type and version.

733. **Enterprise platforms**: Configure Claude Code to use [Amazon Bedrock, Google Vertex AI, or Microsoft Foundry](/en/third-party-integrations) for enterprise deployments with your existing cloud infrastructure.88</Tip>

~~75<Note>~~

~~76 Claude Code securely stores your credentials. See [Credential Management](/en/iam#credential-management) for details.~~

~~77</Note>~~

~~79## Windows setup~~

~~81**Option 1: Claude Code within WSL**~~

~~83* Both WSL 1 and WSL 2 are supported~~

84 89

~~85**Option 2: Claude Code on native Windows with Git Bash**~~90### Platform-specific setup

86 91

~~87* Requires [Git for Windows](https://git-scm.com/downloads/win)~~92**Windows**: Run Claude Code natively (requires [Git Bash](https://git-scm.com/downloads/win)) or inside WSL. Both WSL 1 and WSL 2 are supported, but WSL 1 has limited support and does not support features like Bash tool sandboxing.

~~88* For portable Git installations, specify the path to your `bash.exe`:~~

~~89 ```powershell theme={null}~~

~~90 $env:CLAUDE_CODE_GIT_BASH_PATH="C:\Program Files\Git\bin\bash.exe"~~

~~91 ```~~

92 93

~~93## Alternative installation methods~~94**Alpine Linux and other musl/uClibc-based distributions**:

94 95

~~95Claude Code offers multiple installation methods to suit different environments.~~96The native installer on Alpine and other musl/uClibc-based distributions requires `libgcc`, `libstdc++`, and `ripgrep`. Install these using your distribution's package manager, then set `USE_BUILTIN_RIPGREP=0`.

96 97

~~97If you encounter any issues during installation, consult the [troubleshooting guide](/en/troubleshooting#linux-permission-issues).~~98On Alpine:

98 99

~~99<Tip>~~100```bash theme={null}

100 Run `claude doctor` after installation to check your installation type and version.101apk add libgcc libstdc++ ripgrep

101</Tip>102```

102 103

103### Native installation options104### Authentication

104 105

105The native installation is the recommended method and offers several benefits:106#### For individuals

106 107

107* One self-contained executable1081. **Claude Pro or Max plan** (recommended): Subscribe to Claude's [Pro or Max plan](https://claude.ai/pricing) for a unified subscription that includes both Claude Code and Claude on the web. Manage your account in one place and log in with your Claude.ai account.

108* No Node.js dependency1092. **Claude Console**: Connect through the [Claude Console](https://console.anthropic.com) and complete the OAuth process. Requires active billing in the Anthropic Console. A "Claude Code" workspace is automatically created for usage tracking and cost management. You can't create API keys for the Claude Code workspace; it's dedicated exclusively for Claude Code usage.

109* Improved auto-updater stability

110 110

111If you have an existing installation of Claude Code, use `claude install` to migrate to the native binary installation.111#### For teams and organizations

112 112

113For advanced installation options with the native installer:1131. **Claude for Teams or Enterprise** (recommended): Subscribe to [Claude for Teams](https://claude.com/pricing#team-&-enterprise) or [Claude for Enterprise](https://anthropic.com/contact-sales) for centralized billing, team management, and access to both Claude Code and Claude on the web. Team members log in with their Claude.ai accounts.

1142. **Claude Console with team billing**: Set up a shared [Claude Console](https://console.anthropic.com) organization with team billing. Invite team members and assign roles for usage tracking.

1153. **Cloud providers**: Configure Claude Code to use [Amazon Bedrock, Google Vertex AI, or Microsoft Foundry](/en/third-party-integrations) for deployments with your existing cloud infrastructure.

114 116

115**macOS, Linux, WSL:**117### Install a specific version

116 118

117```bash theme={null}119The native installer accepts either a specific version number or a release channel (`latest` or `stable`). The channel you choose at install time becomes your default for auto-updates. See [Configure release channel](#configure-release-channel) for more information.

118# Install stable version (default)

119curl -fsSL https://claude.ai/install.sh | bash

120 120

121# Install latest version121To install the latest version (default):

122curl -fsSL https://claude.ai/install.sh | bash -s latest

123 122

124# Install specific version number123<Tabs>

125curl -fsSL https://claude.ai/install.sh | bash -s 1.0.58124 <Tab title="macOS, Linux, WSL">

126```125 ```bash theme={null}

126 curl -fsSL https://claude.ai/install.sh | bash

127 ```

128 </Tab>

127 129

128<Note>130 <Tab title="Windows PowerShell">

129 **Alpine Linux and other musl/uClibc-based distributions**: The native build requires `libgcc`, `libstdc++`, and `ripgrep`. For Alpine: `apk add libgcc libstdc++ ripgrep`. Set `USE_BUILTIN_RIPGREP=0`.131 ```powershell theme={null}

130</Note>132 irm https://claude.ai/install.ps1 | iex

133 ```

134 </Tab>

131 135

132**Windows PowerShell:**136 <Tab title="Windows CMD">

137 ```batch theme={null}

138 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

139 ```

140 </Tab>

141</Tabs>

133 142

134```powershell theme={null}143To install the stable version:

135# Install stable version (default)

136irm https://claude.ai/install.ps1 | iex

137 144

138# Install latest version145<Tabs>

139& ([scriptblock]::Create((irm https://claude.ai/install.ps1))) latest146 <Tab title="macOS, Linux, WSL">

147 ```bash theme={null}

148 curl -fsSL https://claude.ai/install.sh | bash -s stable

149 ```

150 </Tab>

140 151

141# Install specific version number152 <Tab title="Windows PowerShell">

142& ([scriptblock]::Create((irm https://claude.ai/install.ps1))) 1.0.58153 ```powershell theme={null}

143```154 & ([scriptblock]::Create((irm https://claude.ai/install.ps1))) stable

155 ```

156 </Tab>

144 157

145**Windows CMD:**158 <Tab title="Windows CMD">

159 ```batch theme={null}

160 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd stable && del install.cmd

161 ```

162 </Tab>

163</Tabs>

146 164

147```batch theme={null}165To install a specific version number:

148REM Install stable version (default)

149curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

150 166

151REM Install latest version167<Tabs>

152curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd latest && del install.cmd168 <Tab title="macOS, Linux, WSL">

169 ```bash theme={null}

170 curl -fsSL https://claude.ai/install.sh | bash -s 1.0.58

171 ```

172 </Tab>

153 173

154REM Install specific version number174 <Tab title="Windows PowerShell">

155curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd 1.0.58 && del install.cmd175 ```powershell theme={null}

156```176 & ([scriptblock]::Create((irm https://claude.ai/install.ps1))) 1.0.58

177 ```

178 </Tab>

157 179

158<Tip>180 <Tab title="Windows CMD">

159 Make sure that you remove any outdated aliases or symlinks before installing.181 ```batch theme={null}

160</Tip>182 curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd 1.0.58 && del install.cmd

183 ```

184 </Tab>

185</Tabs>

161 186

162**Binary integrity and code signing**187### Binary integrity and code signing

163 188

164* SHA256 checksums for all platforms are published in the release manifests, currently located at `https://storage.googleapis.com/claude-code-dist-86c565f3-f756-42ad-8dfa-d59b1c096819/claude-code-releases/{VERSION}/manifest.json` (example: replace `{VERSION}` with `2.0.30`)189* SHA256 checksums for all platforms are published in the release manifests, currently located at `https://storage.googleapis.com/claude-code-dist-86c565f3-f756-42ad-8dfa-d59b1c096819/claude-code-releases/{VERSION}/manifest.json` (example: replace `{VERSION}` with `2.0.30`)

165* Signed binaries are distributed for the following platforms:190* Signed binaries are distributed for the following platforms:

166 * macOS: Signed by "Anthropic PBC" and notarized by Apple191 * macOS: Signed by "Anthropic PBC" and notarized by Apple

167 * Windows: Signed by "Anthropic, PBC"192 * Windows: Signed by "Anthropic, PBC"

168 193

169### NPM installation194## NPM installation (deprecated)

195

196NPM installation is deprecated. Use the [native installation](#installation) method when possible. To migrate an existing npm installation to native, run `claude install`.

170 197

171For environments where NPM is preferred or required:198**Global npm installation**

172 199

173```sh theme={null}200```sh theme={null}

174npm install -g @anthropic-ai/claude-code201npm install -g @anthropic-ai/claude-code

176 203

177<Warning>204<Warning>

178 Do NOT use `sudo npm install -g` as this can lead to permission issues and security risks.205 Do NOT use `sudo npm install -g` as this can lead to permission issues and security risks.

179 If you encounter permission errors, see [configure Claude Code](/en/troubleshooting#linux-permission-issues) for recommended solutions.206 If you encounter permission errors, see [troubleshooting permission errors](/en/troubleshooting#command-not-found-claude-or-permission-errors) for recommended solutions.

180</Warning>207</Warning>

181 208

182## Running on AWS or GCP209## Windows setup

210

211**Option 1: Claude Code within WSL**

183 212

184By default, Claude Code uses the Claude API.213* Both WSL 1 and WSL 2 are supported

214* WSL 2 supports [sandboxing](/en/sandboxing) for enhanced security. WSL 1 does not support sandboxing.

185 215

186For details on running Claude Code on AWS or GCP, see [third-party integrations](/en/third-party-integrations).216**Option 2: Claude Code on native Windows with Git Bash**

217

218* Requires [Git for Windows](https://git-scm.com/downloads/win)

219* For portable Git installations, specify the path to your `bash.exe`:

220 ```powershell theme={null}

221 $env:CLAUDE_CODE_GIT_BASH_PATH="C:\Program Files\Git\bin\bash.exe"

222 ```

187 223

188## Update Claude Code224## Update Claude Code

189 225

196* **Notifications**: You'll see a notification when updates are installed232* **Notifications**: You'll see a notification when updates are installed

197* **Applying updates**: Updates take effect the next time you start Claude Code233* **Applying updates**: Updates take effect the next time you start Claude Code

198 234

199**Disable auto-updates:**235<Note>

236 Homebrew and WinGet installations do not auto-update. Use `brew upgrade claude-code` or `winget upgrade Anthropic.ClaudeCode` to update manually.

237

238 **Known issue:** Claude Code may notify you of updates before the new version is available in these package managers. If an upgrade fails, wait and try again later.

239</Note>

240

241### Configure release channel

242

243Configure which release channel Claude Code follows for both auto-updates and `claude update` with the `autoUpdatesChannel` setting:

244

245* `"latest"` (default): Receive new features as soon as they're released

246* `"stable"`: Use a version that is typically about one week old, skipping releases with major regressions

247

248Configure this via `/config` → **Auto-update channel**, or add it to your [settings.json file](/en/settings):

249

250```json theme={null}

251{

252 "autoUpdatesChannel": "stable"

253}

254```

255

256For enterprise deployments, you can enforce a consistent release channel across your organization using [managed settings](/en/permissions#managed-settings).

257

258### Disable auto-updates

200 259

201Set the `DISABLE_AUTOUPDATER` environment variable in your shell or [settings.json file](/en/settings):260Set the `DISABLE_AUTOUPDATER` environment variable in your shell or [settings.json file](/en/settings):

202 261

245brew uninstall --cask claude-code304brew uninstall --cask claude-code

246```305```

247 306

307### WinGet installation

308

309```powershell theme={null}

310winget uninstall Anthropic.ClaudeCode

311```

312

248### NPM installation313### NPM installation

249 314

250```bash theme={null}315```bash theme={null}

294rmdir /s /q ".claude"359rmdir /s /q ".claude"

295del ".mcp.json"360del ".mcp.json"

296```361```

~~297~~

~~298~~

~~299~~

300> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

skills.md +478 −278

Details

~~1# Agent Skills~~1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

2 4

~~3> Create, manage, and share Skills to extend Claude's capabilities in Claude Code.~~5# Extend Claude with skills

4 6

5This guide shows you how to create, use, and manage Agent Skills in Claude Code. For background on how Skills work across Claude products, see [What are Skills?](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/overview).7> Create, manage, and share skills to extend Claude's capabilities in Claude Code. Includes custom slash commands.

6 8

7A Skill is a markdown file that teaches Claude how to do something specific: reviewing PRs using your team's standards, generating commit messages in your preferred format, or querying your company's database schema. When you ask Claude something that matches a Skill's purpose, Claude automatically applies it.9Skills extend what Claude can do. Create a `SKILL.md` file with instructions, and Claude adds it to its toolkit. Claude uses skills when relevant, or you can invoke one directly with `/skill-name`.

8 10

~~9## Create your first Skill~~11<Note>

12 For built-in commands like `/help` and `/compact`, see [interactive mode](/en/interactive-mode#built-in-commands).

10 13

11This example creates a personal Skill that teaches Claude to explain code using visual diagrams and analogies. Unlike Claude's default explanations, this Skill ensures every explanation includes an ASCII diagram and a real-world analogy.14 **Custom slash commands have been merged into skills.** A file at `.claude/commands/review.md` and a skill at `.claude/skills/review/SKILL.md` both create `/review` and work the same way. Your existing `.claude/commands/` files keep working. Skills add optional features: a directory for supporting files, frontmatter to [control whether you or Claude invokes them](#control-who-invokes-a-skill), and the ability for Claude to load them automatically when relevant.

15</Note>

12 16

~~13<Steps>~~17Claude Code skills follow the [Agent Skills](https://agentskills.io) open standard, which works across multiple AI tools. Claude Code extends the standard with additional features like [invocation control](#control-who-invokes-a-skill), [subagent execution](#run-skills-in-a-subagent), and [dynamic context injection](#inject-dynamic-context).

~~14 <Step title="Check available Skills">~~

~~15 Before creating a Skill, see what Skills Claude already has access to:~~

16 18

~~17 ```~~19## Getting started

~~18 What Skills are available?~~

~~19 ```~~

20 20

~~21 Claude will list any Skills currently loaded. You may see none, or you may see Skills from plugins or your organization.~~21### Create your first skill

~~22 </Step>~~

23 22

~~24 <Step title="Create the Skill directory">~~23This example creates a skill that teaches Claude to explain code using visual diagrams and analogies. Since it uses default frontmatter, Claude can load it automatically when you ask how something works, or you can invoke it directly with `/explain-code`.

25 Create a directory for the Skill in your personal Skills folder. Personal Skills are available across all your projects. (You can also create [project Skills](#where-skills-live) in `.claude/skills/` to share with your team.)24

25<Steps>

26 <Step title="Create the skill directory">

27 Create a directory for the skill in your personal skills folder. Personal skills are available across all your projects.

26 28

27 ```bash theme={null}29 ```bash theme={null}

~~28 mkdir -p ~/.claude/skills/explaining-code~~30 mkdir -p ~/.claude/skills/explain-code

29 ```31 ```

30 </Step>32 </Step>

31 33

32 <Step title="Write SKILL.md">34 <Step title="Write SKILL.md">

33 Every Skill needs a `SKILL.md` file. The file starts with YAML metadata between `---` markers and must include a `name` and `description`, followed by Markdown instructions that Claude follows when the Skill is active.35 Every skill needs a `SKILL.md` file with two parts: YAML frontmatter (between `---` markers) that tells Claude when to use the skill, and markdown content with instructions Claude follows when the skill is invoked. The `name` field becomes the `/slash-command`, and the `description` helps Claude decide when to load it automatically.

34 36

~~35 The `description` is especially important, because Claude uses it to decide when to apply the Skill.~~37 Create `~/.claude/skills/explain-code/SKILL.md`:

~~37 Create `~/.claude/skills/explaining-code/SKILL.md`:~~

38 38

39 ```yaml theme={null}39 ```yaml theme={null}

40 ---40 ---

~~41 name: explaining-code~~41 name: explain-code

42 description: Explains code with visual diagrams and analogies. Use when explaining how code works, teaching about a codebase, or when the user asks "how does this work?"42 description: Explains code with visual diagrams and analogies. Use when explaining how code works, teaching about a codebase, or when the user asks "how does this work?"

43 ---43 ---

44 44

53 ```53 ```

54 </Step>54 </Step>

55 55

~~56 <Step title="Load and verify the Skill">~~56 <Step title="Test the skill">

~~57 Exit and restart Claude Code to load the new Skill. Then verify it appears in the list:~~57 You can test it two ways:

59 **Let Claude invoke it automatically** by asking something that matches the description:

58 60

59 ```61 ```

~~60 What Skills are available?~~62 How does this code work?

61 ```63 ```

62 64

~~63 You should see `explaining-code` in the list with its description.~~65 **Or invoke it directly** with the skill name:

~~64 </Step>~~

~~66 <Step title="Test the Skill">~~

~~67 Open any file in your project and ask Claude a question that matches the Skill's description:~~

68 66

69 ```67 ```

~~70 How does this code work?~~68 /explain-code src/auth/login.ts

71 ```69 ```

72 70

73 Claude should ask to use the `explaining-code` Skill, then include an analogy and ASCII diagram in its explanation. If the Skill doesn't trigger, try rephrasing to include more keywords from the description, like "explain how this works."71 Either way, Claude should include an analogy and ASCII diagram in its explanation.

74 </Step>72 </Step>

75</Steps>73</Steps>

76 74

~~77The rest of this guide covers how Skills work, configuration options, and troubleshooting.~~75### Where skills live

~~79## How Skills work~~

81Skills are **model-invoked**: Claude decides which Skills to use based on your request. You don't need to explicitly call a Skill. Claude automatically applies relevant Skills when your request matches their description.

82 76

~~83When you send a request, Claude follows these steps to find and use relevant Skills:~~77Where you store a skill determines who can use it:

84 78

~~85<Steps>~~79| Location | Path | Applies to |

~~86 <Step title="Discovery">~~80| :--------- | :------------------------------------------------------- | :----------------------------- |

~~87 At startup, Claude loads only the name and description of each available Skill. This keeps startup fast while giving Claude enough context to know when each Skill might be relevant.~~81| Enterprise | See [managed settings](/en/permissions#managed-settings) | All users in your organization |

~~88 </Step>~~82| Personal | `~/.claude/skills/<skill-name>/SKILL.md` | All your projects |

89 83| Project | `.claude/skills/<skill-name>/SKILL.md` | This project only |

~~90 <Step title="Activation">~~84| Plugin | `<plugin>/skills/<skill-name>/SKILL.md` | Where plugin is enabled |

91 When your request matches a Skill's description, Claude asks to use the Skill. You'll see a confirmation prompt before the full `SKILL.md` is loaded into context. Claude matches requests against descriptions using semantic similarity, so [write descriptions](#skill-not-triggering) that include keywords users would naturally say.

~~92 </Step>~~

~~94 <Step title="Execution">~~

~~95 Claude follows the Skill's instructions, loading referenced files or running bundled scripts as needed.~~

~~96 </Step>~~

~~97</Steps>~~

98 85

~~99### Where Skills live~~86When skills share the same name across levels, higher-priority locations win: enterprise > personal > project. Plugin skills use a `plugin-name:skill-name` namespace, so they cannot conflict with other levels. If you have files in `.claude/commands/`, those work the same way, but if a skill and a command share the same name, the skill takes precedence.

100 87

101Where you store a Skill determines who can use it:88#### Automatic discovery from nested directories

102 89

103| Location | Path | Applies to |90When you work with files in subdirectories, Claude Code automatically discovers skills from nested `.claude/skills/` directories. For example, if you're editing a file in `packages/frontend/`, Claude Code also looks for skills in `packages/frontend/.claude/skills/`. This supports monorepo setups where packages have their own skills.

104| :--------- | :---------------------------------------------------------- | :-------------------------------- |

105| Enterprise | See [managed settings](/en/iam#enterprise-managed-settings) | All users in your organization |

106| Personal | `~/.claude/skills/` | You, across all projects |

107| Project | `.claude/skills/` | Anyone working in this repository |

108| Plugin | Bundled with [plugins](/en/plugins) | Anyone with the plugin installed |

109 91

110If two Skills have the same name, the higher row wins: enterprise overrides personal, personal overrides project, and project overrides plugin.92Each skill is a directory with `SKILL.md` as the entrypoint:

111 93

112### When to use Skills versus other options94```

95my-skill/

96├── SKILL.md # Main instructions (required)

97├── template.md # Template for Claude to fill in

98├── examples/

99│ └── sample.md # Example output showing expected format

100└── scripts/

101 └── validate.sh # Script Claude can execute

102```

113 103

114Claude Code offers several ways to customize behavior. The key difference: **Skills are triggered automatically by Claude** based on your request, while slash commands require you to type `/command` explicitly.104The `SKILL.md` contains the main instructions and is required. Other files are optional and let you build more powerful skills: templates for Claude to fill in, example outputs showing the expected format, scripts Claude can execute, or detailed reference documentation. Reference these files from your `SKILL.md` so Claude knows what they contain and when to load them. See [Add supporting files](#add-supporting-files) for more details.

115 105

116| Use this | When you want to... | When it runs |106<Note>

117| :--------------------------------------- | :------------------------------------------------------------------------- | :----------------------------------------- |107 Files in `.claude/commands/` still work and support the same [frontmatter](#frontmatter-reference). Skills are recommended since they support additional features like supporting files.

118| **Skills** | Give Claude specialized knowledge (e.g., "review PRs using our standards") | Claude chooses when relevant |108</Note>

119| **[Slash commands](/en/slash-commands)** | Create reusable prompts (e.g., `/deploy staging`) | You type `/command` to run it |

120| **[CLAUDE.md](/en/memory)** | Set project-wide instructions (e.g., "use TypeScript strict mode") | Loaded into every conversation |

121| **[Subagents](/en/sub-agents)** | Delegate tasks to a separate context with its own tools | Claude delegates, or you invoke explicitly |

122| **[Hooks](/en/hooks)** | Run scripts on events (e.g., lint on file save) | Fires on specific tool events |

123| **[MCP servers](/en/mcp)** | Connect Claude to external tools and data sources | Claude calls MCP tools as needed |

124 109

125**Skills vs. subagents**: Skills add knowledge to the current conversation. Subagents run in a separate context with their own tools. Use Skills for guidance and standards; use subagents when you need isolation or different tool access.110#### Skills from additional directories

126 111

127**Skills vs. MCP**: Skills tell Claude *how* to use tools; MCP *provides* the tools. For example, an MCP server connects Claude to your database, while a Skill teaches Claude your data model and query patterns.112Skills defined in `.claude/skills/` within directories added via `--add-dir` are loaded automatically and picked up by live change detection, so you can edit them during a session without restarting.

128 113

129<Note>114<Note>

130 For a deep dive into the architecture and real-world applications of Agent Skills, read [Equipping agents for the real world with Agent Skills](https://www.anthropic.com/engineering/equipping-agents-for-the-real-world-with-agent-skills).115 CLAUDE.md files from `--add-dir` directories are not loaded by default. To load them, set `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`. See [Load memory from additional directories](/en/memory#load-memory-from-additional-directories).

131</Note>116</Note>

132 117

133## Configure Skills118## Configure skills

119

120Skills are configured through YAML frontmatter at the top of `SKILL.md` and the markdown content that follows.

134 121

135This section covers Skill file structure, supporting files, tool restrictions, and distribution options.122### Types of skill content

136 123

137### Write SKILL.md124Skill files can contain any instructions, but thinking about how you want to invoke them helps guide what to include:

138 125

139The `SKILL.md` file is the only required file in a Skill. It has two parts: YAML metadata (the section between `---` markers) at the top, and Markdown instructions that tell Claude how to use the Skill:126**Reference content** adds knowledge Claude applies to your current work. Conventions, patterns, style guides, domain knowledge. This content runs inline so Claude can use it alongside your conversation context.

140 127

141```yaml theme={null}128```yaml theme={null}

142---129---

143name: your-skill-name130name: api-conventions

144description: Brief description of what this Skill does and when to use it131description: API design patterns for this codebase

145---132---

146 133

147# Your Skill Name134When writing API endpoints:

135- Use RESTful naming conventions

136- Return consistent error formats

137- Include request validation

138```

139

140**Task content** gives Claude step-by-step instructions for a specific action, like deployments, commits, or code generation. These are often actions you want to invoke directly with `/skill-name` rather than letting Claude decide when to run them. Add `disable-model-invocation: true` to prevent Claude from triggering it automatically.

148 141

149## Instructions142```yaml theme={null}

150Provide clear, step-by-step guidance for Claude.143---

144name: deploy

145description: Deploy the application to production

146context: fork

147disable-model-invocation: true

148---

151 149

152## Examples150Deploy the application:

153Show concrete examples of using this Skill.1511. Run the test suite

1522. Build the application

1533. Push to the deployment target

154```154```

155 155

156#### Available metadata fields156Your `SKILL.md` can contain anything, but thinking through how you want the skill invoked (by you, by Claude, or both) and where you want it to run (inline or in a subagent) helps guide what to include. For complex skills, you can also [add supporting files](#add-supporting-files) to keep the main skill focused.

157 157

158You can use the following fields in the YAML frontmatter:158### Frontmatter reference

159 159

160| Field | Required | Description |160Beyond the markdown content, you can configure skill behavior using YAML frontmatter fields between `---` markers at the top of your `SKILL.md` file:

161| :-------------- | :------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

162| `name` | Yes | Skill name. Must use lowercase letters, numbers, and hyphens only (max 64 characters). Should match the directory name. |

163| `description` | Yes | What the Skill does and when to use it (max 1024 characters). Claude uses this to decide when to apply the Skill. |

164| `allowed-tools` | No | Tools Claude can use without asking permission when this Skill is active. See [Restrict tool access](#restrict-tool-access-with-allowed-tools). |

165| `model` | No | [Model](https://docs.claude.com/en/docs/about-claude/models/overview) to use when this Skill is active (e.g., `claude-sonnet-4-20250514`). Defaults to the conversation's model. |

166 161

167See the [best practices guide](https://docs.claude.com/en/docs/agents-and-tools/agent-skills/best-practices) for complete authoring guidance including validation rules.162```yaml theme={null}

163---

164name: my-skill

165description: What this skill does

166disable-model-invocation: true

167allowed-tools: Read, Grep

168---

168 169

169### Update or delete a Skill170Your skill instructions here...

171```

170 172

171To update a Skill, edit its `SKILL.md` file directly. To remove a Skill, delete its directory. Exit and restart Claude Code for changes to take effect.173All fields are optional. Only `description` is recommended so Claude knows when to use the skill.

172 174

173### Add supporting files with progressive disclosure175| Field | Required | Description |

176| :------------------------- | :---------- | :---------------------------------------------------------------------------------------------------------------------------------------------------- |

177| `name` | No | Display name for the skill. If omitted, uses the directory name. Lowercase letters, numbers, and hyphens only (max 64 characters). |

178| `description` | Recommended | What the skill does and when to use it. Claude uses this to decide when to apply the skill. If omitted, uses the first paragraph of markdown content. |

179| `argument-hint` | No | Hint shown during autocomplete to indicate expected arguments. Example: `[issue-number]` or `[filename] [format]`. |

180| `disable-model-invocation` | No | Set to `true` to prevent Claude from automatically loading this skill. Use for workflows you want to trigger manually with `/name`. Default: `false`. |

181| `user-invocable` | No | Set to `false` to hide from the `/` menu. Use for background knowledge users shouldn't invoke directly. Default: `true`. |

182| `allowed-tools` | No | Tools Claude can use without asking permission when this skill is active. |

183| `model` | No | Model to use when this skill is active. |

184| `context` | No | Set to `fork` to run in a forked subagent context. |

185| `agent` | No | Which subagent type to use when `context: fork` is set. |

186| `hooks` | No | Hooks scoped to this skill's lifecycle. See [Hooks in skills and agents](/en/hooks#hooks-in-skills-and-agents) for configuration format. |

187

188#### Available string substitutions

189

190Skills support string substitution for dynamic values in the skill content:

191

192| Variable | Description |

193| :--------------------- | :------------------------------------------------------------------------------------------------------------------------------------------- |

194| `$ARGUMENTS` | All arguments passed when invoking the skill. If `$ARGUMENTS` is not present in the content, arguments are appended as `ARGUMENTS: <value>`. |

195| `$ARGUMENTS[N]` | Access a specific argument by 0-based index, such as `$ARGUMENTS[0]` for the first argument. |

196| `$N` | Shorthand for `$ARGUMENTS[N]`, such as `$0` for the first argument or `$1` for the second. |

197| `${CLAUDE_SESSION_ID}` | The current session ID. Useful for logging, creating session-specific files, or correlating skill output with sessions. |

198

199**Example using substitutions:**

174 200

175Skills share Claude's context window with conversation history, other Skills, and your request. To keep context focused, use **progressive disclosure**: put essential information in `SKILL.md` and detailed reference material in separate files that Claude reads only when needed.201```yaml theme={null}

202---

203name: session-logger

204description: Log activity for this session

205---

176 206

177This approach lets you bundle comprehensive documentation, examples, and scripts without consuming context upfront. Claude loads additional files only when the task requires them.207Log the following to logs/${CLAUDE_SESSION_ID}.log:

178 208

179<Tip>Keep `SKILL.md` under 500 lines for optimal performance. If your content exceeds this, split detailed reference material into separate files.</Tip>209$ARGUMENTS

210```

180 211

181#### Example: multi-file Skill structure212### Add supporting files

182 213

183Claude discovers supporting files through links in your `SKILL.md`. The following example shows a Skill with detailed documentation in separate files and utility scripts that Claude can execute without reading:214Skills can include multiple files in their directory. This keeps `SKILL.md` focused on the essentials while letting Claude access detailed reference material only when needed. Large reference docs, API specifications, or example collections don't need to load into context every time the skill runs.

184 215

185```216```

186my-skill/217my-skill/

191 └── helper.py (utility script - executed, not loaded)222 └── helper.py (utility script - executed, not loaded)

192```223```

193 224

194The `SKILL.md` file references these supporting files so Claude knows they exist:225Reference supporting files from `SKILL.md` so Claude knows what each file contains and when to load it:

~~195~~

196````markdown theme={null}

197## Overview

~~198~~

199[Essential instructions here]

200 226

227```markdown theme={null}

201## Additional resources228## Additional resources

202 229

203- For complete API details, see [reference.md](reference.md)230- For complete API details, see [reference.md](reference.md)

204- For usage examples, see [examples.md](examples.md)231- For usage examples, see [examples.md](examples.md)

~~205~~

206## Utility scripts

~~207~~

208To validate input files, run the helper script. It checks for required fields and returns any validation errors:

209```bash

210python scripts/helper.py input.txt

211```232```

212````

213 233

214<Tip>Keep references one level deep. Link directly from `SKILL.md` to reference files. Deeply nested references (file A links to file B which links to file C) may result in Claude partially reading files.</Tip>234<Tip>Keep `SKILL.md` under 500 lines. Move detailed reference material to separate files.</Tip>

215 235

216**Bundle utility scripts for zero-context execution.** Scripts in your Skill directory can be executed without loading their contents into context. Claude runs the script and only the output consumes tokens. This is useful for:236### Control who invokes a skill

217 237

218* Complex validation logic that would be verbose to describe in prose238By default, both you and Claude can invoke any skill. You can type `/skill-name` to invoke it directly, and Claude can load it automatically when relevant to your conversation. Two frontmatter fields let you restrict this:

219* Data processing that's more reliable as tested code than generated code

220* Operations that benefit from consistency across uses

221 239

222In `SKILL.md`, tell Claude to run the script rather than read it:240* **`disable-model-invocation: true`**: Only you can invoke the skill. Use this for workflows with side effects or that you want to control timing, like `/commit`, `/deploy`, or `/send-slack-message`. You don't want Claude deciding to deploy because your code looks ready.

223 241

224```markdown theme={null}242* **`user-invocable: false`**: Only Claude can invoke the skill. Use this for background knowledge that isn't actionable as a command. A `legacy-system-context` skill explains how an old system works. Claude should know this when relevant, but `/legacy-system-context` isn't a meaningful action for users to take.

225Run the validation script to check the form:

226python scripts/validate_form.py input.pdf

227```

228 243

229For complete guidance on structuring Skills, see the [best practices guide](https://docs.claude.com/en/docs/agents-and-tools/agent-skills/best-practices#progressive-disclosure-patterns).244This example creates a deploy skill that only you can trigger. The `disable-model-invocation: true` field prevents Claude from running it automatically:

~~230~~

231### Restrict tool access with allowed-tools

~~232~~

233Use the `allowed-tools` frontmatter field to limit which tools Claude can use when a Skill is active:

234 245

235```yaml theme={null}246```yaml theme={null}

236---247---

237name: reading-files-safely248name: deploy

238description: Read files without making changes. Use when you need read-only file access.249description: Deploy the application to production

239allowed-tools: Read, Grep, Glob250disable-model-invocation: true

240---251---

241 252

242# Safe File Reader253Deploy $ARGUMENTS to production:

~~243~~

244This Skill provides read-only file access.

245 254

246## Instructions2551. Run the test suite

2471. Use Read to view file contents2562. Build the application

2482. Use Grep to search within files2573. Push to the deployment target

2493. Use Glob to find files by pattern2584. Verify the deployment succeeded

250```259```

251 260

252When this Skill is active, Claude can only use the specified tools (Read, Grep, Glob) without needing to ask for permission. This is useful for:261Here's how the two fields affect invocation and context loading:

253 262

255* Skills with limited scope: for example, only data analysis, no file writing264| :------------------------------- | :------------- | :---------------- | :----------------------------------------------------------- |

256* Security-sensitive workflows where you want to restrict capabilities265| (default) | Yes | Yes | Description always in context, full skill loads when invoked |

~~257~~ 266| `disable-model-invocation: true` | Yes | No | Description not in context, full skill loads when you invoke |

258If `allowed-tools` is omitted, the Skill doesn't restrict tools. Claude uses its standard permission model and may ask you to approve tool usage.267| `user-invocable: false` | No | Yes | Description always in context, full skill loads when invoked |

259 268

260<Note>269<Note>

261 `allowed-tools` is only supported for Skills in Claude Code.270 In a regular session, skill descriptions are loaded into context so Claude knows what's available, but full skill content only loads when invoked. [Subagents with preloaded skills](/en/sub-agents#preload-skills-into-subagents) work differently: the full skill content is injected at startup.

262</Note>271</Note>

263 272

264### Use Skills with subagents273### Restrict tool access

265 274

266[Subagents](/en/sub-agents) do not automatically inherit Skills from the main conversation. To give a custom subagent access to specific Skills, list them in the subagent's `skills` field in `.claude/agents/`:275Use the `allowed-tools` field to limit which tools Claude can use when a skill is active. This skill creates a read-only mode where Claude can explore files but not modify them:

267 276

268```yaml theme={null}277```yaml theme={null}

269# .claude/agents/code-reviewer/AGENT.md

270---278---

271name: code-reviewer279name: safe-reader

272description: Review code for quality and best practices280description: Read files without making changes

273skills: pr-review, security-check281allowed-tools: Read, Grep, Glob

274---282---

275```283```

276 284

277The listed Skills are loaded into the subagent's context when it starts. If the `skills` field is omitted, no Skills are preloaded for that subagent.285### Pass arguments to skills

278 286

279<Note>287Both you and Claude can pass arguments when invoking a skill. Arguments are available via the `$ARGUMENTS` placeholder.

280 Built-in agents (Explore, Plan, Verify) and the Task tool do not have access to your Skills. Only custom subagents you define in `.claude/agents/` with an explicit `skills` field can use Skills.

281</Note>

~~282~~

283### Distribute Skills

284 288

285You can share Skills in several ways:289This skill fixes a GitHub issue by number. The `$ARGUMENTS` placeholder gets replaced with whatever follows the skill name:

286 290

287* **Project Skills**: Commit `.claude/skills/` to version control. Anyone who clones the repository gets the Skills.291```yaml theme={null}

288* **Plugins**: To share Skills across multiple repositories, create a `skills/` directory in your [plugin](/en/plugins) with Skill folders containing `SKILL.md` files. Distribute through a [plugin marketplace](/en/plugin-marketplaces).292---

289* **Enterprise**: Administrators can deploy Skills organization-wide through [managed settings](/en/iam#enterprise-managed-settings). See [Where Skills live](#where-skills-live) for enterprise Skill paths.293name: fix-issue

294description: Fix a GitHub issue

295disable-model-invocation: true

296---

290 297

291## Examples298Fix GitHub issue $ARGUMENTS following our coding standards.

292 299

293These examples show common Skill patterns, from minimal single-file Skills to multi-file Skills with supporting documentation and scripts.3001. Read the issue description

3012. Understand the requirements

3023. Implement the fix

3034. Write tests

3045. Create a commit

305```

294 306

295### Simple Skill (single file)307When you run `/fix-issue 123`, Claude receives "Fix GitHub issue 123 following our coding standards..."

296 308

297A minimal Skill needs only a `SKILL.md` file with frontmatter and instructions. This example helps Claude generate commit messages by examining staged changes:309If you invoke a skill with arguments but the skill doesn't include `$ARGUMENTS`, Claude Code appends `ARGUMENTS: <your input>` to the end of the skill content so Claude still sees what you typed.

298 310

299```311To access individual arguments by position, use `$ARGUMENTS[N]` or the shorter `$N`:

300commit-helper/

301└── SKILL.md

302```

303 312

304```yaml theme={null}313```yaml theme={null}

305---314---

306name: generating-commit-messages315name: migrate-component

307description: Generates clear commit messages from git diffs. Use when writing commit messages or reviewing staged changes.316description: Migrate a component from one framework to another

308---317---

309 318

310# Generating Commit Messages319Migrate the $ARGUMENTS[0] component from $ARGUMENTS[1] to $ARGUMENTS[2].

~~311~~ 320Preserve all existing behavior and tests.

312## Instructions321```

313 322

3141. Run `git diff --staged` to see changes323Running `/migrate-component SearchBar React Vue` replaces `$ARGUMENTS[0]` with `SearchBar`, `$ARGUMENTS[1]` with `React`, and `$ARGUMENTS[2]` with `Vue`. The same skill using the `$N` shorthand:

3152. I'll suggest a commit message with:

316 - Summary under 50 characters

317 - Detailed description

318 - Affected components

319 324

320## Best practices325```yaml theme={null}

326---

327name: migrate-component

328description: Migrate a component from one framework to another

329---

321 330

322- Use present tense331Migrate the $0 component from $1 to $2.

323- Explain what and why, not how332Preserve all existing behavior and tests.

324```333```

325 334

326### Use multiple files335## Advanced patterns

327 336

328For complex Skills, use progressive disclosure to keep the main `SKILL.md` focused while providing detailed documentation in supporting files. This PDF processing Skill includes reference docs, utility scripts, and uses `allowed-tools` to restrict Claude to specific tools:337### Inject dynamic context

329 338

330```339The `!`command\`\` syntax runs shell commands before the skill content is sent to Claude. The command output replaces the placeholder, so Claude receives actual data, not the command itself.

331pdf-processing/

332├── SKILL.md # Overview and quick start

333├── FORMS.md # Form field mappings and filling instructions

334├── REFERENCE.md # API details for pypdf and pdfplumber

335└── scripts/

336 ├── fill_form.py # Utility to populate form fields

337 └── validate.py # Checks PDFs for required fields

338```

339 340

340**`SKILL.md`**:341This skill summarizes a pull request by fetching live PR data with the GitHub CLI. The `!`gh pr diff\`\` and other commands run first, and their output gets inserted into the prompt:

341 342

342````yaml theme={null}343```yaml theme={null}

343---344---

344name: pdf-processing345name: pr-summary

345description: Extract text, fill forms, merge PDFs. Use when working with PDF files, forms, or document extraction. Requires pypdf and pdfplumber packages.346description: Summarize changes in a pull request

346allowed-tools: Read, Bash(python:*)347context: fork

348agent: Explore

349allowed-tools: Bash(gh *)

347---350---

348 351

349# PDF Processing352## Pull request context

353- PR diff: !`gh pr diff`

354- PR comments: !`gh pr view --comments`

355- Changed files: !`gh pr diff --name-only`

350 356

351## Quick start357## Your task

~~352~~ 358Summarize this pull request...

353Extract text:

354```python

355import pdfplumber

356with pdfplumber.open("doc.pdf") as pdf:

357 text = pdf.pages[0].extract_text()

358```359```

359 360

360For form filling, see [FORMS.md](FORMS.md).361When this skill runs:

361For detailed API reference, see [REFERENCE.md](REFERENCE.md).

362 362

363## Requirements3631. Each `!`command\`\` executes immediately (before Claude sees anything)

3642. The output replaces the placeholder in the skill content

3653. Claude receives the fully-rendered prompt with actual PR data

364 366

365Packages must be installed in your environment:367This is preprocessing, not something Claude executes. Claude only sees the final result.

366```bash

367pip install pypdf pdfplumber

368```

369````

370 368

371<Note>369<Tip>

372 If your Skill requires external packages, list them in the description. Packages must be installed in your environment before Claude can use them.370 To enable [extended thinking](/en/common-workflows#use-extended-thinking-thinking-mode) in a skill, include the word "ultrathink" anywhere in your skill content.

373</Note>371</Tip>

374 372

375## Troubleshooting373### Run skills in a subagent

376 374

377### View and test Skills375Add `context: fork` to your frontmatter when you want a skill to run in isolation. The skill content becomes the prompt that drives the subagent. It won't have access to your conversation history.

378 376

379To see which Skills Claude has access to, ask Claude a question like "What Skills are available?" Claude loads all available Skill names and descriptions into the context window when a conversation starts, so it can list the Skills it currently has access to.377<Warning>

378 `context: fork` only makes sense for skills with explicit instructions. If your skill contains guidelines like "use these API conventions" without a task, the subagent receives the guidelines but no actionable prompt, and returns without meaningful output.

379</Warning>

380 380

381To test a specific Skill, ask Claude to do a task that matches the Skill's description. For example, if your Skill has the description "Reviews pull requests for code quality", ask Claude to "Review the changes in my current branch." Claude automatically uses the Skill when the request matches its description.381Skills and [subagents](/en/sub-agents) work together in two directions:

382 382

384| :--------------------------- | :---------------------------------------- | :-------------------------- | :--------------------------- |

384 387

385The description field is how Claude decides whether to use your Skill. Vague descriptions like "Helps with documents" don't give Claude enough information to match your Skill to relevant requests.388With `context: fork`, you write the task in your skill and pick an agent type to execute it. For the inverse (defining a custom subagent that uses skills as reference material), see [Subagents](/en/sub-agents#preload-skills-into-subagents).

386 389

387A good description answers two questions:390#### Example: Research skill using Explore agent

388 391

3891. **What does this Skill do?** List the specific capabilities.392This skill runs research in a forked Explore agent. The skill content becomes the task, and the agent provides read-only tools optimized for codebase exploration:

3902. **When should Claude use it?** Include trigger terms users would mention.

391 393

392```yaml theme={null}394```yaml theme={null}

393description: Extract text and tables from PDF files, fill forms, merge documents. Use when working with PDF files or when the user mentions PDFs, forms, or document extraction.395---

396name: deep-research

397description: Research a topic thoroughly

398context: fork

399agent: Explore

400---

401

402Research $ARGUMENTS thoroughly:

403

4041. Find relevant files using Glob and Grep

4052. Read and analyze the code

4063. Summarize findings with specific file references

394```407```

395 408

396This description works because it names specific actions (extract, fill, merge) and includes keywords users would say (PDF, forms, document extraction).409When this skill runs:

410

4111. A new isolated context is created

4122. The subagent receives the skill content as its prompt ("Research \$ARGUMENTS thoroughly...")

4133. The `agent` field determines the execution environment (model, tools, and permissions)

4144. Results are summarized and returned to your main conversation

397 415

398### Skill doesn't load416The `agent` field specifies which subagent configuration to use. Options include built-in agents (`Explore`, `Plan`, `general-purpose`) or any custom subagent from `.claude/agents/`. If omitted, uses `general-purpose`.

399 417

400**Check the file path.** Skills must be in the correct directory with the exact filename `SKILL.md` (case-sensitive):418### Restrict Claude's skill access

401 419

402| Type | Path |420By default, Claude can invoke any skill that doesn't have `disable-model-invocation: true` set. Skills that define `allowed-tools` grant Claude access to those tools without per-use approval when the skill is active. Your [permission settings](/en/permissions) still govern baseline approval behavior for all other tools. Built-in commands like `/compact` and `/init` are not available through the Skill tool.

403| :--------- | :---------------------------------------------------------------------- |

404| Personal | `~/.claude/skills/my-skill/SKILL.md` |

405| Project | `.claude/skills/my-skill/SKILL.md` |

406| Enterprise | See [Where Skills live](#where-skills-live) for platform-specific paths |

407| Plugin | `skills/my-skill/SKILL.md` inside the plugin directory |

408 421

409**Check the YAML syntax.** Invalid YAML in the frontmatter prevents the Skill from loading. The frontmatter must start with `---` on line 1 (no blank lines before it), end with `---` before the Markdown content, and use spaces for indentation (not tabs).422Three ways to control which skills Claude can invoke:

410 423

411**Run debug mode.** Use `claude --debug` to see Skill loading errors.424**Disable all skills** by denying the Skill tool in `/permissions`:

412 425

413### Skill has errors426```

427# Add to deny rules:

428Skill

429```

430

431**Allow or deny specific skills** using [permission rules](/en/permissions):

432

433```

434# Allow only specific skills

435Skill(commit)

436Skill(review-pr *)

437

438# Deny specific skills

439Skill(deploy *)

440```

441

442Permission syntax: `Skill(name)` for exact match, `Skill(name *)` for prefix match with any arguments.

414 443

415**Check dependencies are installed.** If your Skill uses external packages, they must be installed in your environment before Claude can use them.444**Hide individual skills** by adding `disable-model-invocation: true` to their frontmatter. This removes the skill from Claude's context entirely.

416 445

417**Check script permissions.** Scripts need execute permissions: `chmod +x scripts/*.py`446<Note>

447 The `user-invocable` field only controls menu visibility, not Skill tool access. Use `disable-model-invocation: true` to block programmatic invocation.

448</Note>

418 449

419**Check file paths.** Use forward slashes (Unix style) in all paths. Use `scripts/helper.py`, not `scripts\helper.py`.450## Share skills

420 451

421### Multiple Skills conflict452Skills can be distributed at different scopes depending on your audience:

422 453

423If Claude uses the wrong Skill or seems confused between similar Skills, the descriptions are probably too similar. Make each description distinct by using specific trigger terms.454* **Project skills**: Commit `.claude/skills/` to version control

455* **Plugins**: Create a `skills/` directory in your [plugin](/en/plugins)

456* **Managed**: Deploy organization-wide through [managed settings](/en/permissions#managed-settings)

424 457

425For example, instead of two Skills with "data analysis" in both descriptions, differentiate them: one for "sales data in Excel files and CRM exports" and another for "log files and system metrics". The more specific your trigger terms, the easier it is for Claude to match the right Skill to your request.458### Generate visual output

426 459

427### Plugin Skills not appearing460Skills can bundle and run scripts in any language, giving Claude capabilities beyond what's possible in a single prompt. One powerful pattern is generating visual output: interactive HTML files that open in your browser for exploring data, debugging, or creating reports.

428 461

429**Symptom**: You installed a plugin from a marketplace, but its Skills don't appear when you ask Claude "What Skills are available?"462This example creates a codebase explorer: an interactive tree view where you can expand and collapse directories, see file sizes at a glance, and identify file types by color.

430 463

431**Solution**: Clear the plugin cache and reinstall:464Create the Skill directory:

432 465

433```bash theme={null}466```bash theme={null}

434rm -rf ~/.claude/plugins/cache467mkdir -p ~/.claude/skills/codebase-visualizer/scripts

435```468```

436 469

437Then restart Claude Code and reinstall the plugin:470Create `~/.claude/skills/codebase-visualizer/SKILL.md`. The description tells Claude when to activate this Skill, and the instructions tell Claude to run the bundled script:

438 471

439```shell theme={null}472````yaml theme={null}

440/plugin install plugin-name@marketplace-name473---

441```474name: codebase-visualizer

475description: Generate an interactive collapsible tree visualization of your codebase. Use when exploring a new repo, understanding project structure, or identifying large files.

476allowed-tools: Bash(python *)

477---

478

479# Codebase Visualizer

480

481Generate an interactive HTML tree view that shows your project's file structure with collapsible directories.

442 482

443This forces Claude Code to re-download and re-register the plugin's Skills.483## Usage

444 484

445**If Skills still don't appear**, verify the plugin's directory structure is correct. Skills must be in a `skills/` directory at the plugin root:485Run the visualization script from your project root:

446 486

487```bash

488python ~/.claude/skills/codebase-visualizer/scripts/visualize.py .

447```489```

448my-plugin/490

449├── .claude-plugin/491This creates `codebase-map.html` in the current directory and opens it in your default browser.

450│ └── plugin.json492

451└── skills/493## What the visualization shows

452 └── my-skill/494

453 └── SKILL.md495- **Collapsible directories**: Click folders to expand/collapse

496- **File sizes**: Displayed next to each file

497- **Colors**: Different colors for different file types

498- **Directory totals**: Shows aggregate size of each folder

499````

500

501Create `~/.claude/skills/codebase-visualizer/scripts/visualize.py`. This script scans a directory tree and generates a self-contained HTML file with:

502

503* A **summary sidebar** showing file count, directory count, total size, and number of file types

504* A **bar chart** breaking down the codebase by file type (top 8 by size)

505* A **collapsible tree** where you can expand and collapse directories, with color-coded file type indicators

506

507The script requires Python but uses only built-in libraries, so there are no packages to install:

508

509```python expandable theme={null}

510#!/usr/bin/env python3

511"""Generate an interactive collapsible tree visualization of a codebase."""

512

513import json

514import sys

515import webbrowser

516from pathlib import Path

517from collections import Counter

518

519IGNORE = {'.git', 'node_modules', '__pycache__', '.venv', 'venv', 'dist', 'build'}

520

521def scan(path: Path, stats: dict) -> dict:

522 result = {"name": path.name, "children": [], "size": 0}

523 try:

524 for item in sorted(path.iterdir()):

525 if item.name in IGNORE or item.name.startswith('.'):

526 continue

527 if item.is_file():

528 size = item.stat().st_size

529 ext = item.suffix.lower() or '(no ext)'

530 result["children"].append({"name": item.name, "size": size, "ext": ext})

531 result["size"] += size

532 stats["files"] += 1

533 stats["extensions"][ext] += 1

534 stats["ext_sizes"][ext] += size

535 elif item.is_dir():

536 stats["dirs"] += 1

537 child = scan(item, stats)

538 if child["children"]:

539 result["children"].append(child)

540 result["size"] += child["size"]

541 except PermissionError:

542 pass

543 return result

544

545def generate_html(data: dict, stats: dict, output: Path) -> None:

546 ext_sizes = stats["ext_sizes"]

547 total_size = sum(ext_sizes.values()) or 1

548 sorted_exts = sorted(ext_sizes.items(), key=lambda x: -x[1])[:8]

549 colors = {

550 '.js': '#f7df1e', '.ts': '#3178c6', '.py': '#3776ab', '.go': '#00add8',

551 '.rs': '#dea584', '.rb': '#cc342d', '.css': '#264de4', '.html': '#e34c26',

552 '.json': '#6b7280', '.md': '#083fa1', '.yaml': '#cb171e', '.yml': '#cb171e',

553 '.mdx': '#083fa1', '.tsx': '#3178c6', '.jsx': '#61dafb', '.sh': '#4eaa25',

554 }

555 lang_bars = "".join(

556 f'<div class="bar-row"><span class="bar-label">{ext}</span>'

557 f'<div class="bar" style="width:{(size/total_size)*100}%;background:{colors.get(ext,"#6b7280")}"></div>'

558 f'<span class="bar-pct">{(size/total_size)*100:.1f}%</span></div>'

559 for ext, size in sorted_exts

560 )

561 def fmt(b):

562 if b < 1024: return f"{b} B"

563 if b < 1048576: return f"{b/1024:.1f} KB"

564 return f"{b/1048576:.1f} MB"

565

566 html = f'''<!DOCTYPE html>

567<html><head>

568 <meta charset="utf-8"><title>Codebase Explorer</title>

569 <style>

570 body {{ font: 14px/1.5 system-ui, sans-serif; margin: 0; background: #1a1a2e; color: #eee; }}

571 .container {{ display: flex; height: 100vh; }}

572 .sidebar {{ width: 280px; background: #252542; padding: 20px; border-right: 1px solid #3d3d5c; overflow-y: auto; flex-shrink: 0; }}

573 .main {{ flex: 1; padding: 20px; overflow-y: auto; }}

574 h1 {{ margin: 0 0 10px 0; font-size: 18px; }}

575 h2 {{ margin: 20px 0 10px 0; font-size: 14px; color: #888; text-transform: uppercase; }}

576 .stat {{ display: flex; justify-content: space-between; padding: 8px 0; border-bottom: 1px solid #3d3d5c; }}

577 .stat-value {{ font-weight: bold; }}

578 .bar-row {{ display: flex; align-items: center; margin: 6px 0; }}

579 .bar-label {{ width: 55px; font-size: 12px; color: #aaa; }}

580 .bar {{ height: 18px; border-radius: 3px; }}

581 .bar-pct {{ margin-left: 8px; font-size: 12px; color: #666; }}

582 .tree {{ list-style: none; padding-left: 20px; }}

583 details {{ cursor: pointer; }}

584 summary {{ padding: 4px 8px; border-radius: 4px; }}

585 summary:hover {{ background: #2d2d44; }}

586 .folder {{ color: #ffd700; }}

587 .file {{ display: flex; align-items: center; padding: 4px 8px; border-radius: 4px; }}

588 .file:hover {{ background: #2d2d44; }}

589 .size {{ color: #888; margin-left: auto; font-size: 12px; }}

590 .dot {{ width: 8px; height: 8px; border-radius: 50%; margin-right: 8px; }}

591 </style>

592</head><body>

593 <div class="container">

594 <div class="sidebar">

595 <h1>📊 Summary</h1>

596 <div class="stat"><span>Files</span><span class="stat-value">{stats["files"]:,}</span></div>

597 <div class="stat"><span>Directories</span><span class="stat-value">{stats["dirs"]:,}</span></div>

598 <div class="stat"><span>Total size</span><span class="stat-value">{fmt(data["size"])}</span></div>

599 <div class="stat"><span>File types</span><span class="stat-value">{len(stats["extensions"])}</span></div>

600 <h2>By file type</h2>

601 {lang_bars}

602 </div>

603 <div class="main">

604 <h1>📁 {data["name"]}</h1>

605 <ul class="tree" id="root"></ul>

606 </div>

607 </div>

608 <script>

609 const data = {json.dumps(data)};

610 const colors = {json.dumps(colors)};

611 function fmt(b) {{ if (b < 1024) return b + ' B'; if (b < 1048576) return (b/1024).toFixed(1) + ' KB'; return (b/1048576).toFixed(1) + ' MB'; }}

612 function render(node, parent) {{

613 if (node.children) {{

614 const det = document.createElement('details');

615 det.open = parent === document.getElementById('root');

616 det.innerHTML = `<summary><span class="folder">📁 ${{node.name}}</span><span class="size">${{fmt(node.size)}}</span></summary>`;

617 const ul = document.createElement('ul'); ul.className = 'tree';

618 node.children.sort((a,b) => (b.children?1:0)-(a.children?1:0) || a.name.localeCompare(b.name));

619 node.children.forEach(c => render(c, ul));

620 det.appendChild(ul);

621 const li = document.createElement('li'); li.appendChild(det); parent.appendChild(li);

622 }} else {{

623 const li = document.createElement('li'); li.className = 'file';

624 li.innerHTML = `<span class="dot" style="background:${{colors[node.ext]||'#6b7280'}}"></span>${{node.name}}<span class="size">${{fmt(node.size)}}</span>`;

625 parent.appendChild(li);

626 }}

627 }}

628 data.children.forEach(c => render(c, document.getElementById('root')));

629 </script>

630</body></html>'''

631 output.write_text(html)

632

633if __name__ == '__main__':

634 target = Path(sys.argv[1] if len(sys.argv) > 1 else '.').resolve()

635 stats = {"files": 0, "dirs": 0, "extensions": Counter(), "ext_sizes": Counter()}

636 data = scan(target, stats)

637 out = Path('codebase-map.html')

638 generate_html(data, stats, out)

639 print(f'Generated {out.absolute()}')

640 webbrowser.open(f'file://{out.absolute()}')

454```641```

455 642

456## Next steps643To test, open Claude Code in any project and ask "Visualize this codebase." Claude runs the script, generates `codebase-map.html`, and opens it in your browser.

457 644

458<CardGroup cols={2}>645This pattern works for any visual output: dependency graphs, test coverage reports, API documentation, or database schema visualizations. The bundled script does the heavy lifting while Claude handles orchestration.

459 <Card title="Authoring best practices" icon="lightbulb" href="https://docs.claude.com/en/docs/agents-and-tools/agent-skills/best-practices">

460 Write Skills that Claude can use effectively

461 </Card>

462 646

463 <Card title="Agent Skills overview" icon="book" href="https://docs.claude.com/en/docs/agents-and-tools/agent-skills/overview">647## Troubleshooting

464 Learn how Skills work across Claude products

465 </Card>

466 648

467 <Card title="Use Skills in the Agent SDK" icon="cube" href="https://docs.claude.com/en/docs/agent-sdk/skills">649### Skill not triggering

468 Use Skills programmatically with TypeScript and Python

469 </Card>

470 650

471 <Card title="Get started with Agent Skills" icon="rocket" href="https://docs.claude.com/en/docs/agents-and-tools/agent-skills/quickstart">651If Claude doesn't use your skill when expected:

472 Create your first Skill

473 </Card>

474</CardGroup>

475 652

6531. Check the description includes keywords users would naturally say

6542. Verify the skill appears in `What skills are available?`

6553. Try rephrasing your request to match the description more closely

6564. Invoke it directly with `/skill-name` if the skill is user-invocable

476 657

658### Skill triggers too often

659

660If Claude uses your skill when you don't want it:

661

6621. Make the description more specific

6632. Add `disable-model-invocation: true` if you only want manual invocation

664

665### Claude doesn't see all my skills

666

667Skill descriptions are loaded into context so Claude knows what's available. If you have many skills, they may exceed the character budget. The budget scales dynamically at 2% of the context window, with a fallback of 16,000 characters. Run `/context` to check for a warning about excluded skills.

668

669To override the limit, set the `SLASH_COMMAND_TOOL_CHAR_BUDGET` environment variable.

670

671## Related resources

477 672

478> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt673* **[Subagents](/en/sub-agents)**: delegate tasks to specialized agents

674* **[Plugins](/en/plugins)**: package and distribute skills with other extensions

675* **[Hooks](/en/hooks)**: automate workflows around tool events

676* **[Memory](/en/memory)**: manage CLAUDE.md files for persistent context

677* **[Interactive mode](/en/interactive-mode#built-in-commands)**: built-in commands and shortcuts

678* **[Permissions](/en/permissions)**: control tool and skill access

slack.md +31 −6

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Claude Code in Slack5# Claude Code in Slack

2 6

3> Delegate coding tasks directly from your Slack workspace7> Delegate coding tasks directly from your Slack workspace

60 In Code + Chat mode, if Claude routes a message to Chat but you wanted a coding session, you can click "Retry as Code" to create a Claude Code session instead. Similarly, if it's routed to Code but you wanted a Chat session, you can choose that option in that thread.64 In Code + Chat mode, if Claude routes a message to Chat but you wanted a coding session, you can click "Retry as Code" to create a Claude Code session instead. Similarly, if it's routed to Code but you wanted a Chat session, you can choose that option in that thread.

61 </Note>65 </Note>

62 </Step>66 </Step>

68 <Step title="Add Claude to channels">

69 Claude is not automatically added to any channels after installation. To use Claude in a channel, invite it by typing `/invite @Claude` in that channel. Claude can only respond to @mentions in channels where it has been added.

70 </Step>

63</Steps>71</Steps>

64 72

65## How it works73## How it works

123| Repository Access | Users can only access repositories they've personally connected |131| Repository Access | Users can only access repositories they've personally connected |

124| Session History | Sessions appear in your Claude Code history on claude.ai/code |132| Session History | Sessions appear in your Claude Code history on claude.ai/code |

125 133

126### Workspace admin permissions134### Workspace-level access

135

136Slack workspace administrators control whether the Claude app is available in their workspace:

137

138| Control | Description |

139| :--------------------------- | :---------------------------------------------------------------------------------------------------------------- |

140| App installation | Workspace admins decide whether to install the Claude app from the Slack App Marketplace |

141| Enterprise Grid distribution | For Enterprise Grid organizations, organization admins can control which workspaces have access to the Claude app |

142| App removal | Removing the app from a workspace immediately revokes access for all users in that workspace |

143

144### Channel-based access control

127 145

128Slack workspace administrators control whether the Claude app can be installed in the workspace. Individual users then authenticate with their own Claude accounts to use the integration.146Claude is not automatically added to any channels after installation. Users must explicitly invite Claude to channels where they want to use it:

147

148* **Invite required**: Type `/invite @Claude` in any channel to add Claude to that channel

149* **Channel membership controls access**: Claude can only respond to @mentions in channels where it has been added

150* **Access gating through channels**: Admins can control who uses Claude Code by managing which channels Claude is invited to and who has access to those channels

151* **Private channel support**: Claude works in both public and private channels, giving teams flexibility in controlling visibility

152

153This channel-based model allows teams to restrict Claude Code usage to specific channels, providing an additional layer of access control beyond workspace-level permissions.

129 154

130## What's accessible where155## What's accessible where

131 156

133 158

134**On the web**: The complete Claude Code session with full conversation history, all code changes, file operations, and the ability to continue the session or create pull requests.159**On the web**: The complete Claude Code session with full conversation history, all code changes, file operations, and the ability to continue the session or create pull requests.

135 160

161For Enterprise and Teams accounts, sessions created from Claude in Slack are

162automatically visible to the organization. See [Claude Code on the Web sharing](/en/claude-code-on-the-web#sharing-sessions)

163for more details.

164

136## Best practices165## Best practices

137 166

138### Writing effective requests167### Writing effective requests

204 Get additional support233 Get additional support

205 </Card>234 </Card>

206</CardGroup>235</CardGroup>

~~207~~

~~208~~

~~209~~

210> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

slash-commands.md +0 −506 deleted

File DeletedView Diff

~~1# Slash commands~~

~~3> Control Claude's behavior during an interactive session with slash commands.~~

~~5## Built-in slash commands~~

~~7| Command | Purpose |~~

~~8| :------------------------ | :-------------------------------------------------------------------------------------------------------------------------- |~~

~~9| `/add-dir` | Add additional working directories |~~

~~10| `/agents` | Manage custom AI subagents for specialized tasks |~~

~~11| `/bashes` | List and manage background tasks |~~

~~12| `/bug` | Report bugs (sends conversation to Anthropic) |~~

~~13| `/clear` | Clear conversation history |~~

~~14| `/compact [instructions]` | Compact conversation with optional focus instructions |~~

~~15| `/config` | Open the Settings interface (Config tab) |~~

~~16| `/context` | Visualize current context usage as a colored grid |~~

~~17| `/cost` | Show token usage statistics. See [cost tracking guide](/en/costs#using-the-cost-command) for subscription-specific details. |~~

~~18| `/doctor` | Checks the health of your Claude Code installation |~~

~~19| `/exit` | Exit the REPL |~~

~~20| `/export [filename]` | Export the current conversation to a file or clipboard |~~

~~21| `/help` | Get usage help |~~

~~22| `/hooks` | Manage hook configurations for tool events |~~

~~23| `/ide` | Manage IDE integrations and show status |~~

~~24| `/init` | Initialize project with `CLAUDE.md` guide |~~

~~25| `/install-github-app` | Set up Claude GitHub Actions for a repository |~~

~~26| `/login` | Switch Anthropic accounts |~~

~~27| `/logout` | Sign out from your Anthropic account |~~

~~28| `/mcp` | Manage MCP server connections and OAuth authentication |~~

~~29| `/memory` | Edit `CLAUDE.md` memory files |~~

~~30| `/model` | Select or change the AI model |~~

~~31| `/output-style [style]` | Set the output style directly or from a selection menu |~~

~~32| `/permissions` | View or update [permissions](/en/iam#configuring-permissions) |~~

~~33| `/plugin` | Manage Claude Code plugins |~~

~~34| `/pr-comments` | View pull request comments |~~

~~35| `/privacy-settings` | View and update your privacy settings |~~

~~36| `/release-notes` | View release notes |~~

~~37| `/rename <name>` | Rename the current session for easier identification |~~

~~38| `/resume [session]` | Resume a conversation by ID or name, or open the session picker |~~

~~39| `/review` | Request code review |~~

~~40| `/rewind` | Rewind the conversation and/or code |~~

~~41| `/sandbox` | Enable sandboxed bash tool with filesystem and network isolation for safer, more autonomous execution |~~

~~42| `/security-review` | Complete a security review of pending changes on the current branch |~~

~~43| `/stats` | Visualize daily usage, session history, streaks, and model preferences |~~

~~44| `/status` | Open the Settings interface (Status tab) showing version, model, account, and connectivity |~~

~~45| `/statusline` | Set up Claude Code's status line UI |~~

~~46| `/terminal-setup` | Install Shift+Enter key binding for newlines (iTerm2 and VSCode only) |~~

~~47| `/todos` | List current TODO items |~~

~~48| `/usage` | For subscription plans only: show plan usage limits and rate limit status |~~

~~49| `/vim` | Enter vim mode for alternating insert and command modes |~~

~~51## Custom slash commands~~

53Custom slash commands allow you to define frequently used prompts as Markdown files that Claude Code can execute. Commands are organized by scope (project-specific or personal) and support namespacing through directory structures.

~~55### Syntax~~

~~57```~~

~~58/<command-name> [arguments]~~

~~59```~~

~~61#### Parameters~~

~~63| Parameter | Description |~~

~~64| :--------------- | :---------------------------------------------------------------- |~~

~~65| `<command-name>` | Name derived from the Markdown filename (without `.md` extension) |~~

~~66| `[arguments]` | Optional arguments passed to the command |~~

~~68### Command types~~

~~70#### Project commands~~

~~72Commands stored in your repository and shared with your team. When listed in `/help`, these commands show "(project)" after their description.~~

~~74**Location**: `.claude/commands/`~~

~~76The following example creates the `/optimize` command:~~

~~78```bash theme={null}~~

~~79# Create a project command~~

~~80mkdir -p .claude/commands~~

~~81echo "Analyze this code for performance issues and suggest optimizations:" > .claude/commands/optimize.md~~

~~82```~~

~~84#### Personal commands~~

~~86Commands available across all your projects. When listed in `/help`, these commands show "(user)" after their description.~~

~~88**Location**: `~/.claude/commands/`~~

~~90The following example creates the `/security-review` command:~~

~~92```bash theme={null}~~

~~93# Create a personal command~~

~~94mkdir -p ~/.claude/commands~~

~~95echo "Review this code for security vulnerabilities:" > ~/.claude/commands/security-review.md~~

~~96```~~

~~98### Features~~

100#### Namespacing

~~101~~

102Use subdirectories to group related commands. Subdirectories appear in the command description but don't affect the command name.

~~103~~

104For example:

~~105~~

106* `.claude/commands/frontend/component.md` creates `/component` with description "(project:frontend)"

107* `~/.claude/commands/component.md` creates `/component` with description "(user)"

~~108~~

109If a project command and user command share the same name, the project command takes precedence and the user command is silently ignored. For example, if both `.claude/commands/deploy.md` and `~/.claude/commands/deploy.md` exist, `/deploy` runs the project version.

~~110~~

111Commands in different subdirectories can share names since the subdirectory appears in the description to distinguish them. For example, `.claude/commands/frontend/test.md` and `.claude/commands/backend/test.md` both create `/test`, but show as "(project:frontend)" and "(project:backend)" respectively.

~~112~~

113#### Arguments

~~114~~

115Pass dynamic values to commands using argument placeholders:

~~116~~

117##### All arguments with `$ARGUMENTS`

~~118~~

119The `$ARGUMENTS` placeholder captures all arguments passed to the command:

~~120~~

121```bash theme={null}

122# Command definition

123echo 'Fix issue #$ARGUMENTS following our coding standards' > .claude/commands/fix-issue.md

~~124~~

125# Usage

126> /fix-issue 123 high-priority

127# $ARGUMENTS becomes: "123 high-priority"

128```

~~129~~

130##### Individual arguments with `$1`, `$2`, etc.

~~131~~

132Access specific arguments individually using positional parameters (similar to shell scripts):

~~133~~

134```bash theme={null}

135# Command definition

136echo 'Review PR #$1 with priority $2 and assign to $3' > .claude/commands/review-pr.md

~~137~~

138# Usage

139> /review-pr 456 high alice

140# $1 becomes "456", $2 becomes "high", $3 becomes "alice"

141```

~~142~~

143Use positional arguments when you need to:

~~144~~

145* Access arguments individually in different parts of your command

146* Provide defaults for missing arguments

147* Build more structured commands with specific parameter roles

~~148~~

149#### Bash command execution

~~150~~

151Execute bash commands before the slash command runs using the `!` prefix. The output is included in the command context. You *must* include `allowed-tools` with the `Bash` tool, but you can choose the specific bash commands to allow.

~~152~~

153For example:

~~154~~

155```markdown theme={null}

156allowed-tools: Bash(git add:*), Bash(git status:*), Bash(git commit:*)

157description: Create a git commit

~~158~~

159## Context

~~160~~

161- Current git status: !`git status`

162- Current git diff (staged and unstaged changes): !`git diff HEAD`

163- Current branch: !`git branch --show-current`

164- Recent commits: !`git log --oneline -10`

~~165~~

166## Your task

~~167~~

168Based on the above changes, create a single git commit.

169```

~~170~~

171#### File references

~~172~~

173Include file contents in commands using the `@` prefix to [reference files](/en/common-workflows#reference-files-and-directories).

~~174~~

175For example:

~~176~~

177```markdown theme={null}

178# Reference a specific file

~~179~~

180Review the implementation in @src/utils/helpers.js

~~181~~

182# Reference multiple files

~~183~~

184Compare @src/old-version.js with @src/new-version.js

185```

~~186~~

187#### Thinking mode

~~188~~

189Slash commands can trigger extended thinking by including [extended thinking keywords](/en/common-workflows#use-extended-thinking).

~~190~~

191### Frontmatter

~~192~~

193Command files support frontmatter, useful for specifying metadata about the command:

~~194~~

195| Frontmatter | Purpose | Default |

196| :------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | :---------------------------------- |

197| `allowed-tools` | List of tools the command can use | Inherits from the conversation |

199| `description` | Brief description of the command | Uses the first line from the prompt |

200| `model` | Specific model string (see [Models overview](https://docs.claude.com/en/docs/about-claude/models/overview)) | Inherits from the conversation |

201| `disable-model-invocation` | Whether to prevent `SlashCommand` tool from calling this command | false |

~~202~~

203For example:

~~204~~

205```markdown theme={null}

206allowed-tools: Bash(git add:*), Bash(git status:*), Bash(git commit:*)

207argument-hint: [message]

208description: Create a git commit

209model: claude-3-5-haiku-20241022

~~210~~

211Create a git commit with message: $ARGUMENTS

212```

~~213~~

214Example using positional arguments:

~~215~~

216```markdown theme={null}

217argument-hint: [pr-number] [priority] [assignee]

218description: Review pull request

~~219~~

220Review PR #$1 with priority $2 and assign to $3.

221Focus on security, performance, and code style.

222```

~~223~~

224## Plugin commands

~~225~~

226[Plugins](/en/plugins) can provide custom slash commands that integrate seamlessly with Claude Code. Plugin commands work exactly like user-defined commands but are distributed through [plugin marketplaces](/en/plugin-marketplaces).

~~227~~

228### How plugin commands work

~~229~~

230Plugin commands are:

~~231~~

232* **Namespaced**: Commands can use the format `/plugin-name:command-name` to avoid conflicts (plugin prefix is optional unless there are name collisions)

233* **Automatically available**: Once a plugin is installed and enabled, its commands appear in `/help`

234* **Fully integrated**: Support all command features (arguments, frontmatter, bash execution, file references)

~~235~~

236### Plugin command structure

~~237~~

238**Location**: `commands/` directory in plugin root

~~239~~

240**File format**: Markdown files with frontmatter

~~241~~

242**Basic command structure**:

~~243~~

244```markdown theme={null}

245description: Brief description of what the command does

~~246~~

247# Command Name

~~248~~

249Detailed instructions for Claude on how to execute this command.

250Include specific guidance on parameters, expected outcomes, and any special considerations.

251```

~~252~~

253**Advanced command features**:

~~254~~

255* **Arguments**: Use placeholders like `{arg1}` in command descriptions

256* **Subdirectories**: Organize commands in subdirectories for namespacing

257* **Bash integration**: Commands can execute shell scripts and programs

258* **File references**: Commands can reference and modify project files

~~259~~

260### Invocation patterns

~~261~~

262```shell Direct command (when no conflicts) theme={null}

263/command-name

264```

~~265~~

266```shell Plugin-prefixed (when needed for disambiguation) theme={null}

267/plugin-name:command-name

268```

~~269~~

270```shell With arguments (if command supports them) theme={null}

271/command-name arg1 arg2

272```

~~273~~

274## MCP slash commands

~~275~~

276MCP servers can expose prompts as slash commands that become available in Claude Code. These commands are dynamically discovered from connected MCP servers.

~~277~~

278### Command format

~~279~~

280MCP commands follow the pattern:

~~281~~

282```

283/mcp__<server-name>__<prompt-name> [arguments]

284```

~~285~~

286### Features

~~287~~

288#### Dynamic discovery

~~289~~

290MCP commands are automatically available when:

~~291~~

292* An MCP server is connected and active

293* The server exposes prompts through the MCP protocol

294* The prompts are successfully retrieved during connection

~~295~~

296#### Arguments

~~297~~

298MCP prompts can accept arguments defined by the server:

~~299~~

300```

301# Without arguments

302> /mcp__github__list_prs

~~303~~

304# With arguments

305> /mcp__github__pr_review 456

306> /mcp__jira__create_issue "Bug title" high

307```

~~308~~

309#### Naming conventions

~~310~~

311Server and prompt names are normalized:

~~312~~

313* Spaces and special characters become underscores

314* Names are lowercase for consistency

~~315~~

316### Managing MCP connections

~~317~~

318Use the `/mcp` command to:

~~319~~

320* View all configured MCP servers

321* Check connection status

322* Authenticate with OAuth-enabled servers

323* Clear authentication tokens

324* View available tools and prompts from each server

~~325~~

326### MCP permissions and wildcards

~~327~~

328To approve all tools from an MCP server, use either the server name alone or wildcard syntax:

~~329~~

330* `mcp__github` (approves all GitHub tools)

331* `mcp__github__*` (wildcard syntax, also approves all GitHub tools)

~~332~~

333To approve specific tools, list each one explicitly:

~~334~~

335* `mcp__github__get_issue`

336* `mcp__github__list_issues`

~~337~~

338See [MCP permission rules](/en/iam#tool-specific-permission-rules) for more details.

~~339~~

340## `SlashCommand` tool

~~341~~

342The `SlashCommand` tool allows Claude to execute [custom slash commands](/en/slash-commands#custom-slash-commands) programmatically

343during a conversation. This gives Claude the ability to invoke custom commands

344on your behalf when appropriate.

~~345~~

346To encourage Claude to use the `SlashCommand` tool, reference the command by name, including the slash, in your prompts or `CLAUDE.md` file. For example:

~~347~~

348```

349> Run /write-unit-test when you are about to start writing tests.

350```

~~351~~

352This tool puts each available custom slash command's metadata into context up to the character budget limit. You can use `/context` to monitor token usage and follow the operations below to manage context.

~~353~~

354### `SlashCommand` tool supported commands

~~355~~

356`SlashCommand` tool only supports custom slash commands that:

~~357~~

358* Are user-defined. Built-in commands like `/compact` and `/init` are *not* supported.

359* Have the `description` frontmatter field populated. The description is used in the context.

~~360~~

361For Claude Code versions >= 1.0.124, you can see which custom slash commands

362`SlashCommand` tool can invoke by running `claude --debug` and triggering a query.

~~363~~

364### Disable `SlashCommand` tool

~~365~~

366To prevent Claude from executing any slash commands via the tool:

~~367~~

368```bash theme={null}

369/permissions

370# Add to deny rules: SlashCommand

371```

~~372~~

373This also removes the SlashCommand tool and command descriptions from context.

~~374~~

375### Disable specific commands only

~~376~~

377To prevent a specific slash command from becoming available, add

378`disable-model-invocation: true` to the slash command's frontmatter.

~~379~~

380This also removes the command's metadata from context.

~~381~~

382### `SlashCommand` permission rules

~~383~~

384The permission rules support:

~~385~~

386* **Exact match**: `SlashCommand:/commit` (allows only `/commit` with no arguments)

387* **Prefix match**: `SlashCommand:/review-pr:*` (allows `/review-pr` with any arguments)

~~388~~

389### Character budget limit

~~390~~

391The `SlashCommand` tool includes a character budget to limit the size of command

392descriptions shown to Claude. This prevents token overflow when many commands

393are available.

~~394~~

395The budget includes each custom slash command's name, arguments, and description.

~~396~~

397* **Default limit**: 15,000 characters

398* **Custom limit**: Set via `SLASH_COMMAND_TOOL_CHAR_BUDGET` environment variable

~~399~~

400When the character budget is exceeded, Claude sees only a subset of the available commands. In `/context`, a warning shows "M of N commands".

~~401~~

402## Skills vs slash commands

~~403~~

404**Slash commands** and **Agent Skills** serve different purposes in Claude Code:

~~405~~

406### Use slash commands for

~~407~~

408**Quick, frequently used prompts**:

~~409~~

410* Simple prompt snippets you use often

411* Quick reminders or templates

412* Frequently used instructions that fit in one file

~~413~~

414**Examples**:

~~415~~

416* `/review` → "Review this code for bugs and suggest improvements"

417* `/explain` → "Explain this code in simple terms"

418* `/optimize` → "Analyze this code for performance issues"

~~419~~

420### Use Skills for

~~421~~

422**Comprehensive capabilities with structure**:

~~423~~

424* Complex workflows with multiple steps

425* Capabilities requiring scripts or utilities

426* Knowledge organized across multiple files

427* Team workflows you want to standardize

~~428~~

429**Examples**:

~~430~~

431* PDF processing Skill with form-filling scripts and validation

432* Data analysis Skill with reference docs for different data types

433* Documentation Skill with style guides and templates

~~434~~

435### Key differences

~~436~~

437| Aspect | Slash Commands | Agent Skills |

438| -------------- | -------------------------------- | ----------------------------------- |

439| **Complexity** | Simple prompts | Complex capabilities |

440| **Structure** | Single .md file | Directory with SKILL.md + resources |

441| **Discovery** | Explicit invocation (`/command`) | Automatic (based on context) |

442| **Files** | One file only | Multiple files, scripts, templates |

443| **Scope** | Project or personal | Project or personal |

444| **Sharing** | Via git | Via git |

~~445~~

446### Example comparison

~~447~~

448**As a slash command**:

~~449~~

450```markdown theme={null}

451# .claude/commands/review.md

452Review this code for:

453- Security vulnerabilities

454- Performance issues

455- Code style violations

456```

~~457~~

458Usage: `/review` (manual invocation)

~~459~~

460**As a Skill**:

~~461~~

462```

463.claude/skills/code-review/

464├── SKILL.md (overview and workflows)

465├── SECURITY.md (security checklist)

466├── PERFORMANCE.md (performance patterns)

467├── STYLE.md (style guide reference)

468└── scripts/

469 └── run-linters.sh

470```

~~471~~

472Usage: "Can you review this code?" (automatic discovery)

~~473~~

474The Skill provides richer context, validation scripts, and organized reference material.

~~475~~

476### When to use each

~~477~~

478**Use slash commands**:

~~479~~

480* You invoke the same prompt repeatedly

481* The prompt fits in a single file

482* You want explicit control over when it runs

~~483~~

484**Use Skills**:

~~485~~

486* Claude should discover the capability automatically

487* Multiple files or scripts are needed

488* Complex workflows with validation steps

489* Team needs standardized, detailed guidance

~~490~~

491Both slash commands and Skills can coexist. Use the approach that fits your needs.

~~492~~

493Learn more about [Agent Skills](/en/skills).

~~494~~

495## See also

~~496~~

497* [Plugins](/en/plugins) - Extend Claude Code with custom commands through plugins

498* [Identity and Access Management](/en/iam) - Complete guide to permissions, including MCP tool permissions

499* [Interactive mode](/en/interactive-mode) - Shortcuts, input modes, and interactive features

500* [CLI reference](/en/cli-reference) - Command-line flags and options

501* [Settings](/en/settings) - Configuration options

502* [Memory management](/en/memory) - Managing Claude's memory across sessions

~~503~~

~~504~~

~~505~~

506> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

statusline.md +752 −153

Details

~~1# Status line configuration~~1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

2 4

~~3> Create a custom status line for Claude Code to display contextual information~~5# Customize your status line

4 6

~~5Make Claude Code your own with a custom status line that displays at the bottom of the Claude Code interface, similar to how terminal prompts (PS1) work in shells like Oh-my-zsh.~~7> Configure a custom status bar to monitor context window usage, costs, and git status in Claude Code

6 8

~~7## Create a custom status line~~9The status line is a customizable bar at the bottom of Claude Code that runs any shell script you configure. It receives JSON session data on stdin and displays whatever your script prints, giving you a persistent, at-a-glance view of context usage, costs, git status, or anything else you want to track.

8 10

~~9You can either:~~11Status lines are useful when you:

10 12

11* Run `/statusline` to ask Claude Code to help you set up a custom status line. By default, it will try to reproduce your terminal's prompt, but you can provide additional instructions about the behavior you want to Claude Code, such as `/statusline show the model name in orange`13* Want to monitor context window usage as you work

14* Need to track session costs

15* Work across multiple sessions and need to distinguish them

16* Want git branch and status always visible

12 17

~~13* Directly add a `statusLine` command to your `.claude/settings.json`:~~18Here's an example of a [multi-line status line](#display-multiple-lines) that displays git info on the first line and a color-coded context bar on the second.

20<Frame>

21 <img src="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=60f11387658acc9ff75158ae85f2ac87" alt="A multi-line status line showing model name, directory, git branch on the first line, and a context usage progress bar with cost and duration on the second line" data-og-width="776" width="776" data-og-height="212" height="212" data-path="images/statusline-multiline.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=280&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=2e448b44c332620e6c9c2be4ded992e5 280w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=560&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=f796af2db9c68ab2ddbc5136840b9551 560w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=840&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=d29c13d6164773198a0b2c47b31f6c09 840w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=1100&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=d7720e5f51310185c0c02152f6c10d8b 1100w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=1650&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=b4e008cde27990a8d5783e41e5b93246 1650w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=2500&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=40ab24813303dc2e4c09f2675f3faf6e 2500w" />

22</Frame>

24This page walks through [setting up a basic status line](#set-up-a-status-line), explains [how the data flows](#how-status-lines-work) from Claude Code to your script, lists [all the fields you can display](#available-data), and provides [ready-to-use examples](#examples) for common patterns like git status, cost tracking, and progress bars.

26## Set up a status line

28Use the [`/statusline` command](#use-the-statusline-command) to have Claude Code generate a script for you, or [manually create a script](#manually-configure-a-status-line) and add it to your settings.

30### Use the /statusline command

32The `/statusline` command accepts natural language instructions describing what you want displayed. Claude Code generates a script file in `~/.claude/` and updates your settings automatically:

34```

35/statusline show model name and context percentage with a progress bar

36```

38### Manually configure a status line

40Add a `statusLine` field to your user settings (`~/.claude/settings.json`, where `~` is your home directory) or [project settings](/en/settings#settings-files). Set `type` to `"command"` and point `command` to a script path or an inline shell command. For a full walkthrough of creating a script, see [Build a status line step by step](#build-a-status-line-step-by-step).

14 41

15```json theme={null}42```json theme={null}

16{43{

17 "statusLine": {44 "statusLine": {

18 "type": "command",45 "type": "command",

19 "command": "~/.claude/statusline.sh",46 "command": "~/.claude/statusline.sh",

~~20 "padding": 0 // Optional: set to 0 to let status line go to edge~~47 "padding": 2

48 }

49}

50```

52The `command` field runs in a shell, so you can also use inline commands instead of a script file. This example uses `jq` to parse the JSON input and display the model name and context percentage:

54```json theme={null}

55{

56 "statusLine": {

57 "type": "command",

58 "command": "jq -r '\"[\\(.model.display_name)] \\(.context_window.used_percentage // 0)% context\"'"

21 }59 }

22}60}

23```61```

24 62

~~25## How it Works~~63The optional `padding` field adds extra horizontal spacing (in characters) to the status line content. Defaults to `0`. This padding is in addition to the interface's built-in spacing, so it controls relative indentation rather than absolute distance from the terminal edge.

26 64

~~27* The status line is updated when the conversation messages update~~65### Disable the status line

~~28* Updates run at most every 300 ms~~

~~29* The first line of stdout from your command becomes the status line text~~

~~30* ANSI color codes are supported for styling your status line~~

~~31* Claude Code passes contextual information about the current session (model, directories, etc.) as JSON to your script via stdin~~

32 66

~~33## JSON Input Structure~~67Run `/statusline` and ask it to remove or clear your status line (e.g., `/statusline delete`, `/statusline clear`, `/statusline remove it`). You can also manually delete the `statusLine` field from your settings.json.

34 68

~~35Your status line command receives structured data via stdin in JSON format:~~69## Build a status line step by step

36 70

~~37```json theme={null}~~71This walkthrough shows what's happening under the hood by manually creating a status line that displays the current model, working directory, and context window usage percentage.

~~38{~~72

~~39 "hook_event_name": "Status",~~73<Note>Running [`/statusline`](#use-the-statusline-command) with a description of what you want configures all of this for you automatically.</Note>

~~40 "session_id": "abc123...",~~74

~~41 "transcript_path": "/path/to/transcript.json",~~75These examples use Bash scripts, which work on macOS and Linux. On Windows, you can run Bash scripts through [WSL (Windows Subsystem for Linux)](https://learn.microsoft.com/en-us/windows/wsl/install) or rewrite them in PowerShell.

77<Frame>

78 <img src="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-quickstart.png?fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=696445e59ca0059213250651ad23db6b" alt="A status line showing model name, directory, and context percentage" data-og-width="726" width="726" data-og-height="164" height="164" data-path="images/statusline-quickstart.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-quickstart.png?w=280&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=728c4bd06c8559cb46ddffffad983373 280w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-quickstart.png?w=560&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=f9d28e0f8f48f695167dd1d632a6cf4f 560w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-quickstart.png?w=840&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=57a2803a18cafe8cf1aa05619444f20c 840w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-quickstart.png?w=1100&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=52cdd52865842f0cda24489dd5310d3b 1100w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-quickstart.png?w=1650&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=f8876ea1f72bf40bd0aeec483ee20164 1650w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-quickstart.png?w=2500&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=6b1524305c7c71122cde65d0c3822374 2500w" />

79</Frame>

81<Steps>

82 <Step title="Create a script that reads JSON and prints output">

83 Claude Code sends JSON data to your script via stdin. This script uses [`jq`](https://jqlang.github.io/jq/), a command-line JSON parser you may need to install, to extract the model name, directory, and context percentage, then prints a formatted line.

85 Save this to `~/.claude/statusline.sh` (where `~` is your home directory, such as `/Users/username` on macOS or `/home/username` on Linux):

87 ```bash theme={null}

88 #!/bin/bash

89 # Read JSON data that Claude Code sends to stdin

90 input=$(cat)

92 # Extract fields using jq

93 MODEL=$(echo "$input" | jq -r '.model.display_name')

94 DIR=$(echo "$input" | jq -r '.workspace.current_dir')

95 # The "// 0" provides a fallback if the field is null

96 PCT=$(echo "$input" | jq -r '.context_window.used_percentage // 0' | cut -d. -f1)

98 # Output the status line - ${DIR##*/} extracts just the folder name

99 echo "[$MODEL] 📁 ${DIR##*/} | ${PCT}% context"

100 ```

101 </Step>

102

103 <Step title="Make it executable">

104 Mark the script as executable so your shell can run it:

105

106 ```bash theme={null}

107 chmod +x ~/.claude/statusline.sh

108 ```

109 </Step>

110

111 <Step title="Add to settings">

112 Tell Claude Code to run your script as the status line. Add this configuration to `~/.claude/settings.json`, which sets `type` to `"command"` (meaning "run this shell command") and points `command` to your script:

113

114 ```json theme={null}

115 {

116 "statusLine": {

117 "type": "command",

118 "command": "~/.claude/statusline.sh"

119 }

120 }

121 ```

122

123 Your status line appears at the bottom of the interface. Settings reload automatically, but changes won't appear until your next interaction with Claude Code.

124 </Step>

125</Steps>

126

127## How status lines work

128

129Claude Code runs your script and pipes [JSON session data](#available-data) to it via stdin. Your script reads the JSON, extracts what it needs, and prints text to stdout. Claude Code displays whatever your script prints.

130

131**When it updates**

132

133Your script runs after each new assistant message, when the permission mode changes, or when vim mode toggles. Updates are debounced at 300ms, meaning rapid changes batch together and your script runs once things settle. If a new update triggers while your script is still running, the in-flight execution is cancelled. If you edit your script, the changes won't appear until your next interaction with Claude Code triggers an update.

134

135**What your script can output**

136

137* **Multiple lines**: each `echo` or `print` statement displays as a separate row. See the [multi-line example](#display-multiple-lines).

138* **Colors**: use [ANSI escape codes](https://en.wikipedia.org/wiki/ANSI_escape_code#Colors) like `\033[32m` for green (terminal must support them). See the [git status example](#git-status-with-colors).

139* **Links**: use [OSC 8 escape sequences](https://en.wikipedia.org/wiki/ANSI_escape_code#OSC) to make text clickable (Cmd+click on macOS, Ctrl+click on Windows/Linux). Requires a terminal that supports hyperlinks like iTerm2, Kitty, or WezTerm. See the [clickable links example](#clickable-links).

140

141<Note>The status line runs locally and does not consume API tokens. It temporarily hides during certain UI interactions, including autocomplete suggestions, the help menu, and permission prompts.</Note>

142

143## Available data

144

145Claude Code sends the following JSON fields to your script via stdin:

146

147| Field | Description |

148| ------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

149| `model.id`, `model.display_name` | Current model identifier and display name |

150| `cwd`, `workspace.current_dir` | Current working directory. Both fields contain the same value; `workspace.current_dir` is preferred for consistency with `workspace.project_dir`. |

151| `workspace.project_dir` | Directory where Claude Code was launched, which may differ from `cwd` if the working directory changes during a session |

152| `cost.total_cost_usd` | Total session cost in USD |

153| `cost.total_duration_ms` | Total wall-clock time since the session started, in milliseconds |

154| `cost.total_api_duration_ms` | Total time spent waiting for API responses in milliseconds |

155| `cost.total_lines_added`, `cost.total_lines_removed` | Lines of code changed |

156| `context_window.total_input_tokens`, `context_window.total_output_tokens` | Cumulative token counts across the session |

157| `context_window.context_window_size` | Maximum context window size in tokens. 200000 by default, or 1000000 for models with extended context. |

158| `context_window.used_percentage` | Pre-calculated percentage of context window used |

159| `context_window.remaining_percentage` | Pre-calculated percentage of context window remaining |

160| `context_window.current_usage` | Token counts from the last API call, described in [context window fields](#context-window-fields) |

161| `exceeds_200k_tokens` | Whether the total token count (input, cache, and output tokens combined) from the most recent API response exceeds 200k. This is a fixed threshold regardless of actual context window size. |

162| `session_id` | Unique session identifier |

163| `transcript_path` | Path to conversation transcript file |

164| `version` | Claude Code version |

165| `output_style.name` | Name of the current output style |

166| `vim.mode` | Current vim mode (`NORMAL` or `INSERT`) when [vim mode](/en/interactive-mode#vim-editor-mode) is enabled |

167| `agent.name` | Agent name when running with the `--agent` flag or agent settings configured |

168

169<Accordion title="Full JSON schema">

170 Your status line command receives this JSON structure via stdin:

171

172 ```json theme={null}

173 {

42 "cwd": "/current/working/directory",174 "cwd": "/current/working/directory",

175 "session_id": "abc123...",

176 "transcript_path": "/path/to/transcript.jsonl",

43 "model": {177 "model": {

~~44 "id": "claude-opus-4-1",~~178 "id": "claude-opus-4-6",

45 "display_name": "Opus"179 "display_name": "Opus"

46 },180 },

47 "workspace": {181 "workspace": {

63 "total_input_tokens": 15234,197 "total_input_tokens": 15234,

64 "total_output_tokens": 4521,198 "total_output_tokens": 4521,

65 "context_window_size": 200000,199 "context_window_size": 200000,

200 "used_percentage": 8,

201 "remaining_percentage": 92,

66 "current_usage": {202 "current_usage": {

67 "input_tokens": 8500,203 "input_tokens": 8500,

68 "output_tokens": 1200,204 "output_tokens": 1200,

69 "cache_creation_input_tokens": 5000,205 "cache_creation_input_tokens": 5000,

70 "cache_read_input_tokens": 2000206 "cache_read_input_tokens": 2000

71 }207 }

208 },

209 "exceeds_200k_tokens": false,

210 "vim": {

211 "mode": "NORMAL"

212 },

213 "agent": {

214 "name": "security-reviewer"

72 }215 }

~~73}~~216 }

~~74```~~217 ```

75 218

~~76## Example Scripts~~219 **Fields that may be absent** (not present in JSON):

77 220

~~78### Simple Status Line~~221 * `vim`: appears only when vim mode is enabled

222 * `agent`: appears only when running with the `--agent` flag or agent settings configured

79 223

~~80```bash theme={null}~~224 **Fields that may be `null`**:

~~81#!/bin/bash~~

~~82# Read JSON input from stdin~~

~~83input=$(cat)~~

84 225

~~85# Extract values using jq~~226 * `context_window.current_usage`: `null` before the first API call in a session

~~86MODEL_DISPLAY=$(echo "$input" | jq -r '.model.display_name')~~227 * `context_window.used_percentage`, `context_window.remaining_percentage`: may be `null` early in the session

~~87CURRENT_DIR=$(echo "$input" | jq -r '.workspace.current_dir')~~

88 228

~~89echo "[$MODEL_DISPLAY] 📁 ${CURRENT_DIR##*/}"~~229 Handle missing fields with conditional access and null values with fallback defaults in your scripts.

~~90```~~230</Accordion>

231

232### Context window fields

233

234The `context_window` object provides two ways to track context usage:

235

236* **Cumulative totals** (`total_input_tokens`, `total_output_tokens`): sum of all tokens across the entire session, useful for tracking total consumption

237* **Current usage** (`current_usage`): token counts from the most recent API call, use this for accurate context percentage since it reflects the actual context state

238

239The `current_usage` object contains:

240

241* `input_tokens`: input tokens in current context

242* `output_tokens`: output tokens generated

243* `cache_creation_input_tokens`: tokens written to cache

244* `cache_read_input_tokens`: tokens read from cache

245

246The `used_percentage` field is calculated from input tokens only: `input_tokens + cache_creation_input_tokens + cache_read_input_tokens`. It does not include `output_tokens`.

247

248If you calculate context percentage manually from `current_usage`, use the same input-only formula to match `used_percentage`.

249

250The `current_usage` object is `null` before the first API call in a session.

251

252## Examples

253

254These examples show common status line patterns. To use any example:

255

2561. Save the script to a file like `~/.claude/statusline.sh` (or `.py`/`.js`)

2572. Make it executable: `chmod +x ~/.claude/statusline.sh`

2583. Add the path to your [settings](#manually-configure-a-status-line)

259

260The Bash examples use [`jq`](https://jqlang.github.io/jq/) to parse JSON. Python and Node.js have built-in JSON parsing.

261

262### Context window usage

263

264Display the current model and context window usage with a visual progress bar. Each script reads JSON from stdin, extracts the `used_percentage` field, and builds a 10-character bar where filled blocks (▓) represent usage:

265

266<Frame>

267 <img src="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-context-window-usage.png?fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=15b58ab3602f036939145dde3165c6f7" alt="A status line showing model name and a progress bar with percentage" data-og-width="448" width="448" data-og-height="152" height="152" data-path="images/statusline-context-window-usage.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-context-window-usage.png?w=280&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=a18fecd31f06b16e984b1ab3310acbc0 280w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-context-window-usage.png?w=560&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=2f4b3caff156efede2ded995dbaf167f 560w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-context-window-usage.png?w=840&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=8f6b8c7e7d3a999c570e96ad2ea13d5a 840w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-context-window-usage.png?w=1100&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=d9334e6a08e6f11a253733c8592774a9 1100w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-context-window-usage.png?w=1650&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=e79490da8f62952e4d92837c408e63dc 1650w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-context-window-usage.png?w=2500&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=6f7c9ef8e629a794969c54b24163f92d 2500w" />

268</Frame>

269

270<CodeGroup>

271 ```bash Bash theme={null}

272 #!/bin/bash

273 # Read all of stdin into a variable

274 input=$(cat)

275

276 # Extract fields with jq, "// 0" provides fallback for null

277 MODEL=$(echo "$input" | jq -r '.model.display_name')

278 PCT=$(echo "$input" | jq -r '.context_window.used_percentage // 0' | cut -d. -f1)

279

280 # Build progress bar: printf creates spaces, tr replaces with blocks

281 BAR_WIDTH=10

282 FILLED=$((PCT * BAR_WIDTH / 100))

283 EMPTY=$((BAR_WIDTH - FILLED))

284 BAR=""

285 [ "$FILLED" -gt 0 ] && BAR=$(printf "%${FILLED}s" | tr ' ' '▓')

286 [ "$EMPTY" -gt 0 ] && BAR="${BAR}$(printf "%${EMPTY}s" | tr ' ' '░')"

287

288 echo "[$MODEL] $BAR $PCT%"

289 ```

91 290

~~92### Git-Aware Status Line~~291 ```python Python theme={null}

292 #!/usr/bin/env python3

293 import json, sys

93 294

~~94```bash theme={null}~~295 # json.load reads and parses stdin in one step

~~95#!/bin/bash~~296 data = json.load(sys.stdin)

~~96# Read JSON input from stdin~~297 model = data['model']['display_name']

~~97input=$(cat)~~298 # "or 0" handles null values

299 pct = int(data.get('context_window', {}).get('used_percentage', 0) or 0)

98 300

~~99# Extract values using jq~~301 # String multiplication builds the bar

100MODEL_DISPLAY=$(echo "$input" | jq -r '.model.display_name')302 filled = pct * 10 // 100

101CURRENT_DIR=$(echo "$input" | jq -r '.workspace.current_dir')303 bar = '▓' * filled + '░' * (10 - filled)

102 304

103# Show git branch if in a git repo305 print(f"[{model}] {bar} {pct}%")

104GIT_BRANCH=""306 ```

105if git rev-parse --git-dir > /dev/null 2>&1; then307

308 ```javascript Node.js theme={null}

309 #!/usr/bin/env node

310 // Node.js reads stdin asynchronously with events

311 let input = '';

312 process.stdin.on('data', chunk => input += chunk);

313 process.stdin.on('end', () => {

314 const data = JSON.parse(input);

315 const model = data.model.display_name;

316 // Optional chaining (?.) safely handles null fields

317 const pct = Math.floor(data.context_window?.used_percentage || 0);

318

319 // String.repeat() builds the bar

320 const filled = Math.floor(pct * 10 / 100);

321 const bar = '▓'.repeat(filled) + '░'.repeat(10 - filled);

322

323 console.log(`[${model}] ${bar} ${pct}%`);

324 });

325 ```

326</CodeGroup>

327

328### Git status with colors

329

330Show git branch with color-coded indicators for staged and modified files. This script uses [ANSI escape codes](https://en.wikipedia.org/wiki/ANSI_escape_code#Colors) for terminal colors: `\033[32m` is green, `\033[33m` is yellow, and `\033[0m` resets to default.

331

332<Frame>

333 <img src="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-git-context.png?fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=e656f34f90d1d9a1d0e220988914345f" alt="A status line showing model, directory, git branch, and colored indicators for staged and modified files" data-og-width="742" width="742" data-og-height="178" height="178" data-path="images/statusline-git-context.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-git-context.png?w=280&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=c1bced5f46afdc9aae549702591f8457 280w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-git-context.png?w=560&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=debe46a7a888234ec692751243bba492 560w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-git-context.png?w=840&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=3a069d5c8b0395908e42f0e295fd4854 840w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-git-context.png?w=1100&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=26aff0978865756d5ea299a22e5e9afd 1100w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-git-context.png?w=1650&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=d5ac1d59881e6f2032af053557dc4590 1650w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-git-context.png?w=2500&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=46febbf34b0ee646502d095433132709 2500w" />

334</Frame>

335

336Each script checks if the current directory is a git repository, counts staged and modified files, and displays color-coded indicators:

337

338<CodeGroup>

339 ```bash Bash theme={null}

340 #!/bin/bash

341 input=$(cat)

342

343 MODEL=$(echo "$input" | jq -r '.model.display_name')

344 DIR=$(echo "$input" | jq -r '.workspace.current_dir')

345

346 GREEN='\033[32m'

347 YELLOW='\033[33m'

348 RESET='\033[0m'

349

350 if git rev-parse --git-dir > /dev/null 2>&1; then

106 BRANCH=$(git branch --show-current 2>/dev/null)351 BRANCH=$(git branch --show-current 2>/dev/null)

107 if [ -n "$BRANCH" ]; then352 STAGED=$(git diff --cached --numstat 2>/dev/null | wc -l | tr -d ' ')

108 GIT_BRANCH=" | 🌿 $BRANCH"353 MODIFIED=$(git diff --numstat 2>/dev/null | wc -l | tr -d ' ')

109 fi

110fi

111 354

112echo "[$MODEL_DISPLAY] 📁 ${CURRENT_DIR##*/}$GIT_BRANCH"355 GIT_STATUS=""

113```356 [ "$STAGED" -gt 0 ] && GIT_STATUS="${GREEN}+${STAGED}${RESET}"

357 [ "$MODIFIED" -gt 0 ] && GIT_STATUS="${GIT_STATUS}${YELLOW}~${MODIFIED}${RESET}"

114 358

115### Python Example359 echo -e "[$MODEL] 📁 ${DIR##*/} | 🌿 $BRANCH $GIT_STATUS"

360 else

361 echo "[$MODEL] 📁 ${DIR##*/}"

362 fi

363 ```

116 364

117```python theme={null}365 ```python Python theme={null}

118#!/usr/bin/env python3366 #!/usr/bin/env python3

119import json367 import json, sys, subprocess, os

120import sys

121import os

122 368

123# Read JSON from stdin369 data = json.load(sys.stdin)

124data = json.load(sys.stdin)370 model = data['model']['display_name']

371 directory = os.path.basename(data['workspace']['current_dir'])

125 372

126# Extract values373 GREEN, YELLOW, RESET = '\033[32m', '\033[33m', '\033[0m'

127model = data['model']['display_name']

128current_dir = os.path.basename(data['workspace']['current_dir'])

129 374

130# Check for git branch

131git_branch = ""

132if os.path.exists('.git'):

133 try:375 try:

134 with open('.git/HEAD', 'r') as f:376 subprocess.check_output(['git', 'rev-parse', '--git-dir'], stderr=subprocess.DEVNULL)

135 ref = f.read().strip()377 branch = subprocess.check_output(['git', 'branch', '--show-current'], text=True).strip()

136 if ref.startswith('ref: refs/heads/'):378 staged_output = subprocess.check_output(['git', 'diff', '--cached', '--numstat'], text=True).strip()

137 git_branch = f" | 🌿 {ref.replace('ref: refs/heads/', '')}"379 modified_output = subprocess.check_output(['git', 'diff', '--numstat'], text=True).strip()

380 staged = len(staged_output.split('\n')) if staged_output else 0

381 modified = len(modified_output.split('\n')) if modified_output else 0

382

383 git_status = f"{GREEN}+{staged}{RESET}" if staged else ""

384 git_status += f"{YELLOW}~{modified}{RESET}" if modified else ""

385

386 print(f"[{model}] 📁 {directory} | 🌿 {branch} {git_status}")

138 except:387 except:

139 pass388 print(f"[{model}] 📁 {directory}")

389 ```

140 390

141print(f"[{model}] 📁 {current_dir}{git_branch}")391 ```javascript Node.js theme={null}

142```392 #!/usr/bin/env node

393 const { execSync } = require('child_process');

394 const path = require('path');

395

396 let input = '';

397 process.stdin.on('data', chunk => input += chunk);

398 process.stdin.on('end', () => {

399 const data = JSON.parse(input);

400 const model = data.model.display_name;

401 const dir = path.basename(data.workspace.current_dir);

143 402

144### Node.js Example403 const GREEN = '\x1b[32m', YELLOW = '\x1b[33m', RESET = '\x1b[0m';

145 404

146```javascript theme={null}405 try {

147#!/usr/bin/env node406 execSync('git rev-parse --git-dir', { stdio: 'ignore' });

407 const branch = execSync('git branch --show-current', { encoding: 'utf8' }).trim();

408 const staged = execSync('git diff --cached --numstat', { encoding: 'utf8' }).trim().split('\n').filter(Boolean).length;

409 const modified = execSync('git diff --numstat', { encoding: 'utf8' }).trim().split('\n').filter(Boolean).length;

148 410

149const fs = require('fs');411 let gitStatus = staged ? `${GREEN}+${staged}${RESET}` : '';

150const path = require('path');412 gitStatus += modified ? `${YELLOW}~${modified}${RESET}` : '';

151 413

152// Read JSON from stdin414 console.log(`[${model}] 📁 ${dir} | 🌿 ${branch} ${gitStatus}`);

153let input = '';415 } catch {

154process.stdin.on('data', chunk => input += chunk);416 console.log(`[${model}] 📁 ${dir}`);

155process.stdin.on('end', () => {417 }

418 });

419 ```

420</CodeGroup>

421

422### Cost and duration tracking

423

424Track your session's API costs and elapsed time. The `cost.total_cost_usd` field accumulates the cost of all API calls in the current session. The `cost.total_duration_ms` field measures total elapsed time since the session started, while `cost.total_api_duration_ms` tracks only the time spent waiting for API responses.

425

426Each script formats cost as currency and converts milliseconds to minutes and seconds:

427

428<Frame>

429 <img src="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-cost-tracking.png?fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=e3444a51fe6f3440c134bd5f1f08ad29" alt="A status line showing model name, session cost, and duration" data-og-width="588" width="588" data-og-height="180" height="180" data-path="images/statusline-cost-tracking.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-cost-tracking.png?w=280&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=b1d35fa8acd792f559b6b1662ed10204 280w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-cost-tracking.png?w=560&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=a3ed4330c3645fc28b87a6cab55be0b7 560w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-cost-tracking.png?w=840&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=386ee2ed68a7d520eba20eac54f7fe52 840w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-cost-tracking.png?w=1100&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=479c2515e53f46d5d1da3b87a6dd993a 1100w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-cost-tracking.png?w=1650&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=1340c7589a4cb89ec071234aba3571d1 1650w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-cost-tracking.png?w=2500&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=69056cf4fe3271770cac4dc1704bcd0a 2500w" />

430</Frame>

431

432<CodeGroup>

433 ```bash Bash theme={null}

434 #!/bin/bash

435 input=$(cat)

436

437 MODEL=$(echo "$input" | jq -r '.model.display_name')

438 COST=$(echo "$input" | jq -r '.cost.total_cost_usd // 0')

439 DURATION_MS=$(echo "$input" | jq -r '.cost.total_duration_ms // 0')

440

441 COST_FMT=$(printf '$%.2f' "$COST")

442 DURATION_SEC=$((DURATION_MS / 1000))

443 MINS=$((DURATION_SEC / 60))

444 SECS=$((DURATION_SEC % 60))

445

446 echo "[$MODEL] 💰 $COST_FMT | ⏱️ ${MINS}m ${SECS}s"

447 ```

448

449 ```python Python theme={null}

450 #!/usr/bin/env python3

451 import json, sys

452

453 data = json.load(sys.stdin)

454 model = data['model']['display_name']

455 cost = data.get('cost', {}).get('total_cost_usd', 0) or 0

456 duration_ms = data.get('cost', {}).get('total_duration_ms', 0) or 0

457

458 duration_sec = duration_ms // 1000

459 mins, secs = duration_sec // 60, duration_sec % 60

460

461 print(f"[{model}] 💰 ${cost:.2f} | ⏱️ {mins}m {secs}s")

462 ```

463

464 ```javascript Node.js theme={null}

465 #!/usr/bin/env node

466 let input = '';

467 process.stdin.on('data', chunk => input += chunk);

468 process.stdin.on('end', () => {

156 const data = JSON.parse(input);469 const data = JSON.parse(input);

470 const model = data.model.display_name;

471 const cost = data.cost?.total_cost_usd || 0;

472 const durationMs = data.cost?.total_duration_ms || 0;

473

474 const durationSec = Math.floor(durationMs / 1000);

475 const mins = Math.floor(durationSec / 60);

476 const secs = durationSec % 60;

477

478 console.log(`[${model}] 💰 $${cost.toFixed(2)} | ⏱️ ${mins}m ${secs}s`);

479 });

480 ```

481</CodeGroup>

482

483### Display multiple lines

484

485Your script can output multiple lines to create a richer display. Each `echo` statement produces a separate row in the status area.

486

487<Frame>

488 <img src="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=60f11387658acc9ff75158ae85f2ac87" alt="A multi-line status line showing model name, directory, git branch on the first line, and a context usage progress bar with cost and duration on the second line" data-og-width="776" width="776" data-og-height="212" height="212" data-path="images/statusline-multiline.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=280&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=2e448b44c332620e6c9c2be4ded992e5 280w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=560&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=f796af2db9c68ab2ddbc5136840b9551 560w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=840&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=d29c13d6164773198a0b2c47b31f6c09 840w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=1100&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=d7720e5f51310185c0c02152f6c10d8b 1100w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=1650&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=b4e008cde27990a8d5783e41e5b93246 1650w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-multiline.png?w=2500&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=40ab24813303dc2e4c09f2675f3faf6e 2500w" />

489</Frame>

490

491This example combines several techniques: threshold-based colors (green under 70%, yellow 70-89%, red 90%+), a progress bar, and git branch info. Each `print` or `echo` statement creates a separate row:

492

493<CodeGroup>

494 ```bash Bash theme={null}

495 #!/bin/bash

496 input=$(cat)

497

498 MODEL=$(echo "$input" | jq -r '.model.display_name')

499 DIR=$(echo "$input" | jq -r '.workspace.current_dir')

500 COST=$(echo "$input" | jq -r '.cost.total_cost_usd // 0')

501 PCT=$(echo "$input" | jq -r '.context_window.used_percentage // 0' | cut -d. -f1)

502 DURATION_MS=$(echo "$input" | jq -r '.cost.total_duration_ms // 0')

503

504 CYAN='\033[36m'; GREEN='\033[32m'; YELLOW='\033[33m'; RED='\033[31m'; RESET='\033[0m'

505

506 # Pick bar color based on context usage

507 if [ "$PCT" -ge 90 ]; then BAR_COLOR="$RED"

508 elif [ "$PCT" -ge 70 ]; then BAR_COLOR="$YELLOW"

509 else BAR_COLOR="$GREEN"; fi

510

511 FILLED=$((PCT / 10)); EMPTY=$((10 - FILLED))

512 BAR=$(printf "%${FILLED}s" | tr ' ' '█')$(printf "%${EMPTY}s" | tr ' ' '░')

513

514 MINS=$((DURATION_MS / 60000)); SECS=$(((DURATION_MS % 60000) / 1000))

515

516 BRANCH=""

517 git rev-parse --git-dir > /dev/null 2>&1 && BRANCH=" | 🌿 $(git branch --show-current 2>/dev/null)"

518

519 echo -e "${CYAN}[$MODEL]${RESET} 📁 ${DIR##*/}$BRANCH"

520 COST_FMT=$(printf '$%.2f' "$COST")

521 echo -e "${BAR_COLOR}${BAR}${RESET} ${PCT}% | ${YELLOW}${COST_FMT}${RESET} | ⏱️ ${MINS}m ${SECS}s"

522 ```

523

524 ```python Python theme={null}

525 #!/usr/bin/env python3

526 import json, sys, subprocess, os

527

528 data = json.load(sys.stdin)

529 model = data['model']['display_name']

530 directory = os.path.basename(data['workspace']['current_dir'])

531 cost = data.get('cost', {}).get('total_cost_usd', 0) or 0

532 pct = int(data.get('context_window', {}).get('used_percentage', 0) or 0)

533 duration_ms = data.get('cost', {}).get('total_duration_ms', 0) or 0

534

535 CYAN, GREEN, YELLOW, RED, RESET = '\033[36m', '\033[32m', '\033[33m', '\033[31m', '\033[0m'

536

537 bar_color = RED if pct >= 90 else YELLOW if pct >= 70 else GREEN

538 filled = pct // 10

539 bar = '█' * filled + '░' * (10 - filled)

540

541 mins, secs = duration_ms // 60000, (duration_ms % 60000) // 1000

542

543 try:

544 branch = subprocess.check_output(['git', 'branch', '--show-current'], text=True, stderr=subprocess.DEVNULL).strip()

545 branch = f" | 🌿 {branch}" if branch else ""

546 except:

547 branch = ""

548

549 print(f"{CYAN}[{model}]{RESET} 📁 {directory}{branch}")

550 print(f"{bar_color}{bar}{RESET} {pct}% | {YELLOW}${cost:.2f}{RESET} | ⏱️ {mins}m {secs}s")

551 ```

552

553 ```javascript Node.js theme={null}

554 #!/usr/bin/env node

555 const { execSync } = require('child_process');

556 const path = require('path');

557

558 let input = '';

559 process.stdin.on('data', chunk => input += chunk);

560 process.stdin.on('end', () => {

561 const data = JSON.parse(input);

562 const model = data.model.display_name;

563 const dir = path.basename(data.workspace.current_dir);

564 const cost = data.cost?.total_cost_usd || 0;

565 const pct = Math.floor(data.context_window?.used_percentage || 0);

566 const durationMs = data.cost?.total_duration_ms || 0;

567

568 const CYAN = '\x1b[36m', GREEN = '\x1b[32m', YELLOW = '\x1b[33m', RED = '\x1b[31m', RESET = '\x1b[0m';

569

570 const barColor = pct >= 90 ? RED : pct >= 70 ? YELLOW : GREEN;

571 const filled = Math.floor(pct / 10);

572 const bar = '█'.repeat(filled) + '░'.repeat(10 - filled);

573

574 const mins = Math.floor(durationMs / 60000);

575 const secs = Math.floor((durationMs % 60000) / 1000);

576

577 let branch = '';

578 try {

579 branch = execSync('git branch --show-current', { encoding: 'utf8', stdio: ['pipe', 'pipe', 'ignore'] }).trim();

580 branch = branch ? ` | 🌿 ${branch}` : '';

581 } catch {}

582

583 console.log(`${CYAN}[${model}]${RESET} 📁 ${dir}${branch}`);

584 console.log(`${barColor}${bar}${RESET} ${pct}% | ${YELLOW}$${cost.toFixed(2)}${RESET} | ⏱️ ${mins}m ${secs}s`);

585 });

586 ```

587</CodeGroup>

588

589### Clickable links

590

591This example creates a clickable link to your GitHub repository. It reads the git remote URL, converts SSH format to HTTPS with `sed`, and wraps the repo name in OSC 8 escape codes. Hold Cmd (macOS) or Ctrl (Windows/Linux) and click to open the link in your browser.

592

593<Frame>

594 <img src="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-links.png?fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=4bcc6e7deb7cf52f41ab85a219b52661" alt="A status line showing a clickable link to a GitHub repository" data-og-width="726" width="726" data-og-height="198" height="198" data-path="images/statusline-links.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-links.png?w=280&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=9386f78056f7be99599bcefe9e838180 280w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-links.png?w=560&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=d748012a0866c37dddc6babd4b7a88c4 560w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-links.png?w=840&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=bade8fbfcde957c1033c376c58b89131 840w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-links.png?w=1100&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=9f7e0c729ea093c3b39682619fd3f201 1100w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-links.png?w=1650&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=ccec17e90a89d82381888a4a9a8fa40e 1650w, https://mintcdn.com/claude-code/nibzesLaJVh4ydOq/images/statusline-links.png?w=2500&fit=max&auto=format&n=nibzesLaJVh4ydOq&q=85&s=4d2e34a4d2f24e174cae1256c84f9a52 2500w" />

595</Frame>

596

597Each script gets the git remote URL, converts SSH format to HTTPS, and wraps the repo name in OSC 8 escape codes. The Bash version uses `printf '%b'` which interprets backslash escapes more reliably than `echo -e` across different shells:

157 598

158 // Extract values599<CodeGroup>

600 ```bash Bash theme={null}

601 #!/bin/bash

602 input=$(cat)

603

604 MODEL=$(echo "$input" | jq -r '.model.display_name')

605

606 # Convert git SSH URL to HTTPS

607 REMOTE=$(git remote get-url origin 2>/dev/null | sed 's/git@github.com:/https:\/\/github.com\//' | sed 's/\.git$//')

608

609 if [ -n "$REMOTE" ]; then

610 REPO_NAME=$(basename "$REMOTE")

611 # OSC 8 format: \e]8;;URL\a then TEXT then \e]8;;\a

612 # printf %b interprets escape sequences reliably across shells

613 printf '%b' "[$MODEL] 🔗 \e]8;;${REMOTE}\a${REPO_NAME}\e]8;;\a\n"

614 else

615 echo "[$MODEL]"

616 fi

617 ```

618

619 ```python Python theme={null}

620 #!/usr/bin/env python3

621 import json, sys, subprocess, re, os

622

623 data = json.load(sys.stdin)

624 model = data['model']['display_name']

625

626 # Get git remote URL

627 try:

628 remote = subprocess.check_output(

629 ['git', 'remote', 'get-url', 'origin'],

630 stderr=subprocess.DEVNULL, text=True

631 ).strip()

632 # Convert SSH to HTTPS format

633 remote = re.sub(r'^git@github\.com:', 'https://github.com/', remote)

634 remote = re.sub(r'\.git$', '', remote)

635 repo_name = os.path.basename(remote)

636 # OSC 8 escape sequences

637 link = f"\033]8;;{remote}\a{repo_name}\033]8;;\a"

638 print(f"[{model}] 🔗 {link}")

639 except:

640 print(f"[{model}]")

641 ```

642

643 ```javascript Node.js theme={null}

644 #!/usr/bin/env node

645 const { execSync } = require('child_process');

646 const path = require('path');

647

648 let input = '';

649 process.stdin.on('data', chunk => input += chunk);

650 process.stdin.on('end', () => {

651 const data = JSON.parse(input);

159 const model = data.model.display_name;652 const model = data.model.display_name;

160 const currentDir = path.basename(data.workspace.current_dir);

161 653

162 // Check for git branch

163 let gitBranch = '';

164 try {654 try {

165 const headContent = fs.readFileSync('.git/HEAD', 'utf8').trim();655 let remote = execSync('git remote get-url origin', { encoding: 'utf8', stdio: ['pipe', 'pipe', 'ignore'] }).trim();

166 if (headContent.startsWith('ref: refs/heads/')) {656 // Convert SSH to HTTPS format

167 gitBranch = ` | 🌿 ${headContent.replace('ref: refs/heads/', '')}`;657 remote = remote.replace(/^git@github\.com:/, 'https://github.com/').replace(/\.git$/, '');

658 const repoName = path.basename(remote);

659 // OSC 8 escape sequences

660 const link = `\x1b]8;;${remote}\x07${repoName}\x1b]8;;\x07`;

661 console.log(`[${model}] 🔗 ${link}`);

662 } catch {

663 console.log(`[${model}]`);

168 }664 }

169 } catch (e) {665 });

170 // Not a git repo or can't read HEAD666 ```

667</CodeGroup>

668

669### Cache expensive operations

670

671Your status line script runs frequently during active sessions. Commands like `git status` or `git diff` can be slow, especially in large repositories. This example caches git information to a temp file and only refreshes it every 5 seconds.

672

673Use a stable, fixed filename for the cache file like `/tmp/statusline-git-cache`. Each status line invocation runs as a new process, so process-based identifiers like `$$`, `os.getpid()`, or `process.pid` produce a different value every time and the cache is never reused.

674

675Each script checks if the cache file is missing or older than 5 seconds before running git commands:

676

677<CodeGroup>

678 ```bash Bash theme={null}

679 #!/bin/bash

680 input=$(cat)

681

682 MODEL=$(echo "$input" | jq -r '.model.display_name')

683 DIR=$(echo "$input" | jq -r '.workspace.current_dir')

684

685 CACHE_FILE="/tmp/statusline-git-cache"

686 CACHE_MAX_AGE=5 # seconds

687

688 cache_is_stale() {

689 [ ! -f "$CACHE_FILE" ] || \

690 # stat -f %m is macOS, stat -c %Y is Linux

691 [ $(($(date +%s) - $(stat -f %m "$CACHE_FILE" 2>/dev/null || stat -c %Y "$CACHE_FILE" 2>/dev/null || echo 0))) -gt $CACHE_MAX_AGE ]

171 }692 }

172 693

173 console.log(`[${model}] 📁 ${currentDir}${gitBranch}`);694 if cache_is_stale; then

174});695 if git rev-parse --git-dir > /dev/null 2>&1; then

175```696 BRANCH=$(git branch --show-current 2>/dev/null)

697 STAGED=$(git diff --cached --numstat 2>/dev/null | wc -l | tr -d ' ')

698 MODIFIED=$(git diff --numstat 2>/dev/null | wc -l | tr -d ' ')

699 echo "$BRANCH|$STAGED|$MODIFIED" > "$CACHE_FILE"

700 else

701 echo "||" > "$CACHE_FILE"

702 fi

703 fi

176 704

177### Helper Function Approach705 IFS='|' read -r BRANCH STAGED MODIFIED < "$CACHE_FILE"

~~178~~

179For more complex bash scripts, you can create helper functions:

~~180~~

181```bash theme={null}

182#!/bin/bash

183# Read JSON input once

184input=$(cat)

~~185~~

186# Helper functions for common extractions

187get_model_name() { echo "$input" | jq -r '.model.display_name'; }

188get_current_dir() { echo "$input" | jq -r '.workspace.current_dir'; }

189get_project_dir() { echo "$input" | jq -r '.workspace.project_dir'; }

190get_version() { echo "$input" | jq -r '.version'; }

191get_cost() { echo "$input" | jq -r '.cost.total_cost_usd'; }

192get_duration() { echo "$input" | jq -r '.cost.total_duration_ms'; }

193get_lines_added() { echo "$input" | jq -r '.cost.total_lines_added'; }

194get_lines_removed() { echo "$input" | jq -r '.cost.total_lines_removed'; }

195get_input_tokens() { echo "$input" | jq -r '.context_window.total_input_tokens'; }

196get_output_tokens() { echo "$input" | jq -r '.context_window.total_output_tokens'; }

197get_context_window_size() { echo "$input" | jq -r '.context_window.context_window_size'; }

~~198~~

199# Use the helpers

200MODEL=$(get_model_name)

201DIR=$(get_current_dir)

202echo "[$MODEL] 📁 ${DIR##*/}"

203```

204 706

205### Context Window Usage707 if [ -n "$BRANCH" ]; then

708 echo "[$MODEL] 📁 ${DIR##*/} | 🌿 $BRANCH +$STAGED ~$MODIFIED"

709 else

710 echo "[$MODEL] 📁 ${DIR##*/}"

711 fi

712 ```

206 713

207Display the percentage of context window consumed. The `context_window` object contains:714 ```python Python theme={null}

715 #!/usr/bin/env python3

716 import json, sys, subprocess, os, time

208 717

209* `total_input_tokens` / `total_output_tokens`: Cumulative totals across the entire session718 data = json.load(sys.stdin)

210* `current_usage`: Current context window usage from the last API call (may be `null` if no messages yet)719 model = data['model']['display_name']

211 * `input_tokens`: Input tokens in current context720 directory = os.path.basename(data['workspace']['current_dir'])

212 * `output_tokens`: Output tokens generated

213 * `cache_creation_input_tokens`: Tokens written to cache

214 * `cache_read_input_tokens`: Tokens read from cache

215 721

216For accurate context percentage, use `current_usage` which reflects the actual context window state:722 CACHE_FILE = "/tmp/statusline-git-cache"

723 CACHE_MAX_AGE = 5 # seconds

217 724

218```bash theme={null}725 def cache_is_stale():

219#!/bin/bash726 if not os.path.exists(CACHE_FILE):

220input=$(cat)727 return True

728 return time.time() - os.path.getmtime(CACHE_FILE) > CACHE_MAX_AGE

221 729

222MODEL=$(echo "$input" | jq -r '.model.display_name')730 if cache_is_stale():

223CONTEXT_SIZE=$(echo "$input" | jq -r '.context_window.context_window_size')731 try:

224USAGE=$(echo "$input" | jq '.context_window.current_usage')732 subprocess.check_output(['git', 'rev-parse', '--git-dir'], stderr=subprocess.DEVNULL)

733 branch = subprocess.check_output(['git', 'branch', '--show-current'], text=True).strip()

734 staged = subprocess.check_output(['git', 'diff', '--cached', '--numstat'], text=True).strip()

735 modified = subprocess.check_output(['git', 'diff', '--numstat'], text=True).strip()

736 staged_count = len(staged.split('\n')) if staged else 0

737 modified_count = len(modified.split('\n')) if modified else 0

738 with open(CACHE_FILE, 'w') as f:

739 f.write(f"{branch}|{staged_count}|{modified_count}")

740 except:

741 with open(CACHE_FILE, 'w') as f:

742 f.write("||")

743

744 with open(CACHE_FILE) as f:

745 branch, staged, modified = f.read().strip().split('|')

746

747 if branch:

748 print(f"[{model}] 📁 {directory} | 🌿 {branch} +{staged} ~{modified}")

749 else:

750 print(f"[{model}] 📁 {directory}")

751 ```

752

753 ```javascript Node.js theme={null}

754 #!/usr/bin/env node

755 const { execSync } = require('child_process');

756 const fs = require('fs');

757 const path = require('path');

758

759 let input = '';

760 process.stdin.on('data', chunk => input += chunk);

761 process.stdin.on('end', () => {

762 const data = JSON.parse(input);

763 const model = data.model.display_name;

764 const dir = path.basename(data.workspace.current_dir);

225 765

226if [ "$USAGE" != "null" ]; then766 const CACHE_FILE = '/tmp/statusline-git-cache';

227 # Calculate current context from current_usage fields767 const CACHE_MAX_AGE = 5; // seconds

228 CURRENT_TOKENS=$(echo "$USAGE" | jq '.input_tokens + .cache_creation_input_tokens + .cache_read_input_tokens')768

229 PERCENT_USED=$((CURRENT_TOKENS * 100 / CONTEXT_SIZE))769 const cacheIsStale = () => {

230 echo "[$MODEL] Context: ${PERCENT_USED}%"770 if (!fs.existsSync(CACHE_FILE)) return true;

231else771 return (Date.now() / 1000) - fs.statSync(CACHE_FILE).mtimeMs / 1000 > CACHE_MAX_AGE;

232 echo "[$MODEL] Context: 0%"772 };

233fi773

234```774 if (cacheIsStale()) {

775 try {

776 execSync('git rev-parse --git-dir', { stdio: 'ignore' });

777 const branch = execSync('git branch --show-current', { encoding: 'utf8' }).trim();

778 const staged = execSync('git diff --cached --numstat', { encoding: 'utf8' }).trim().split('\n').filter(Boolean).length;

779 const modified = execSync('git diff --numstat', { encoding: 'utf8' }).trim().split('\n').filter(Boolean).length;

780 fs.writeFileSync(CACHE_FILE, `${branch}|${staged}|${modified}`);

781 } catch {

782 fs.writeFileSync(CACHE_FILE, '||');

783 }

784 }

785

786 const [branch, staged, modified] = fs.readFileSync(CACHE_FILE, 'utf8').trim().split('|');

787

788 if (branch) {

789 console.log(`[${model}] 📁 ${dir} | 🌿 ${branch} +${staged} ~${modified}`);

790 } else {

791 console.log(`[${model}] 📁 ${dir}`);

792 }

793 });

794 ```

795</CodeGroup>

235 796

236## Tips797## Tips

237 798

238* Keep your status line concise - it should fit on one line799* **Test with mock input**: `echo '{"model":{"display_name":"Opus"},"context_window":{"used_percentage":25}}' | ./statusline.sh`

239* Use emojis (if your terminal supports them) and colors to make information scannable800* **Keep output short**: the status bar has limited width, so long output may get truncated or wrap awkwardly

240* Use `jq` for JSON parsing in Bash (see examples above)801* **Cache slow operations**: your script runs frequently during active sessions, so commands like `git status` can cause lag. See the [caching example](#cache-expensive-operations) for how to handle this.

241* Test your script by running it manually with mock JSON input: `echo '{"model":{"display_name":"Test"},"workspace":{"current_dir":"/test"}}' | ./statusline.sh`802

242* Consider caching expensive operations (like git status) if needed803Community projects like [ccstatusline](https://github.com/sirmalloc/ccstatusline) and [starship-claude](https://github.com/martinemde/starship-claude) provide pre-built configurations with themes and additional features.

243 804

244## Troubleshooting805## Troubleshooting

245 806

246* If your status line doesn't appear, check that your script is executable (`chmod +x`)807**Status line not appearing**

247* Ensure your script outputs to stdout (not stderr)808

809* Verify your script is executable: `chmod +x ~/.claude/statusline.sh`

810* Check that your script outputs to stdout, not stderr

811* Run your script manually to verify it produces output

812* If `disableAllHooks` is set to `true` in your settings, the status line is also disabled. Remove this setting or set it to `false` to re-enable.

813

814**Status line shows `--` or empty values**

815

816* Fields may be `null` before the first API response completes

817* Handle null values in your script with fallbacks such as `// 0` in jq

818* Restart Claude Code if values remain empty after multiple messages

819

820**Context percentage shows unexpected values**

821

822* Use `used_percentage` for accurate context state rather than cumulative totals

823* The `total_input_tokens` and `total_output_tokens` are cumulative across the session and may exceed the context window size

824* Context percentage may differ from `/context` output due to when each is calculated

825

826**OSC 8 links not clickable**

827

828* Verify your terminal supports OSC 8 hyperlinks (iTerm2, Kitty, WezTerm)

829* Terminal.app does not support clickable links

830* SSH and tmux sessions may strip OSC sequences depending on configuration

831* If escape sequences appear as literal text like `\e]8;;`, use `printf '%b'` instead of `echo -e` for more reliable escape handling

832

833**Display glitches with escape sequences**

834

835* Complex escape sequences (ANSI colors, OSC 8 links) can occasionally cause garbled output if they overlap with other UI updates

836* If you see corrupted text, try simplifying your script to plain text output

837* Multi-line status lines with escape codes are more prone to rendering issues than single-line plain text

838

839**Script errors or hangs**

248 840

841* Scripts that exit with non-zero codes or produce no output cause the status line to go blank

842* Slow scripts block the status line from updating until they complete. Keep scripts fast to avoid stale output.

843* If a new update triggers while a slow script is running, the in-flight script is cancelled

844* Test your script independently with mock input before configuring it

249 845

846**Notifications share the status line row**

250 847

251> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt848* System notifications like MCP server errors, auto-updates, and token warnings display on the right side of the same row as your status line

849* Enabling verbose mode adds a token counter to this area

850* On narrow terminals, these notifications may truncate your status line output

sub-agents.md +531 −288

Details

~~1# Subagents~~1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

5# Create custom subagents

2 6

3> Create and use specialized AI subagents in Claude Code for task-specific workflows and improved context management.7> Create and use specialized AI subagents in Claude Code for task-specific workflows and improved context management.

4 8

5Custom subagents in Claude Code are specialized AI assistants that can be invoked to handle specific types of tasks. They enable more efficient problem-solving by providing task-specific configurations with customized system prompts, tools and a separate context window.9Subagents are specialized AI assistants that handle specific types of tasks. Each subagent runs in its own context window with a custom system prompt, specific tool access, and independent permissions. When Claude encounters a task that matches a subagent's description, it delegates to that subagent, which works independently and returns results.

11<Note>

12 If you need multiple agents working in parallel and communicating with each other, see [agent teams](/en/agent-teams) instead. Subagents work within a single session; agent teams coordinate across separate sessions.

13</Note>

15Subagents help you:

17* **Preserve context** by keeping exploration and implementation out of your main conversation

18* **Enforce constraints** by limiting which tools a subagent can use

19* **Reuse configurations** across projects with user-level subagents

20* **Specialize behavior** with focused system prompts for specific domains

21* **Control costs** by routing tasks to faster, cheaper models like Haiku

23Claude uses each subagent's description to decide when to delegate tasks. When you create a subagent, write a clear description so Claude knows when to use it.

25Claude Code includes several built-in subagents like **Explore**, **Plan**, and **general-purpose**. You can also create custom subagents to handle specific tasks. This page covers the [built-in subagents](#built-in-subagents), [how to create your own](#quickstart-create-your-first-subagent), [full configuration options](#configure-subagents), [patterns for working with subagents](#work-with-subagents), and [example subagents](#example-subagents).

27## Built-in subagents

29Claude Code includes built-in subagents that Claude automatically uses when appropriate. Each inherits the parent conversation's permissions with additional tool restrictions.

31<Tabs>

32 <Tab title="Explore">

33 A fast, read-only agent optimized for searching and analyzing codebases.

6 34

~~7## What are subagents?~~35 * **Model**: Haiku (fast, low-latency)

36 * **Tools**: Read-only tools (denied access to Write and Edit tools)

37 * **Purpose**: File discovery, code search, codebase exploration

8 38

~~9Subagents are pre-configured AI personalities that Claude Code can delegate tasks to. Each subagent:~~39 Claude delegates to Explore when it needs to search or understand a codebase without making changes. This keeps exploration results out of your main conversation context.

10 40

~~11* Has a specific purpose and expertise area~~41 When invoking Explore, Claude specifies a thoroughness level: **quick** for targeted lookups, **medium** for balanced exploration, or **very thorough** for comprehensive analysis.

~~12* Uses its own context window separate from the main conversation~~42 </Tab>

~~13* Can be configured with specific tools it's allowed to use~~

~~14* Includes a custom system prompt that guides its behavior~~

15 43

~~16When Claude Code encounters a task that matches a subagent's expertise, it can delegate that task to the specialized subagent, which works independently and returns results.~~44 <Tab title="Plan">

45 A research agent used during [plan mode](/en/common-workflows#use-plan-mode-for-safe-code-analysis) to gather context before presenting a plan.

17 46

~~18## Key benefits~~47 * **Model**: Inherits from main conversation

48 * **Tools**: Read-only tools (denied access to Write and Edit tools)

49 * **Purpose**: Codebase research for planning

19 50

~~20<CardGroup cols={2}>~~51 When you're in plan mode and Claude needs to understand your codebase, it delegates research to the Plan subagent. This prevents infinite nesting (subagents cannot spawn other subagents) while still gathering necessary context.

~~21 <Card title="Context preservation" icon="layer-group">~~52 </Tab>

~~22 Each subagent operates in its own context, preventing pollution of the main conversation and keeping it focused on high-level objectives.~~

~~23 </Card>~~

24 53

~~25 <Card title="Specialized expertise" icon="brain">~~54 <Tab title="General-purpose">

~~26 Subagents can be fine-tuned with detailed instructions for specific domains, leading to higher success rates on designated tasks.~~55 A capable agent for complex, multi-step tasks that require both exploration and action.

~~27 </Card>~~

28 56

~~29 <Card title="Reusability" icon="rotate">~~57 * **Model**: Inherits from main conversation

~~30 Once created, you can use subagents across different projects and share them with your team for consistent workflows.~~58 * **Tools**: All tools

~~31 </Card>~~59 * **Purpose**: Complex research, multi-step operations, code modifications

32 60

~~33 <Card title="Flexible permissions" icon="shield-check">~~61 Claude delegates to general-purpose when the task requires both exploration and modification, complex reasoning to interpret results, or multiple dependent steps.

~~34 Each subagent can have different tool access levels, allowing you to limit powerful tools to specific subagent types.~~62 </Tab>

~~35 </Card>~~

~~36</CardGroup>~~

37 63

~~38## Quick start~~64 <Tab title="Other">

65 Claude Code includes additional helper agents for specific tasks. These are typically invoked automatically, so you don't need to use them directly.

39 66

~~40To create your first subagent:~~67 | Agent | Model | When Claude uses it |

68 | :---------------- | :------- | :------------------------------------------------------- |

69 | Bash | Inherits | Running terminal commands in a separate context |

70 | statusline-setup | Sonnet | When you run `/statusline` to configure your status line |

71 | Claude Code Guide | Haiku | When you ask questions about Claude Code features |

72 </Tab>

73</Tabs>

75Beyond these built-in subagents, you can create your own with custom prompts, tool restrictions, permission modes, hooks, and skills. The following sections show how to get started and customize subagents.

77## Quickstart: create your first subagent

79Subagents are defined in Markdown files with YAML frontmatter. You can [create them manually](#write-subagent-files) or use the `/agents` command.

81This walkthrough guides you through creating a user-level subagent with the `/agent` command. The subagent reviews code and suggests improvements for the codebase.

41 82

42<Steps>83<Steps>

43 <Step title="Open the subagents interface">84 <Step title="Open the subagents interface">

~~44 Run the following command:~~85 In Claude Code, run:

45 86

46 ```87 ```

47 /agents88 /agents

48 ```89 ```

49 </Step>90 </Step>

50 91

~~51 <Step title="Select 'Create New Agent'">~~92 <Step title="Create a new user-level agent">

~~52 Choose whether to create a project-level or user-level subagent~~93 Select **Create new agent**, then choose **User-level**. This saves the subagent to `~/.claude/agents/` so it's available in all your projects.

94 </Step>

96 <Step title="Generate with Claude">

97 Select **Generate with Claude**. When prompted, describe the subagent:

99 ```

100 A code improvement agent that scans files and suggests improvements

101 for readability, performance, and best practices. It should explain

102 each issue, show the current code, and provide an improved version.

103 ```

104

105 Claude generates the system prompt and configuration. Press `e` to open it in your editor if you want to customize it.

53 </Step>106 </Step>

54 107

~~55 <Step title="Define the subagent">~~108 <Step title="Select tools">

~~56 * **Recommended**: generate with Claude first, then customize to make it yours~~109 For a read-only reviewer, deselect everything except **Read-only tools**. If you keep all tools selected, the subagent inherits all tools available to the main conversation.

~~57 * Describe your subagent in detail, including when Claude should use it~~

~~58 * Select the tools you want to grant access to, or leave this blank to inherit all tools~~

~~59 * The interface shows all available tools~~

~~60 * If you're generating with Claude, you can also edit the system prompt in your own editor by pressing `e`~~

61 </Step>110 </Step>

62 111

~~63 <Step title="Save and use">~~112 <Step title="Select model">

~~64 Your subagent is now available. Claude uses it automatically when appropriate, or you can invoke it explicitly:~~113 Choose which model the subagent uses. For this example agent, select **Sonnet**, which balances capability and speed for analyzing code patterns.

114 </Step>

115

116 <Step title="Choose a color">

117 Pick a background color for the subagent. This helps you identify which subagent is running in the UI.

118 </Step>

119

120 <Step title="Save and try it out">

121 Save the subagent. It's available immediately (no restart needed). Try it:

65 122

66 ```123 ```

~~67 > Use the code-reviewer subagent to check my recent changes~~124 Use the code-improver agent to suggest improvements in this project

68 ```125 ```

126

127 Claude delegates to your new subagent, which scans the codebase and returns improvement suggestions.

69 </Step>128 </Step>

70</Steps>129</Steps>

71 130

~~72## Subagent configuration~~131You now have a subagent you can use in any project on your machine to analyze codebases and suggest improvements.

73 132

~~74### File locations~~133You can also create subagents manually as Markdown files, define them via CLI flags, or distribute them through plugins. The following sections cover all configuration options.

75 134

~~76Subagents are stored as Markdown files with YAML frontmatter in two possible locations:~~135## Configure subagents

77 136

~~79| :-------------------- | :------------------ | :---------------------------- | :------- |~~

82 138

~~83When subagent names conflict, project-level subagents take precedence over user-level subagents.~~139The `/agents` command provides an interactive interface for managing subagents. Run `/agents` to:

84 140

~~85### Plugin agents~~141* View all available subagents (built-in, user, project, and plugin)

142* Create new subagents with guided setup or Claude generation

143* Edit existing subagent configuration and tool access

144* Delete custom subagents

145* See which subagents are active when duplicates exist

86 146

~~87[Plugins](/en/plugins) can provide custom subagents that integrate seamlessly with Claude Code. Plugin agents work identically to user-defined agents and appear in the `/agents` interface.~~147This is the recommended way to create and manage subagents. For manual creation or automation, you can also add subagent files directly.

88 148

~~89**Plugin agent locations**: plugins include agents in their `agents/` directory (or custom paths specified in the plugin manifest).~~149### Choose the subagent scope

90 150

~~91**Using plugin agents**:~~151Subagents are Markdown files with YAML frontmatter. Store them in different locations depending on scope. When multiple subagents share the same name, the higher-priority location wins.

92 152

~~94* Can be invoked explicitly: "Use the code-reviewer agent from the security-plugin"~~154| :--------------------------- | :---------------------- | :---------- | :------------------------------------ |

97 159

~~98See the [plugin components reference](/en/plugins-reference#agents) for details on creating plugin agents.~~160**Project subagents** (`.claude/agents/`) are ideal for subagents specific to a codebase. Check them into version control so your team can use and improve them collaboratively.

99 161

100### CLI-based configuration162**User subagents** (`~/.claude/agents/`) are personal subagents available in all your projects.

101 163

102You can also define subagents dynamically using the `--agents` CLI flag, which accepts a JSON object:164**CLI-defined subagents** are passed as JSON when launching Claude Code. They exist only for that session and aren't saved to disk, making them useful for quick testing or automation scripts:

103 165

104```bash theme={null}166```bash theme={null}

105claude --agents '{167claude --agents '{

112}'174}'

113```175```

114 176

115**Priority**: CLI-defined subagents have lower priority than project-level subagents but higher priority than user-level subagents.177The `--agents` flag accepts JSON with the same [frontmatter](#supported-frontmatter-fields) fields as file-based subagents: `description`, `prompt`, `tools`, `disallowedTools`, `model`, `permissionMode`, `mcpServers`, `hooks`, `maxTurns`, `skills`, and `memory`. Use `prompt` for the system prompt, equivalent to the markdown body in file-based subagents. See the [CLI reference](/en/cli-reference#agents-flag-format) for the full JSON format.

~~116~~

117**Use case**: This approach is useful for:

118 178

119* Quick testing of subagent configurations179**Plugin subagents** come from [plugins](/en/plugins) you've installed. They appear in `/agents` alongside your custom subagents. See the [plugin components reference](/en/plugins-reference#agents) for details on creating plugin subagents.

120* Session-specific subagents that don't need to be saved

121* Automation scripts that need custom subagents

122* Sharing subagent definitions in documentation or scripts

123 180

124For detailed information about the JSON format and all available options, see the [CLI reference documentation](/en/cli-reference#agents-flag-format).181### Write subagent files

125 182

126### File format183Subagent files use YAML frontmatter for configuration, followed by the system prompt in Markdown:

127 184

128Each subagent is defined in a Markdown file with this structure:185<Note>

186 Subagents are loaded at session start. If you create a subagent by manually adding a file, restart your session or use `/agents` to load it immediately.

187</Note>

129 188

130```markdown theme={null}189```markdown theme={null}

131---190---

132name: your-sub-agent-name191name: code-reviewer

133description: Description of when this subagent should be invoked192description: Reviews code for quality and best practices

134tools: tool1, tool2, tool3 # Optional - inherits all tools if omitted193tools: Read, Glob, Grep

135model: sonnet # Optional - specify model alias or 'inherit'194model: sonnet

136permissionMode: default # Optional - permission mode for the subagent

137skills: skill1, skill2 # Optional - skills to auto-load

138---195---

139 196

140Your subagent's system prompt goes here. This can be multiple paragraphs197You are a code reviewer. When invoked, analyze the code and provide

141and should clearly define the subagent's role, capabilities, and approach198specific, actionable feedback on quality, security, and best practices.

142to solving problems.

~~143~~

144Include specific instructions, best practices, and any constraints

145the subagent should follow.

146```199```

147 200

148#### Configuration fields201The frontmatter defines the subagent's metadata and configuration. The body becomes the system prompt that guides the subagent's behavior. Subagents receive only this system prompt (plus basic environment details like working directory), not the full Claude Code system prompt.

202

203#### Supported frontmatter fields

204

205The following fields can be used in the YAML frontmatter. Only `name` and `description` are required.

149 206

151| :--------------- | :------- | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |208| :---------------- | :------- | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

152| `name` | Yes | Unique identifier using lowercase letters and hyphens |209| `name` | Yes | Unique identifier using lowercase letters and hyphens |

153| `description` | Yes | Natural language description of the subagent's purpose |210| `description` | Yes | When Claude should delegate to this subagent |

154| `tools` | No | Comma-separated list of specific tools. If omitted, inherits all tools from the main thread |211| `tools` | No | [Tools](#available-tools) the subagent can use. Inherits all tools if omitted |

155| `model` | No | Model to use for this subagent. Can be a model alias (`sonnet`, `opus`, `haiku`) or `'inherit'` to use the main conversation's model. If omitted, defaults to the [configured subagent model](/en/model-config) |212| `disallowedTools` | No | Tools to deny, removed from inherited or specified list |

156| `permissionMode` | No | Permission mode for the subagent. Valid values: `default`, `acceptEdits`, `dontAsk`, `bypassPermissions`, `plan`, `ignore`. Controls how the subagent handles permission requests |213| `model` | No | [Model](#choose-a-model) to use: `sonnet`, `opus`, `haiku`, or `inherit`. Defaults to `inherit` |

157| `skills` | No | Comma-separated list of skill names to auto-load when the subagent starts. Subagents do not inherit Skills from the parent conversation. If omitted, no Skills are preloaded. |214| `permissionMode` | No | [Permission mode](#permission-modes): `default`, `acceptEdits`, `delegate`, `dontAsk`, `bypassPermissions`, or `plan` |

215| `maxTurns` | No | Maximum number of agentic turns before the subagent stops |

216| `skills` | No | [Skills](/en/skills) to load into the subagent's context at startup. The full skill content is injected, not just made available for invocation. Subagents don't inherit skills from the parent conversation |

217| `mcpServers` | No | [MCP servers](/en/mcp) available to this subagent. Each entry is either a server name referencing an already-configured server (e.g., `"slack"`) or an inline definition with the server name as key and a full [MCP server config](/en/mcp#configure-mcp-servers) as value |

218| `hooks` | No | [Lifecycle hooks](#define-hooks-for-subagents) scoped to this subagent |

219| `memory` | No | [Persistent memory scope](#enable-persistent-memory): `user`, `project`, or `local`. Enables cross-session learning |

158 220

159### Model selection221### Choose a model

160 222

161The `model` field allows you to control which [AI model](/en/model-config) the subagent uses:223The `model` field controls which [AI model](/en/model-config) the subagent uses:

162 224

163* **Model alias**: Use one of the available aliases: `sonnet`, `opus`, or `haiku`225* **Model alias**: Use one of the available aliases: `sonnet`, `opus`, or `haiku`

164* **`'inherit'`**: Use the same model as the main conversation (useful for consistency)226* **inherit**: Use the same model as the main conversation

165* **Omitted**: If not specified, uses the default model configured for subagents (`sonnet`)227* **Omitted**: If not specified, defaults to `inherit` (uses the same model as the main conversation)

166 228

167<Note>229### Control subagent capabilities

168 Using `'inherit'` is particularly useful when you want your subagents to adapt to the model choice of the main conversation, ensuring consistent capabilities and response style throughout your session.

169</Note>

170 230

171### Available tools231You can control what subagents can do through tool access, permission modes, and conditional rules.

172 232

173Subagents can be granted access to any of Claude Code's internal tools. See the [tools documentation](/en/settings#tools-available-to-claude) for a complete list of available tools.233#### Available tools

174 234

175<Tip>235Subagents can use any of Claude Code's [internal tools](/en/settings#tools-available-to-claude). By default, subagents inherit all tools from the main conversation, including MCP tools.

176 **Recommended:** Use the `/agents` command to modify tool access - it provides an interactive interface that lists all available tools, including any connected MCP server tools, making it easier to select the ones you need.

177</Tip>

178 236

179You have two options for configuring tools:237To restrict tools, use the `tools` field (allowlist) or `disallowedTools` field (denylist):

180 238

181* **Omit the `tools` field** to inherit all tools from the main thread (default), including MCP tools239```yaml theme={null}

182* **Specify individual tools** as a comma-separated list for more granular control (can be edited manually or via `/agents`)240---

241name: safe-researcher

242description: Research agent with restricted capabilities

243tools: Read, Grep, Glob, Bash

244disallowedTools: Write, Edit

245---

246```

183 247

184**MCP Tools**: Subagents can access MCP tools from configured MCP servers. When the `tools` field is omitted, subagents inherit all MCP tools available to the main thread.248#### Restrict which subagents can be spawned

185 249

186## Managing subagents250When an agent runs as the main thread with `claude --agent`, it can spawn subagents using the Task tool. To restrict which subagent types it can spawn, use `Task(agent_type)` syntax in the `tools` field:

187 251

188### Using the /agents command (Recommended)252```yaml theme={null}

253---

254name: coordinator

255description: Coordinates work across specialized agents

256tools: Task(worker, researcher), Read, Bash

257---

258```

189 259

190The `/agents` command provides a comprehensive interface for subagent management:260This is an allowlist: only the `worker` and `researcher` subagents can be spawned. If the agent tries to spawn any other type, the request fails and the agent sees only the allowed types in its prompt. To block specific agents while allowing all others, use [`permissions.deny`](#disable-specific-subagents) instead.

191 261

192```262To allow spawning any subagent without restrictions, use `Task` without parentheses:

193/agents263

264```yaml theme={null}

265tools: Task, Read, Bash

194```266```

195 267

196This opens an interactive menu where you can:268If `Task` is omitted from the `tools` list entirely, the agent cannot spawn any subagents. This restriction only applies to agents running as the main thread with `claude --agent`. Subagents cannot spawn other subagents, so `Task(agent_type)` has no effect in subagent definitions.

197 269

198* View all available subagents (built-in, user, and project)270#### Permission modes

199* Create new subagents with guided setup

200* Edit existing custom subagents, including their tool access

201* Delete custom subagents

202* See which subagents are active when duplicates exist

203* **Manage tool permissions** with a complete list of available tools

204 271

205### Direct file management272The `permissionMode` field controls how the subagent handles permission prompts. Subagents inherit the permission context from the main conversation but can override the mode.

206 273

207You can also manage subagents by working directly with their files:274| Mode | Behavior |

275| :------------------ | :------------------------------------------------------------------------------------------------------------------- |

276| `default` | Standard permission checking with prompts |

277| `acceptEdits` | Auto-accept file edits |

278| `dontAsk` | Auto-deny permission prompts (explicitly allowed tools still work) |

279| `delegate` | Coordination-only mode for [agent team](/en/agent-teams#use-delegate-mode) leads. Restricts to team management tools |

280| `bypassPermissions` | Skip all permission checks |

281| `plan` | Plan mode (read-only exploration) |

208 282

209```bash theme={null}283<Warning>

210# Create a project subagent284 Use `bypassPermissions` with caution. It skips all permission checks, allowing the subagent to execute any operation without approval.

211mkdir -p .claude/agents285</Warning>

212echo '---286

213name: test-runner287If the parent uses `bypassPermissions`, this takes precedence and cannot be overridden.

214description: Use proactively to run tests and fix failures288

289#### Preload skills into subagents

215 290

216You are a test automation expert. When you see code changes, proactively run the appropriate tests. If tests fail, analyze the failures and fix them while preserving the original test intent.' > .claude/agents/test-runner.md291Use the `skills` field to inject skill content into a subagent's context at startup. This gives the subagent domain knowledge without requiring it to discover and load skills during execution.

292

293```yaml theme={null}

294---

295name: api-developer

296description: Implement API endpoints following team conventions

297skills:

298 - api-conventions

299 - error-handling-patterns

300---

217 301

218# Create a user subagent302Implement API endpoints. Follow the conventions and patterns from the preloaded skills.

219mkdir -p ~/.claude/agents

220# ... create subagent file

221```303```

222 304

305The full content of each skill is injected into the subagent's context, not just made available for invocation. Subagents don't inherit skills from the parent conversation; you must list them explicitly.

306

223<Note>307<Note>

224 Subagents created by manually adding files will be loaded the next time you start a Claude Code session. To create and use a subagent immediately without restarting, use the `/agents` command instead.308 This is the inverse of [running a skill in a subagent](/en/skills#run-skills-in-a-subagent). With `skills` in a subagent, the subagent controls the system prompt and loads skill content. With `context: fork` in a skill, the skill content is injected into the agent you specify. Both use the same underlying system.

225</Note>309</Note>

226 310

227## Using subagents effectively311#### Enable persistent memory

228 312

229### Automatic delegation313The `memory` field gives the subagent a persistent directory that survives across conversations. The subagent uses this directory to build up knowledge over time, such as codebase patterns, debugging insights, and architectural decisions.

230 314

231Claude Code proactively delegates tasks based on:315```yaml theme={null}

316---

317name: code-reviewer

318description: Reviews code for quality and best practices

319memory: user

320---

232 321

233* The task description in your request322You are a code reviewer. As you review code, update your agent memory with

234* The `description` field in subagent configurations323patterns, conventions, and recurring issues you discover.

235* Current context and available tools324```

236 325

237<Tip>326Choose a scope based on how broadly the memory should apply:

238 To encourage more proactive subagent use, include phrases like "use PROACTIVELY" or "MUST BE USED" in your `description` field.327

239</Tip>328| Scope | Location | Use when |

329| :-------- | :-------------------------------------------- | :------------------------------------------------------------------------------------------ |

330| `user` | `~/.claude/agent-memory/<name-of-agent>/` | the subagent should remember learnings across all projects |

331| `project` | `.claude/agent-memory/<name-of-agent>/` | the subagent's knowledge is project-specific and shareable via version control |

332| `local` | `.claude/agent-memory-local/<name-of-agent>/` | the subagent's knowledge is project-specific but should not be checked into version control |

333

334When memory is enabled:

335

336* The subagent's system prompt includes instructions for reading and writing to the memory directory.

337* The subagent's system prompt also includes the first 200 lines of `MEMORY.md` in the memory directory, with instructions to curate `MEMORY.md` if it exceeds 200 lines.

338* Read, Write, and Edit tools are automatically enabled so the subagent can manage its memory files.

339

340##### Persistent memory tips

240 341

241### Explicit invocation342* `user` is the recommended default scope. Use `project` or `local` when the subagent's knowledge is only relevant to a specific codebase.

343* Ask the subagent to consult its memory before starting work: "Review this PR, and check your memory for patterns you've seen before."

344* Ask the subagent to update its memory after completing a task: "Now that you're done, save what you learned to your memory." Over time, this builds a knowledge base that makes the subagent more effective.

345* Include memory instructions directly in the subagent's markdown file so it proactively maintains its own knowledge base:

242 346

243Request a specific subagent by mentioning it in your command:347 ```markdown theme={null}

348 Update your agent memory as you discover codepaths, patterns, library

349 locations, and key architectural decisions. This builds up institutional

350 knowledge across conversations. Write concise notes about what you found

351 and where.

352 ```

353

354#### Conditional rules with hooks

355

356For more dynamic control over tool usage, use `PreToolUse` hooks to validate operations before they execute. This is useful when you need to allow some operations of a tool while blocking others.

244 357

358This example creates a subagent that only allows read-only database queries. The `PreToolUse` hook runs the script specified in `command` before each Bash command executes:

359

360```yaml theme={null}

361---

362name: db-reader

363description: Execute read-only database queries

364tools: Bash

365hooks:

366 PreToolUse:

367 - matcher: "Bash"

368 hooks:

369 - type: command

370 command: "./scripts/validate-readonly-query.sh"

371---

245```372```

246> Use the test-runner subagent to fix failing tests373

247> Have the code-reviewer subagent look at my recent changes374Claude Code [passes hook input as JSON](/en/hooks#pretooluse-input) via stdin to hook commands. The validation script reads this JSON, extracts the Bash command, and [exits with code 2](/en/hooks#exit-code-2-behavior-per-event) to block write operations:

248> Ask the debugger subagent to investigate this error375

376```bash theme={null}

377#!/bin/bash

378# ./scripts/validate-readonly-query.sh

379

380INPUT=$(cat)

381COMMAND=$(echo "$INPUT" | jq -r '.tool_input.command // empty')

382

383# Block SQL write operations (case-insensitive)

385 echo "Blocked: Only SELECT queries are allowed" >&2

386 exit 2

387fi

388

389exit 0

249```390```

250 391

251## Built-in subagents392See [Hook input](/en/hooks#pretooluse-input) for the complete input schema and [exit codes](/en/hooks#exit-code-output) for how exit codes affect behavior.

393

394#### Disable specific subagents

395

396You can prevent Claude from using specific subagents by adding them to the `deny` array in your [settings](/en/settings#permission-settings). Use the format `Task(subagent-name)` where `subagent-name` matches the subagent's name field.

397

398```json theme={null}

399{

400 "permissions": {

401 "deny": ["Task(Explore)", "Task(my-custom-agent)"]

402 }

403}

404```

405

406This works for both built-in and custom subagents. You can also use the `--disallowedTools` CLI flag:

407

408```bash theme={null}

409claude --disallowedTools "Task(Explore)"

410```

411

412See [Permissions documentation](/en/permissions#tool-specific-permission-rules) for more details on permission rules.

413

414### Define hooks for subagents

415

416Subagents can define [hooks](/en/hooks) that run during the subagent's lifecycle. There are two ways to configure hooks:

417

4181. **In the subagent's frontmatter**: Define hooks that run only while that subagent is active

4192. **In `settings.json`**: Define hooks that run in the main session when subagents start or stop

420

421#### Hooks in subagent frontmatter

252 422

253Claude Code includes built-in subagents that are available out of the box:423Define hooks directly in the subagent's markdown file. These hooks only run while that specific subagent is active and are cleaned up when it finishes.

254 424

255### General-purpose subagent425All [hook events](/en/hooks#hook-events) are supported. The most common events for subagents are:

256 426

257The general-purpose subagent is a capable agent for complex, multi-step tasks that require both exploration and action. Unlike the Explore subagent, it can modify files and execute a wider range of operations.427| Event | Matcher input | When it fires |

428| :------------ | :------------ | :------------------------------------------------------------------ |

429| `PreToolUse` | Tool name | Before the subagent uses a tool |

430| `PostToolUse` | Tool name | After the subagent uses a tool |

431| `Stop` | (none) | When the subagent finishes (converted to `SubagentStop` at runtime) |

258 432

259**Key characteristics:**433This example validates Bash commands with the `PreToolUse` hook and runs a linter after file edits with `PostToolUse`:

260 434

261* **Model**: Uses Sonnet for more capable reasoning435```yaml theme={null}

262* **Tools**: Has access to all tools436---

263* **Mode**: Can read and write files, execute commands, make changes437name: code-reviewer

264* **Purpose**: Complex research tasks, multi-step operations, code modifications438description: Review code changes with automatic linting

439hooks:

440 PreToolUse:

441 - matcher: "Bash"

442 hooks:

443 - type: command

444 command: "./scripts/validate-command.sh $TOOL_INPUT"

445 PostToolUse:

446 - matcher: "Edit|Write"

447 hooks:

448 - type: command

449 command: "./scripts/run-linter.sh"

450---

451```

452

453`Stop` hooks in frontmatter are automatically converted to `SubagentStop` events.

265 454

266**When Claude uses it:**455#### Project-level hooks for subagent events

267 456

268Claude delegates to the general-purpose subagent when:457Configure hooks in `settings.json` that respond to subagent lifecycle events in the main session.

269 458

270* The task requires both exploration and modification459| Event | Matcher input | When it fires |

271* Complex reasoning is needed to interpret search results460| :-------------- | :-------------- | :------------------------------- |

272* Multiple strategies may be needed if initial searches fail461| `SubagentStart` | Agent type name | When a subagent begins execution |

273* The task has multiple steps that depend on each other462| `SubagentStop` | Agent type name | When a subagent completes |

274 463

275**Example scenario:**464Both events support matchers to target specific agent types by name. This example runs a setup script only when the `db-agent` subagent starts, and a cleanup script when any subagent stops:

276 465

466```json theme={null}

467{

468 "hooks": {

469 "SubagentStart": [

470 {

471 "matcher": "db-agent",

472 "hooks": [

473 { "type": "command", "command": "./scripts/setup-db-connection.sh" }

474 ]

475 }

476 ],

477 "SubagentStop": [

478 {

479 "hooks": [

480 { "type": "command", "command": "./scripts/cleanup-db-connection.sh" }

481 ]

482 }

483 ]

484 }

485}

277```486```

278User: Find all the places where we handle authentication and update them to use the new token format

279 487

280Claude: [Invokes general-purpose subagent]488See [Hooks](/en/hooks) for the complete hook configuration format.

281[Agent searches for auth-related code across codebase]489

282[Agent reads and analyzes multiple files]490## Work with subagents

283[Agent makes necessary edits]491

284[Returns detailed writeup of changes made]492### Understand automatic delegation

493

494Claude automatically delegates tasks based on the task description in your request, the `description` field in subagent configurations, and current context. To encourage proactive delegation, include phrases like "use proactively" in your subagent's description field.

495

496You can also request a specific subagent explicitly:

497

285```498```

499Use the test-runner subagent to fix failing tests

500Have the code-reviewer subagent look at my recent changes

501```

502

503### Run subagents in foreground or background

504

505Subagents can run in the foreground (blocking) or background (concurrent):

286 506

287### Plan subagent507* **Foreground subagents** block the main conversation until complete. Permission prompts and clarifying questions (like [`AskUserQuestion`](/en/settings#tools-available-to-claude)) are passed through to you.

508* **Background subagents** run concurrently while you continue working. Before launching, Claude Code prompts for any tool permissions the subagent will need, ensuring it has the necessary approvals upfront. Once running, the subagent inherits these permissions and auto-denies anything not pre-approved. If a background subagent needs to ask clarifying questions, that tool call fails but the subagent continues. MCP tools are not available in background subagents.

288 509

289The Plan subagent is a specialized built-in agent designed for use during plan mode. When Claude is operating in plan mode (non-execution mode), it uses the Plan subagent to conduct research and gather information about your codebase before presenting a plan.510If a background subagent fails due to missing permissions, you can [resume it](#resume-subagents) in the foreground to retry with interactive prompts.

290 511

291**Key characteristics:**512Claude decides whether to run subagents in the foreground or background based on the task. You can also:

292 513

293* **Model**: Uses Sonnet for more capable analysis514* Ask Claude to "run this in the background"

294* **Tools**: Has access to Read, Glob, Grep, and Bash tools for codebase exploration515* Press **Ctrl+B** to background a running task

295* **Purpose**: Searches files, analyzes code structure, and gathers context

296* **Automatic invocation**: Claude automatically uses this agent when in plan mode and needs to research the codebase

297 516

298**How it works:**517To disable all background task functionality, set the `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to `1`. See [Environment variables](/en/settings#environment-variables).

299When you're in plan mode and Claude needs to understand your codebase to create a plan, it delegates research tasks to the Plan subagent. This prevents infinite nesting of agents (subagents cannot spawn other subagents) while still allowing Claude to gather the necessary context.

300 518

301**Example scenario:**519### Common patterns

302 520

521#### Isolate high-volume operations

522

523One of the most effective uses for subagents is isolating operations that produce large amounts of output. Running tests, fetching documentation, or processing log files can consume significant context. By delegating these to a subagent, the verbose output stays in the subagent's context while only the relevant summary returns to your main conversation.

524

525```

526Use a subagent to run the test suite and report only the failing tests with their error messages

303```527```

304User: [In plan mode] Help me refactor the authentication module

305 528

306Claude: Let me research your authentication implementation first...529#### Run parallel research

307[Internally invokes Plan subagent to explore auth-related files]530

308[Plan subagent searches codebase and returns findings]531For independent investigations, spawn multiple subagents to work simultaneously:

309Claude: Based on my research, here's my proposed plan...532

533```

534Research the authentication, database, and API modules in parallel using separate subagents

310```535```

311 536

312<Tip>537Each subagent explores its area independently, then Claude synthesizes the findings. This works best when the research paths don't depend on each other.

313 The Plan subagent is only used in plan mode. In normal execution mode, Claude uses the general-purpose agent or other custom subagents you've created.

314</Tip>

315 538

316### Explore subagent539<Warning>

540 When subagents complete, their results return to your main conversation. Running many subagents that each return detailed results can consume significant context.

541</Warning>

317 542

318The Explore subagent is a fast, lightweight agent optimized for searching and analyzing codebases. It operates in strict read-only mode and is designed for rapid file discovery and code exploration.543For tasks that need sustained parallelism or exceed your context window, [agent teams](/en/agent-teams) give each worker its own independent context.

319 544

320**Key characteristics:**545#### Chain subagents

321 546

322* **Model**: Uses Haiku for fast, low-latency searches547For multi-step workflows, ask Claude to use subagents in sequence. Each subagent completes its task and returns results to Claude, which then passes relevant context to the next subagent.

323* **Mode**: Strictly read-only - cannot create, modify, or delete files

324* **Tools available**:

325 * Glob - File pattern matching

326 * Grep - Content searching with regular expressions

327 * Read - Reading file contents

328 * Bash - Read-only commands only (ls, git status, git log, git diff, find, cat, head, tail)

329 548

330**When Claude uses it:**549```

550Use the code-reviewer subagent to find performance issues, then use the optimizer subagent to fix them

551```

331 552

332Claude will delegate to the Explore subagent when it needs to search or understand a codebase but doesn't need to make changes. This is more efficient than the main agent running multiple search commands directly, as content found during the exploration process doesn't bloat the main conversation.553### Choose between subagents and main conversation

333 554

334**Thoroughness levels:**555Use the **main conversation** when:

335 556

336When invoking the Explore subagent, Claude specifies a thoroughness level:557* The task needs frequent back-and-forth or iterative refinement

558* Multiple phases share significant context (planning → implementation → testing)

559* You're making a quick, targeted change

560* Latency matters. Subagents start fresh and may need time to gather context

337 561

338* **Quick** - Fast searches with minimal exploration. Good for targeted lookups.562Use **subagents** when:

339* **Medium** - Moderate exploration. Balances speed and thoroughness.

340* **Very thorough** - Comprehensive analysis across multiple locations and naming conventions. Used when the target might be in unexpected places.

341 563

342**Example scenarios:**564* The task produces verbose output you don't need in your main context

565* You want to enforce specific tool restrictions or permissions

566* The work is self-contained and can return a summary

343 567

344```568Consider [Skills](/en/skills) instead when you want reusable prompts or workflows that run in the main conversation context rather than isolated subagent context.

345User: Where are errors from the client handled?569

570<Note>

571 Subagents cannot spawn other subagents. If your workflow requires nested delegation, use [Skills](/en/skills) or [chain subagents](#chain-subagents) from the main conversation.

572</Note>

573

574### Manage subagent context

575

576#### Resume subagents

577

578Each subagent invocation creates a new instance with fresh context. To continue an existing subagent's work instead of starting over, ask Claude to resume it.

579

580Resumed subagents retain their full conversation history, including all previous tool calls, results, and reasoning. The subagent picks up exactly where it stopped rather than starting fresh.

581

582When a subagent completes, Claude receives its agent ID. To resume a subagent, ask Claude to continue the previous work:

346 583

347Claude: [Invokes Explore subagent with "medium" thoroughness]

348[Explore uses Grep to search for error handling patterns]

349[Explore uses Read to examine promising files]

350[Returns findings with absolute file paths]

351Claude: Client errors are handled in src/services/process.ts:712...

352```584```

585Use the code-reviewer subagent to review the authentication module

586[Agent completes]

353 587

588Continue that code review and now analyze the authorization logic

589[Claude resumes the subagent with full context from previous conversation]

354```590```

355User: What's the codebase structure?

356 591

357Claude: [Invokes Explore subagent with "quick" thoroughness]592You can also ask Claude for the agent ID if you want to reference it explicitly, or find IDs in the transcript files at `~/.claude/projects/{project}/{sessionId}/subagents/`. Each transcript is stored as `agent-{agentId}.jsonl`.

358[Explore uses Glob and ls to map directory structure]593

359[Returns overview of key directories and their purposes]594Subagent transcripts persist independently of the main conversation:

595

596* **Main conversation compaction**: When the main conversation compacts, subagent transcripts are unaffected. They're stored in separate files.

597* **Session persistence**: Subagent transcripts persist within their session. You can [resume a subagent](#resume-subagents) after restarting Claude Code by resuming the same session.

598* **Automatic cleanup**: Transcripts are cleaned up based on the `cleanupPeriodDays` setting (default: 30 days).

599

600#### Auto-compaction

601

602Subagents support automatic compaction using the same logic as the main conversation. By default, auto-compaction triggers at approximately 95% capacity. To trigger compaction earlier, set `CLAUDE_AUTOCOMPACT_PCT_OVERRIDE` to a lower percentage (for example, `50`). See [environment variables](/en/settings#environment-variables) for details.

603

604Compaction events are logged in subagent transcript files:

605

606```json theme={null}

607{

608 "type": "system",

609 "subtype": "compact_boundary",

610 "compactMetadata": {

611 "trigger": "auto",

612 "preTokens": 167189

613 }

614}

360```615```

361 616

617The `preTokens` value shows how many tokens were used before compaction occurred.

618

362## Example subagents619## Example subagents

363 620

621These examples demonstrate effective patterns for building subagents. Use them as starting points, or generate a customized version with Claude.

622

623<Tip>

624 **Best practices:**

625

626 * **Design focused subagents:** each subagent should excel at one specific task

627 * **Write detailed descriptions:** Claude uses the description to decide when to delegate

628 * **Limit tool access:** grant only necessary permissions for security and focus

629 * **Check into version control:** share project subagents with your team

630</Tip>

631

364### Code reviewer632### Code reviewer

365 633

634A read-only subagent that reviews code without modifying it. This example shows how to design a focused subagent with limited tool access (no Edit or Write) and a detailed prompt that specifies exactly what to look for and how to format output.

635

366```markdown theme={null}636```markdown theme={null}

367---637---

368name: code-reviewer638name: code-reviewer

399 668

400### Debugger669### Debugger

401 670

671A subagent that can both analyze and fix issues. Unlike the code reviewer, this one includes Edit because fixing bugs requires modifying code. The prompt provides a clear workflow from diagnosis to verification.

672

402```markdown theme={null}673```markdown theme={null}

403---674---

404name: debugger675name: debugger

434 705

435### Data scientist706### Data scientist

436 707

708A domain-specific subagent for data analysis work. This example shows how to create subagents for specialized workflows outside of typical coding tasks. It explicitly sets `model: sonnet` for more capable analysis.

709

437```markdown theme={null}710```markdown theme={null}

438---711---

439name: data-scientist712name: data-scientist

467Always ensure queries are efficient and cost-effective.740Always ensure queries are efficient and cost-effective.

468```741```

469 742

470## Best practices743### Database query validator

~~471~~

472* **Start with Claude-generated agents**: We highly recommend generating your initial subagent with Claude and then iterating on it to make it personally yours. This approach gives you the best results - a solid foundation that you can customize to your specific needs.

~~473~~

474* **Design focused subagents**: Create subagents with single, clear responsibilities rather than trying to make one subagent do everything. This improves performance and makes subagents more predictable.

~~475~~

476* **Write detailed prompts**: Include specific instructions, examples, and constraints in your system prompts. The more guidance you provide, the better the subagent will perform.

~~477~~

478* **Limit tool access**: Only grant tools that are necessary for the subagent's purpose. This improves security and helps the subagent focus on relevant actions.

479 744

480* **Version control**: Check project subagents into version control so your team can benefit from and improve them collaboratively.745A subagent that allows Bash access but validates commands to permit only read-only SQL queries. This example shows how to use `PreToolUse` hooks for conditional validation when you need finer control than the `tools` field provides.

481 746

482## Advanced usage747```markdown theme={null}

748---

749name: db-reader

750description: Execute read-only database queries. Use when analyzing data or generating reports.

751tools: Bash

752hooks:

753 PreToolUse:

754 - matcher: "Bash"

755 hooks:

756 - type: command

757 command: "./scripts/validate-readonly-query.sh"

758---

483 759

484### Chaining subagents760You are a database analyst with read-only access. Execute SELECT queries to answer questions about the data.

485 761

486For complex workflows, you can chain multiple subagents:762When asked to analyze data:

7631. Identify which tables contain the relevant data

7642. Write efficient SELECT queries with appropriate filters

7653. Present results clearly with context

487 766

767You cannot modify data. If asked to INSERT, UPDATE, DELETE, or modify schema, explain that you only have read access.

488```768```

489> First use the code-analyzer subagent to find performance issues, then use the optimizer subagent to fix them

490```

~~491~~

492### Dynamic subagent selection

493 769

494Claude Code intelligently selects subagents based on context. Make your `description` fields specific and action-oriented for best results.770Claude Code [passes hook input as JSON](/en/hooks#pretooluse-input) via stdin to hook commands. The validation script reads this JSON, extracts the command being executed, and checks it against a list of SQL write operations. If a write operation is detected, the script [exits with code 2](/en/hooks#exit-code-2-behavior-per-event) to block execution and returns an error message to Claude via stderr.

495 771

496### Resumable subagents772Create the validation script anywhere in your project. The path must match the `command` field in your hook configuration:

497 773

498Subagents can be resumed to continue previous conversations, which is particularly useful for long-running research or analysis tasks that need to be continued across multiple invocations.774```bash theme={null}

~~499~~ 775#!/bin/bash

500**How it works:**776# Blocks SQL write operations, allows SELECT queries

~~501~~

502* Each subagent execution is assigned a unique `agentId`

503* The agent's conversation is stored in a separate transcript file: `agent-{agentId}.jsonl`

504* You can resume a previous agent by providing its `agentId` via the `resume` parameter

505* When resumed, the agent continues with full context from its previous conversation

~~506~~

507**Example workflow:**

~~508~~

509Initial invocation:

510 777

511```778# Read JSON input from stdin

512> Use the code-analyzer agent to start reviewing the authentication module779INPUT=$(cat)

513 780

514[Agent completes initial analysis and returns agentId: "abc123"]781# Extract the command field from tool_input using jq

515```782COMMAND=$(echo "$INPUT" | jq -r '.tool_input.command // empty')

516 783

517Resume the agent:784if [ -z "$COMMAND" ]; then

785 exit 0

786fi

518 787

519```788# Block write operations (case-insensitive)

790 echo "Blocked: Write operations not allowed. Use SELECT queries only." >&2

791 exit 2

792fi

521 793

522[Agent continues with full context from previous conversation]794exit 0

523```795```

524 796

525**Use cases:**797Make the script executable:

~~526~~

527* **Long-running research**: Break down large codebase analysis into multiple sessions

528* **Iterative refinement**: Continue refining a subagent's work without losing context

529* **Multi-step workflows**: Have a subagent work on related tasks sequentially while maintaining context

~~530~~

531**Technical details:**

~~532~~

533* Agent transcripts are stored in your project directory

534* Recording is disabled during resume to avoid duplicating messages

535* Both synchronous and asynchronous agents can be resumed

536* The `resume` parameter accepts the agent ID from a previous execution

~~537~~

538**Programmatic usage:**

539 798

540If you're using the Agent SDK or interacting with the AgentTool directly, you can pass the `resume` parameter:799```bash theme={null}

~~541~~ 800chmod +x ./scripts/validate-readonly-query.sh

542```typescript theme={null}

543{

544 "description": "Continue analysis",

545 "prompt": "Now examine the error handling patterns",

546 "subagent_type": "code-analyzer",

547 "resume": "abc123" // Agent ID from previous execution

548}

549```801```

550 802

551<Tip>803The hook receives JSON via stdin with the Bash command in `tool_input.command`. Exit code 2 blocks the operation and feeds the error message back to Claude. See [Hooks](/en/hooks#exit-code-output) for details on exit codes and [Hook input](/en/hooks#pretooluse-input) for the complete input schema.

552 Keep track of agent IDs for tasks you may want to resume later. Claude Code displays the agent ID when a subagent completes its work.

553</Tip>

554 804

555## Performance considerations805## Next steps

556 806

557* **Context efficiency**: Agents help preserve main context, enabling longer overall sessions807Now that you understand subagents, explore these related features:

558* **Latency**: Subagents start off with a clean slate each time they are invoked and may add latency as they gather context that they require to do their job effectively.

~~559~~

560## Related documentation

~~561~~

562* [Plugins](/en/plugins) - Extend Claude Code with custom agents through plugins

563* [Slash commands](/en/slash-commands) - Learn about other built-in commands

564* [Settings](/en/settings) - Configure Claude Code behavior

565* [Hooks](/en/hooks) - Automate workflows with event handlers

~~566~~

~~567~~

568 808

569> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt809* [Distribute subagents with plugins](/en/plugins) to share subagents across teams or projects

810* [Run Claude Code programmatically](/en/headless) with the Agent SDK for CI/CD and automation

811* [Use MCP servers](/en/mcp) to give subagents access to external tools and data

terminal-config.md +19 −8

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Optimize your terminal setup5# Optimize your terminal setup

2 6

3> Claude Code works best when your terminal is properly configured. Follow these guidelines to optimize your experience.7> Claude Code works best when your terminal is properly configured. Follow these guidelines to optimize your experience.

13You have several options for entering line breaks into Claude Code:17You have several options for entering line breaks into Claude Code:

14 18

15* **Quick escape**: Type `\` followed by Enter to create a newline19* **Quick escape**: Type `\` followed by Enter to create a newline

~~16* **Keyboard shortcut**: Set up a keybinding to insert a newline~~20* **Shift+Enter**: Works out of the box in iTerm2, WezTerm, Ghostty, and Kitty

21* **Keyboard shortcut**: Set up a keybinding to insert a newline in other terminals

23**Set up Shift+Enter for other terminals**

17 24

~~18#### Set up Shift+Enter (VS Code or iTerm2):~~25Run `/terminal-setup` within Claude Code to automatically configure Shift+Enter for VS Code, Alacritty, Zed, and Warp.

19 26

~~20Run `/terminal-setup` within Claude Code to automatically configure Shift+Enter.~~27<Note>

28 The `/terminal-setup` command is only visible in terminals that require manual configuration. If you're using iTerm2, WezTerm, Ghostty, or Kitty, you won't see this command because Shift+Enter already works natively.

29</Note>

21 30

~~22#### Set up Option+Enter (VS Code, iTerm2 or macOS Terminal.app):~~31**Set up Option+Enter (VS Code, iTerm2 or macOS Terminal.app)**

23 32

24**For Mac Terminal.app:**33**For Mac Terminal.app:**

25 34

65The supported subset includes:74The supported subset includes:

66 75

67* Mode switching: `Esc` (to NORMAL), `i`/`I`, `a`/`A`, `o`/`O` (to INSERT)76* Mode switching: `Esc` (to NORMAL), `i`/`I`, `a`/`A`, `o`/`O` (to INSERT)

~~68* Navigation: `h`/`j`/`k`/`l`, `w`/`e`/`b`, `0`/`$`/`^`, `gg`/`G`~~77* Navigation: `h`/`j`/`k`/`l`, `w`/`e`/`b`, `0`/`$`/`^`, `gg`/`G`, `f`/`F`/`t`/`T` with `;`/`,` repeat

69* Editing: `x`, `dw`/`de`/`db`/`dd`/`D`, `cw`/`ce`/`cb`/`cc`/`C`, `.` (repeat)78* Editing: `x`, `dw`/`de`/`db`/`dd`/`D`, `cw`/`ce`/`cb`/`cc`/`C`, `.` (repeat)

79* Yank/paste: `yy`/`Y`, `yw`/`ye`/`yb`, `p`/`P`

80* Text objects: `iw`/`aw`, `iW`/`aW`, `i"`/`a"`, `i'`/`a'`, `i(`/`a(`, `i[`/`a[`, `i{`/`a{`

81* Indentation: `>>`/`<<`

82* Line operations: `J` (join lines)

70 83

71 84See [Interactive mode](/en/interactive-mode#vim-editor-mode) for the complete reference.

~~73> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt~~

third-party-integrations.md +147 −153

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Enterprise deployment overview5# Enterprise deployment overview

2 6

3> Learn how Claude Code can integrate with various third-party services and infrastructure to meet enterprise deployment requirements.7> Learn how Claude Code can integrate with various third-party services and infrastructure to meet enterprise deployment requirements.

4 8

~~5This page provides an overview of available deployment options and helps you choose the right configuration for your organization.~~9Organizations can deploy Claude Code through Anthropic directly or through a cloud provider. This page helps you choose the right configuration.

11## Compare deployment options

13For most organizations, Claude for Teams or Claude for Enterprise provides the best experience. Team members get access to both Claude Code and Claude on the web with a single subscription, centralized billing, and no infrastructure setup required.

15**Claude for Teams** is self-service and includes collaboration features, admin tools, and billing management. Best for smaller teams that need to get started quickly.

17**Claude for Enterprise** adds SSO and domain capture, role-based permissions, compliance API access, and managed policy settings for deploying organization-wide Claude Code configurations. Best for larger organizations with security and compliance requirements.

19Learn more about [Team plans](https://support.claude.com/en/articles/9266767-what-is-the-team-plan) and [Enterprise plans](https://support.claude.com/en/articles/9797531-what-is-the-enterprise-plan).

6 20

~~7## Provider comparison~~21If your organization has specific infrastructure requirements, compare the options below:

8 22

9<table>23<table>

10 <thead>24 <thead>

11 <tr>25 <tr>

12 <th>Feature</th>26 <th>Feature</th>

~~13 <th>Anthropic</th>~~27 <th>Claude for Teams/Enterprise</th>

28 <th>Anthropic Console</th>

14 <th>Amazon Bedrock</th>29 <th>Amazon Bedrock</th>

15 <th>Google Vertex AI</th>30 <th>Google Vertex AI</th>

16 <th>Microsoft Foundry</th>31 <th>Microsoft Foundry</th>

18 </thead>33 </thead>

19 34

20 <tbody>35 <tbody>

36 <tr>

37 <td>Best for</td>

38 <td>Most organizations (recommended)</td>

39 <td>Individual developers</td>

40 <td>AWS-native deployments</td>

41 <td>GCP-native deployments</td>

42 <td>Azure-native deployments</td>

43 </tr>

45 <tr>

46 <td>Billing</td>

47 <td><strong>Teams:</strong> \$150/seat (Premium) with PAYG available<br /><strong>Enterprise:</strong> <a href="https://claude.com/contact-sales">Contact Sales</a></td>

48 <td>PAYG</td>

49 <td>PAYG through AWS</td>

50 <td>PAYG through GCP</td>

51 <td>PAYG through Azure</td>

52 </tr>

21 <tr>54 <tr>

22 <td>Regions</td>55 <td>Regions</td>

23 <td>Supported [countries](https://www.anthropic.com/supported-countries)</td>56 <td>Supported [countries](https://www.anthropic.com/supported-countries)</td>

57 <td>Supported [countries](https://www.anthropic.com/supported-countries)</td>

24 <td>Multiple AWS [regions](https://docs.aws.amazon.com/bedrock/latest/userguide/models-regions.html)</td>58 <td>Multiple AWS [regions](https://docs.aws.amazon.com/bedrock/latest/userguide/models-regions.html)</td>

25 <td>Multiple GCP [regions](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/locations)</td>59 <td>Multiple GCP [regions](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/locations)</td>

26 <td>Multiple Azure [regions](https://azure.microsoft.com/en-us/explore/global-infrastructure/products-by-region/)</td>60 <td>Multiple Azure [regions](https://azure.microsoft.com/en-us/explore/global-infrastructure/products-by-region/)</td>

32 <td>Enabled by default</td>66 <td>Enabled by default</td>

33 <td>Enabled by default</td>67 <td>Enabled by default</td>

34 <td>Enabled by default</td>68 <td>Enabled by default</td>

69 <td>Enabled by default</td>

35 </tr>70 </tr>

36 71

37 <tr>72 <tr>

38 <td>Authentication</td>73 <td>Authentication</td>

74 <td>Claude.ai SSO or email</td>

39 <td>API key</td>75 <td>API key</td>

40 <td>API key or AWS credentials</td>76 <td>API key or AWS credentials</td>

41 <td>GCP credentials</td>77 <td>GCP credentials</td>

44 80

45 <tr>81 <tr>

46 <td>Cost tracking</td>82 <td>Cost tracking</td>

~~47 <td>Dashboard</td>~~83 <td>Usage dashboard</td>

84 <td>Usage dashboard</td>

48 <td>AWS Cost Explorer</td>85 <td>AWS Cost Explorer</td>

49 <td>GCP Billing</td>86 <td>GCP Billing</td>

50 <td>Azure Cost Management</td>87 <td>Azure Cost Management</td>

51 </tr>88 </tr>

52 89

90 <tr>

91 <td>Includes Claude on web</td>

92 <td>Yes</td>

93 <td>No</td>

94 <td>No</td>

95 <td>No</td>

96 <td>No</td>

97 </tr>

53 <tr>99 <tr>

54 <td>Enterprise features</td>100 <td>Enterprise features</td>

~~55 <td>Teams, usage monitoring</td>~~101 <td>Team management, SSO, usage monitoring</td>

102 <td>None</td>

56 <td>IAM policies, CloudTrail</td>103 <td>IAM policies, CloudTrail</td>

57 <td>IAM roles, Cloud Audit Logs</td>104 <td>IAM roles, Cloud Audit Logs</td>

58 <td>RBAC policies, Azure Monitor</td>105 <td>RBAC policies, Azure Monitor</td>

60 </tbody>107 </tbody>

61</table>108</table>

62 109

~~63## Cloud providers~~110Select a deployment option to view setup instructions:

~~65<CardGroup cols={3}>~~

~~66 <Card title="Amazon Bedrock" icon="aws" href="/en/amazon-bedrock">~~

~~67 Use Claude models through AWS infrastructure with API key or IAM-based authentication and AWS-native monitoring~~

~~68 </Card>~~

~~70 <Card title="Google Vertex AI" icon="google" href="/en/google-vertex-ai">~~

~~71 Access Claude models via Google Cloud Platform with enterprise-grade security and compliance~~

~~72 </Card>~~

~~74 <Card title="Microsoft Foundry" icon="microsoft" href="/en/microsoft-foundry">~~

~~75 Access Claude through Azure with API key or Microsoft Entra ID authentication and Azure billing~~

~~76 </Card>~~

~~77</CardGroup>~~

~~79## Corporate infrastructure~~

~~81<CardGroup cols={2}>~~

~~82 <Card title="Enterprise Network" icon="shield" href="/en/network-config">~~

~~83 Configure Claude Code to work with your organization's proxy servers and SSL/TLS requirements~~

~~84 </Card>~~

85 111

~~86 <Card title="LLM Gateway" icon="server" href="/en/llm-gateway">~~112* [Claude for Teams or Enterprise](/en/authentication#claude-for-teams-or-enterprise)

~~87 Deploy centralized model access with usage tracking, budgeting, and audit logging~~113* [Anthropic Console](/en/authentication#claude-console-authentication)

~~88 </Card>~~114* [Amazon Bedrock](/en/amazon-bedrock)

~~89</CardGroup>~~115* [Google Vertex AI](/en/google-vertex-ai)

116* [Microsoft Foundry](/en/microsoft-foundry)

90 117

~~91## Configuration overview~~118## Configure proxies and gateways

92 119

~~93Claude Code supports flexible configuration options that allow you to combine different providers and infrastructure:~~120Most organizations can use a cloud provider directly without additional configuration. However, you may need to configure a corporate proxy or LLM gateway if your organization has specific network or management requirements. These are different configurations that can be used together:

94 121

~~95<Note>~~122* **Corporate proxy**: Routes traffic through an HTTP/HTTPS proxy. Use this if your organization requires all outbound traffic to pass through a proxy server for security monitoring, compliance, or network policy enforcement. Configure with the `HTTPS_PROXY` or `HTTP_PROXY` environment variables. Learn more in [Enterprise network configuration](/en/network-config).

~~96 Understand the difference between:~~123* **LLM Gateway**: A service that sits between Claude Code and the cloud provider to handle authentication and routing. Use this if you need centralized usage tracking across teams, custom rate limiting or budgets, or centralized authentication management. Configure with the `ANTHROPIC_BASE_URL`, `ANTHROPIC_BEDROCK_BASE_URL`, or `ANTHROPIC_VERTEX_BASE_URL` environment variables. Learn more in [LLM gateway configuration](/en/llm-gateway).

97 124

~~98 * **Corporate proxy**: An HTTP/HTTPS proxy for routing traffic (set via `HTTPS_PROXY` or `HTTP_PROXY`)~~125The following examples show the environment variables to set in your shell or shell profile (`.bashrc`, `.zshrc`). See [Settings](/en/settings) for other configuration methods.

~~99 * **LLM Gateway**: A service that handles authentication and provides provider-compatible endpoints (set via `ANTHROPIC_BASE_URL`, `ANTHROPIC_BEDROCK_BASE_URL`, or `ANTHROPIC_VERTEX_BASE_URL`)~~

100 126

101 Both configurations can be used in tandem.127### Amazon Bedrock

102</Note>

103 128

104### Using Bedrock with corporate proxy129<Tabs>

130 <Tab title="Corporate proxy">

131 Route Bedrock traffic through your corporate proxy by setting the following [environment variables](/en/settings#environment-variables):

105 132

106Route Bedrock traffic through a corporate HTTP/HTTPS proxy:133 ```bash theme={null}

134 # Enable Bedrock

135 export CLAUDE_CODE_USE_BEDROCK=1

136 export AWS_REGION=us-east-1

107 137

108```bash theme={null}138 # Configure corporate proxy

109# Enable Bedrock139 export HTTPS_PROXY='https://proxy.example.com:8080'

110export CLAUDE_CODE_USE_BEDROCK=1140 ```

111export AWS_REGION=us-east-1141 </Tab>

112 142

113# Configure corporate proxy143 <Tab title="LLM Gateway">

114export HTTPS_PROXY='https://proxy.example.com:8080'144 Route Bedrock traffic through your LLM gateway by setting the following [environment variables](/en/settings#environment-variables):

115```

116 145

117### Using Bedrock with LLM Gateway146 ```bash theme={null}

147 # Enable Bedrock

148 export CLAUDE_CODE_USE_BEDROCK=1

118 149

119Use a gateway service that provides Bedrock-compatible endpoints:150 # Configure LLM gateway

151 export ANTHROPIC_BEDROCK_BASE_URL='https://your-llm-gateway.com/bedrock'

152 export CLAUDE_CODE_SKIP_BEDROCK_AUTH=1 # If gateway handles AWS auth

153 ```

154 </Tab>

155</Tabs>

120 156

121```bash theme={null}157### Microsoft Foundry

122# Enable Bedrock

123export CLAUDE_CODE_USE_BEDROCK=1

124 158

125# Configure LLM gateway159<Tabs>

126export ANTHROPIC_BEDROCK_BASE_URL='https://your-llm-gateway.com/bedrock'160 <Tab title="Corporate proxy">

127export CLAUDE_CODE_SKIP_BEDROCK_AUTH=1 # If gateway handles AWS auth161 Route Foundry traffic through your corporate proxy by setting the following [environment variables](/en/settings#environment-variables):

128```

129 162

130### Using Foundry with corporate proxy163 ```bash theme={null}

164 # Enable Microsoft Foundry

165 export CLAUDE_CODE_USE_FOUNDRY=1

166 export ANTHROPIC_FOUNDRY_RESOURCE=your-resource

167 export ANTHROPIC_FOUNDRY_API_KEY=your-api-key # Or omit for Entra ID auth

131 168

132Route Azure traffic through a corporate HTTP/HTTPS proxy:169 # Configure corporate proxy

170 export HTTPS_PROXY='https://proxy.example.com:8080'

171 ```

172 </Tab>

133 173

134```bash theme={null}174 <Tab title="LLM Gateway">

135# Enable Microsoft Foundry175 Route Foundry traffic through your LLM gateway by setting the following [environment variables](/en/settings#environment-variables):

136export CLAUDE_CODE_USE_FOUNDRY=1

137export ANTHROPIC_FOUNDRY_RESOURCE=your-resource

138export ANTHROPIC_FOUNDRY_API_KEY=your-api-key # Or omit for Entra ID auth

139 176

140# Configure corporate proxy177 ```bash theme={null}

141export HTTPS_PROXY='https://proxy.example.com:8080'178 # Enable Microsoft Foundry

142```179 export CLAUDE_CODE_USE_FOUNDRY=1

143 180

144### Using Foundry with LLM Gateway181 # Configure LLM gateway

182 export ANTHROPIC_FOUNDRY_BASE_URL='https://your-llm-gateway.com'

183 export CLAUDE_CODE_SKIP_FOUNDRY_AUTH=1 # If gateway handles Azure auth

184 ```

185 </Tab>

186</Tabs>

145 187

146Use a gateway service that provides Azure-compatible endpoints:188### Google Vertex AI

147 189

148```bash theme={null}190<Tabs>

149# Enable Microsoft Foundry191 <Tab title="Corporate proxy">

150export CLAUDE_CODE_USE_FOUNDRY=1192 Route Vertex AI traffic through your corporate proxy by setting the following [environment variables](/en/settings#environment-variables):

151 193

152# Configure LLM gateway194 ```bash theme={null}

153export ANTHROPIC_FOUNDRY_BASE_URL='https://your-llm-gateway.com'195 # Enable Vertex

154export CLAUDE_CODE_SKIP_FOUNDRY_AUTH=1 # If gateway handles Azure auth196 export CLAUDE_CODE_USE_VERTEX=1

155```197 export CLOUD_ML_REGION=us-east5

198 export ANTHROPIC_VERTEX_PROJECT_ID=your-project-id

156 199

157### Using Vertex AI with corporate proxy200 # Configure corporate proxy

201 export HTTPS_PROXY='https://proxy.example.com:8080'

202 ```

203 </Tab>

158 204

159Route Vertex AI traffic through a corporate HTTP/HTTPS proxy:205 <Tab title="LLM Gateway">

206 Route Vertex AI traffic through your LLM gateway by setting the following [environment variables](/en/settings#environment-variables):

160 207

161```bash theme={null}208 ```bash theme={null}

162# Enable Vertex209 # Enable Vertex

163export CLAUDE_CODE_USE_VERTEX=1210 export CLAUDE_CODE_USE_VERTEX=1

164export CLOUD_ML_REGION=us-east5

165export ANTHROPIC_VERTEX_PROJECT_ID=your-project-id

166 211

167# Configure corporate proxy212 # Configure LLM gateway

168export HTTPS_PROXY='https://proxy.example.com:8080'213 export ANTHROPIC_VERTEX_BASE_URL='https://your-llm-gateway.com/vertex'

169```214 export CLAUDE_CODE_SKIP_VERTEX_AUTH=1 # If gateway handles GCP auth

215 ```

216 </Tab>

217</Tabs>

170 218

171### Using Vertex AI with LLM Gateway219<Tip>

~~172~~ 220 Use `/status` in Claude Code to verify your proxy and gateway configuration is applied correctly.

173Combine Google Vertex AI models with an LLM gateway for centralized management:221</Tip>

~~174~~

175```bash theme={null}

176# Enable Vertex

177export CLAUDE_CODE_USE_VERTEX=1

~~178~~

179# Configure LLM gateway

180export ANTHROPIC_VERTEX_BASE_URL='https://your-llm-gateway.com/vertex'

181export CLAUDE_CODE_SKIP_VERTEX_AUTH=1 # If gateway handles GCP auth

182```

~~183~~

184### Authentication configuration

~~185~~

186Claude Code uses the `ANTHROPIC_AUTH_TOKEN` for the `Authorization` header when needed. The `SKIP_AUTH` flags (`CLAUDE_CODE_SKIP_BEDROCK_AUTH`, `CLAUDE_CODE_SKIP_VERTEX_AUTH`) are used in LLM gateway scenarios where the gateway handles provider authentication.

~~187~~

188## Choosing the right deployment configuration

~~189~~

190Consider these factors when selecting your deployment approach:

~~191~~

192### Direct provider access

~~193~~

194Best for organizations that:

~~195~~

196* Want the simplest setup

197* Have existing AWS or GCP infrastructure

198* Need provider-native monitoring and compliance

~~199~~

200### Corporate proxy

~~201~~

202Best for organizations that:

~~203~~

204* Have existing corporate proxy requirements

205* Need traffic monitoring and compliance

206* Must route all traffic through specific network paths

~~207~~

208### LLM Gateway

~~209~~

210Best for organizations that:

~~211~~

212* Need usage tracking across teams

213* Want to dynamically switch between models

214* Require custom rate limiting or budgets

215* Need centralized authentication management

~~216~~

217## Debugging

~~218~~

219When debugging your deployment:

~~220~~

221* Use the `claude /status` [slash command](/en/slash-commands). This command provides observability into any applied authentication, proxy, and URL settings.

222* Set environment variable `export ANTHROPIC_LOG=debug` to log requests.

223 222

224## Best practices for organizations223## Best practices for organizations

225 224

226### 1. Invest in documentation and memory225### Invest in documentation and memory

227 226

228We strongly recommend investing in documentation so that Claude Code understands your codebase. Organizations can deploy CLAUDE.md files at multiple levels:227We strongly recommend investing in documentation so that Claude Code understands your codebase. Organizations can deploy CLAUDE.md files at multiple levels:

229 228

230* **Organization-wide**: Deploy to system directories like `/Library/Application Support/ClaudeCode/CLAUDE.md` (macOS) for company-wide standards229* **Organization-wide**: Deploy to system directories like `/Library/Application Support/ClaudeCode/CLAUDE.md` (macOS) for company-wide standards

231* **Repository-level**: Create `CLAUDE.md` files in repository roots containing project architecture, build commands, and contribution guidelines. Check these into source control so all users benefit230* **Repository-level**: Create `CLAUDE.md` files in repository roots containing project architecture, build commands, and contribution guidelines. Check these into source control so all users benefit

232 231

233 [Learn more](/en/memory).232Learn more in [Memory and CLAUDE.md files](/en/memory).

234 233

235### 2. Simplify deployment234### Simplify deployment

236 235

237If you have a custom development environment, we find that creating a "one click" way to install Claude Code is key to growing adoption across an organization.236If you have a custom development environment, we find that creating a "one click" way to install Claude Code is key to growing adoption across an organization.

238 237

239### 3. Start with guided usage238### Start with guided usage

240 239

241Encourage new users to try Claude Code for codebase Q\&A, or on smaller bug fixes or feature requests. Ask Claude Code to make a plan. Check Claude's suggestions and give feedback if it's off-track. Over time, as users understand this new paradigm better, then they'll be more effective at letting Claude Code run more agentically.240Encourage new users to try Claude Code for codebase Q\&A, or on smaller bug fixes or feature requests. Ask Claude Code to make a plan. Check Claude's suggestions and give feedback if it's off-track. Over time, as users understand this new paradigm better, then they'll be more effective at letting Claude Code run more agentically.

242 241

243### 4. Configure security policies242### Configure security policies

244 243

245Security teams can configure managed permissions for what Claude Code is and is not allowed to do, which cannot be overwritten by local configuration. [Learn more](/en/security).244Security teams can configure managed permissions for what Claude Code is and is not allowed to do, which cannot be overwritten by local configuration. [Learn more](/en/security).

246 245

247### 5. Leverage MCP for integrations246### Leverage MCP for integrations

248 247

249MCP is a great way to give Claude Code more information, such as connecting to ticket management systems or error logs. We recommend that one central team configures MCP servers and checks a `.mcp.json` configuration into the codebase so that all users benefit. [Learn more](/en/mcp).248MCP is a great way to give Claude Code more information, such as connecting to ticket management systems or error logs. We recommend that one central team configures MCP servers and checks a `.mcp.json` configuration into the codebase so that all users benefit. [Learn more](/en/mcp).

250 249

252 251

253## Next steps252## Next steps

254 253

255* [Set up Amazon Bedrock](/en/amazon-bedrock) for AWS-native deployment254Once you've chosen a deployment option and configured access for your team:

256* [Configure Google Vertex AI](/en/google-vertex-ai) for GCP deployment

257* [Set up Microsoft Foundry](/en/microsoft-foundry) for Azure deployment

258* [Configure Enterprise Network](/en/network-config) for network requirements

259* [Deploy LLM Gateway](/en/llm-gateway) for enterprise management

260* [Settings](/en/settings) for configuration options and environment variables

~~261~~

~~262~~

263 255

264> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt2561. **Roll out to your team**: Share installation instructions and have team members [install Claude Code](/en/setup) and authenticate with their credentials.

2572. **Set up shared configuration**: Create a [CLAUDE.md file](/en/memory) in your repositories to help Claude Code understand your codebase and coding standards.

2583. **Configure permissions**: Review [security settings](/en/security) to define what Claude Code can and cannot do in your environment.

troubleshooting.md +43 −14

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Troubleshooting5# Troubleshooting

2 6

3> Discover solutions to common issues with Claude Code installation and usage.7> Discover solutions to common issues with Claude Code installation and usage.

53 Avoid disabling Windows PATH importing (`appendWindowsPath = false`) as this breaks the ability to call Windows executables from WSL. Similarly, avoid uninstalling Node.js from Windows if you use it for Windows development.57 Avoid disabling Windows PATH importing (`appendWindowsPath = false`) as this breaks the ability to call Windows executables from WSL. Similarly, avoid uninstalling Node.js from Windows if you use it for Windows development.

54</Warning>58</Warning>

55 59

60### WSL2 sandbox setup

62[Sandboxing](/en/sandboxing) is supported on WSL2 but requires installing additional packages. If you see an error like "Sandbox requires socat and bubblewrap" when running `/sandbox`, install the dependencies:

64<Tabs>

65 <Tab title="Ubuntu/Debian">

66 ```bash theme={null}

67 sudo apt-get install bubblewrap socat

68 ```

69 </Tab>

71 <Tab title="Fedora">

72 ```bash theme={null}

73 sudo dnf install bubblewrap socat

74 ```

75 </Tab>

76</Tabs>

78WSL1 does not support sandboxing. If you see "Sandboxing requires WSL2", you need to upgrade to WSL2 or run Claude Code without sandboxing.

56### Linux and Mac installation issues: permission or command not found errors80### Linux and Mac installation issues: permission or command not found errors

57 81

58When installing Claude Code with npm, `PATH` problems may prevent access to `claude`.82When installing Claude Code with npm, `PATH` problems may prevent access to `claude`.

143### Repeated permission prompts167### Repeated permission prompts

144 168

145If you find yourself repeatedly approving the same commands, you can allow specific tools169If you find yourself repeatedly approving the same commands, you can allow specific tools

146to run without approval using the `/permissions` command. See [Permissions docs](/en/iam#configuring-permissions).170to run without approval using the `/permissions` command. See [Permissions docs](/en/permissions#manage-permissions).

147 171

148### Authentication issues172### Authentication issues

149 173

1532. Close Claude Code1772. Close Claude Code

1543. Restart with `claude` and complete the authentication process again1783. Restart with `claude` and complete the authentication process again

155 179

180If the browser doesn't open automatically during login, press `c` to copy the OAuth URL to your clipboard, then paste it into your browser manually.

181

156If problems persist, try:182If problems persist, try:

157 183

158```bash theme={null}184```bash theme={null}

167Claude Code stores configuration in several locations:193Claude Code stores configuration in several locations:

168 194

169| File | Purpose |195| File | Purpose |

170| :---------------------------- | :--------------------------------------------------------------------- |196| :---------------------------- | :------------------------------------------------------- |

178 204

179On Windows, `~` refers to your user home directory, such as `C:\Users\YourName`.205On Windows, `~` refers to your user home directory, such as `C:\Users\YourName`.

180 206

181**Enterprise managed file locations:**207**Managed file locations:**

182 208

183* macOS: `/Library/Application Support/ClaudeCode/`209* macOS: `/Library/Application Support/ClaudeCode/`

184* Linux/WSL: `/etc/claude-code/`210* Linux/WSL: `/etc/claude-code/`

201```227```

202 228

203<Warning>229<Warning>

204 This will remove all your settings, allowed tools, MCP server configurations, and session history.230 This will remove all your settings, MCP server configurations, and session history.

205</Warning>231</Warning>

206 232

207## Performance and stability233## Performance and stability

223 249

224### Search and discovery issues250### Search and discovery issues

225 251

226If Search tool, `@file` mentions, custom agents, and custom slash commands aren't working, install system `ripgrep`:252If Search tool, `@file` mentions, custom agents, and custom skills aren't working, install system `ripgrep`:

227 253

228```bash theme={null}254```bash theme={null}

229# macOS (Homebrew) 255# macOS (Homebrew)

357 383

3581. **Ask Claude to add language tags**: Request "Add appropriate language tags to all code blocks in this markdown file."3841. **Ask Claude to add language tags**: Request "Add appropriate language tags to all code blocks in this markdown file."

359 385

3602. **Use post-processing hooks**: Set up automatic formatting hooks to detect and add missing language tags. See the [markdown formatting hook example](/en/hooks-guide#markdown-formatting-hook) for implementation details.3862. **Use post-processing hooks**: Set up automatic formatting hooks to detect and add missing language tags. See [Auto-format code after edits](/en/hooks-guide#auto-format-code-after-edits) for an example of a PostToolUse formatting hook.

361 387

3623. **Manual verification**: After generating markdown files, review them for proper code block formatting and request corrections if needed.3883. **Manual verification**: After generating markdown files, review them for proper code block formatting and request corrections if needed.

363 389

387 413

3881. Use the `/bug` command within Claude Code to report problems directly to Anthropic4141. Use the `/bug` command within Claude Code to report problems directly to Anthropic

3892. Check the [GitHub repository](https://github.com/anthropics/claude-code) for known issues4152. Check the [GitHub repository](https://github.com/anthropics/claude-code) for known issues

3903. Run `/doctor` to check the health of your Claude Code installation4163. Run `/doctor` to diagnose issues. It checks:

417 * Installation type, version, and search functionality

418 * Auto-update status and available versions

419 * Invalid settings files (malformed JSON, incorrect types)

420 * MCP server configuration errors

421 * Keybinding configuration problems

422 * Context usage warnings (large CLAUDE.md files, high MCP token usage, unreachable permission rules)

423 * Plugin and agent loading errors

3914. Ask Claude directly about its capabilities and features - Claude has built-in access to its documentation4244. Ask Claude directly about its capabilities and features - Claude has built-in access to its documentation

~~392~~

~~393~~

~~394~~

395> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

vs-code.md +260 −83

Details

1> ## Documentation Index

2> Fetch the complete documentation index at: https://code.claude.com/docs/llms.txt

3> Use this file to discover all available pages before exploring further.

1# Use Claude Code in VS Code5# Use Claude Code in VS Code

2 6

3> Install and configure the Claude Code extension for VS Code. Get AI coding assistance with inline diffs, @-mentions, plan review, and keyboard shortcuts.7> Install and configure the Claude Code extension for VS Code. Get AI coding assistance with inline diffs, @-mentions, plan review, and keyboard shortcuts.

13* VS Code 1.98.0 or higher17* VS Code 1.98.0 or higher

14* An Anthropic account (you'll sign in when you first open the extension). If you're using a third-party provider like Amazon Bedrock or Google Vertex AI, see [Use third-party providers](#use-third-party-providers) instead.18* An Anthropic account (you'll sign in when you first open the extension). If you're using a third-party provider like Amazon Bedrock or Google Vertex AI, see [Use third-party providers](#use-third-party-providers) instead.

15 19

16You don't need to install the Claude Code CLI first. However, some features like MCP server configuration require the CLI. See [VS Code extension vs. Claude Code CLI](#vs-code-extension-vs-claude-code-cli) for details.20<Tip>

21 The extension includes the CLI (command-line interface), which you can access from VS Code's integrated terminal for advanced features. See [VS Code extension vs. Claude Code CLI](#vs-code-extension-vs-claude-code-cli) for details.

22</Tip>

17 23

18## Install the extension24## Install the extension

19 25

24 30

25Or in VS Code, press `Cmd+Shift+X` (Mac) or `Ctrl+Shift+X` (Windows/Linux) to open the Extensions view, search for "Claude Code", and click **Install**.31Or in VS Code, press `Cmd+Shift+X` (Mac) or `Ctrl+Shift+X` (Windows/Linux) to open the Extensions view, search for "Claude Code", and click **Install**.

26 32

~~27<Note>You may need to restart VS Code or run "Developer: Reload Window" from the Command Palette after installation.</Note>~~33<Note>If the extension doesn't appear after installation, restart VS Code or run "Developer: Reload Window" from the Command Palette.</Note>

28 34

29## Get started35## Get started

30 36

43 * **Command Palette**: `Cmd+Shift+P` (Mac) or `Ctrl+Shift+P` (Windows/Linux), type "Claude Code", and select an option like "Open in New Tab"49 * **Command Palette**: `Cmd+Shift+P` (Mac) or `Ctrl+Shift+P` (Windows/Linux), type "Claude Code", and select an option like "Open in New Tab"

44 * **Status Bar**: Click **✱ Claude Code** in the bottom-right corner of the window. This works even when no file is open.50 * **Status Bar**: Click **✱ Claude Code** in the bottom-right corner of the window. This works even when no file is open.

45 51

52 When you first open the panel, a **Learn Claude Code** checklist appears. Work through each item by clicking **Show me**, or dismiss it with the X. To reopen it later, uncheck **Hide Onboarding** in VS Code settings under Extensions → Claude Code.

46 You can drag the Claude panel to reposition it anywhere in VS Code. See [Customize your workflow](#customize-your-workflow) for details.54 You can drag the Claude panel to reposition it anywhere in VS Code. See [Customize your workflow](#customize-your-workflow) for details.

47 </Step>55 </Step>

48 56

49 <Step title="Send a prompt">57 <Step title="Send a prompt">

50 Ask Claude to help with your code or files, whether that's explaining how something works, debugging an issue, or making changes.58 Ask Claude to help with your code or files, whether that's explaining how something works, debugging an issue, or making changes.

51 59

~~52 <Tip>Select text in the editor and press `Alt+K` to insert an @-mention with the file path and line numbers directly into your prompt.</Tip>~~60 <Tip>Claude automatically sees your selected text. Press `Option+K` (Mac) / `Alt+K` (Windows/Linux) to also insert an @-mention reference (like `@file.ts#5-10`) into your prompt.</Tip>

53 61

54 Here's an example of asking about a particular line in a file:62 Here's an example of asking about a particular line in a file:

55 63

57 </Step>65 </Step>

58 66

59 <Step title="Review changes">67 <Step title="Review changes">

~~60 When Claude wants to edit a file, it shows you a diff and asks for permission. You can accept, reject, or tell Claude what to do instead.~~68 When Claude wants to edit a file, it shows a side-by-side comparison of the original and proposed changes, then asks for permission. You can accept, reject, or tell Claude what to do instead.

61 69

62 <img src="https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=e005f9b41c541c5c7c59c082f7c4841c" alt="VS Code showing a diff of Claude's proposed changes with a permission prompt asking whether to make the edit" data-og-width="3292" width="3292" data-og-height="1876" height="1876" data-path="images/vs-code-edits.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=280&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=cb5d41b81087f79b842a56b5a3304660 280w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=560&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=90bb691960decdc06393c3c21cd62c75 560w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=840&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=9a11bf878ba619e850380904ff4f38e8 840w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=1100&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=6dddbf596b4f69ec6245bdc5eb6dd487 1100w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=1650&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=ef2713b8cbfd2cee97af817d813d64c7 1650w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=2500&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=1f7e1c52919cdfddf295f32a2ec7ae59 2500w" />70 <img src="https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=e005f9b41c541c5c7c59c082f7c4841c" alt="VS Code showing a diff of Claude's proposed changes with a permission prompt asking whether to make the edit" data-og-width="3292" width="3292" data-og-height="1876" height="1876" data-path="images/vs-code-edits.png" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=280&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=cb5d41b81087f79b842a56b5a3304660 280w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=560&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=90bb691960decdc06393c3c21cd62c75 560w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=840&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=9a11bf878ba619e850380904ff4f38e8 840w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=1100&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=6dddbf596b4f69ec6245bdc5eb6dd487 1100w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=1650&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=ef2713b8cbfd2cee97af817d813d64c7 1650w, https://mintcdn.com/claude-code/FVYz38sRY-VuoGHA/images/vs-code-edits.png?w=2500&fit=max&auto=format&n=FVYz38sRY-VuoGHA&q=85&s=1f7e1c52919cdfddf295f32a2ec7ae59 2500w" />

63 </Step>71 </Step>

65 73

66For more ideas on what you can do with Claude Code, see [Common workflows](/en/common-workflows).74For more ideas on what you can do with Claude Code, see [Common workflows](/en/common-workflows).

67 75

76<Tip>

77 Run "Claude Code: Open Walkthrough" from the Command Palette for a guided tour of the basics.

78</Tip>

80## Use the prompt box

82The prompt box supports several features:

84* **Permission modes**: Click the mode indicator at the bottom of the prompt box to switch modes. In normal mode, Claude asks permission before each action. In Plan mode, Claude describes what it will do and waits for approval before making changes. In auto-accept mode, Claude makes edits without asking. Set the default in VS Code settings under `claudeCode.initialPermissionMode`.

85* **Command menu**: Click `/` or type `/` to open the command menu. Options include attaching files, switching models, toggling extended thinking, and viewing plan usage (`/usage`). The Customize section provides access to MCP servers, hooks, memory, permissions, and plugins. Items with a terminal icon open in the integrated terminal.

86* **Context indicator**: The prompt box shows how much of Claude's context window you're using. Claude automatically compacts when needed, or you can run `/compact` manually.

87* **Extended thinking**: Lets Claude spend more time reasoning through complex problems. Toggle it on via the command menu (`/`). See [Extended thinking](/en/common-workflows#use-extended-thinking-thinking-mode) for details.

88* **Multi-line input**: Press `Shift+Enter` to add a new line without sending. This also works in the "Other" free-text input of question dialogs.

90### Reference files and folders

92Use @-mentions to give Claude context about specific files or folders. When you type `@` followed by a file or folder name, Claude reads that content and can answer questions about it or make changes to it. Claude Code supports fuzzy matching, so you can type partial names to find what you need:

94```

95> Explain the logic in @auth (fuzzy matches auth.js, AuthService.ts, etc.)

96> What's in @src/components/ (include a trailing slash for folders)

97```

99For large PDFs, you can ask Claude to read specific pages instead of the whole file: a single page, a range like pages 1-10, or an open-ended range like page 3 onward.

100

101When you select text in the editor, Claude can see your highlighted code automatically. The prompt box footer shows how many lines are selected. Press `Option+K` (Mac) / `Alt+K` (Windows/Linux) to insert an @-mention with the file path and line numbers (e.g., `@app.ts#5-10`). Click the selection indicator to toggle whether Claude can see your highlighted text - the eye-slash icon means the selection is hidden from Claude.

102

103You can also hold `Shift` while dragging files into the prompt box to add them as attachments. Click the X on any attachment to remove it from context.

104

105### Resume past conversations

106

107Click the dropdown at the top of the Claude Code panel to access your conversation history. You can search by keyword or browse by time (Today, Yesterday, Last 7 days, etc.). Click any conversation to resume it with the full message history. For more on resuming sessions, see [Common workflows](/en/common-workflows#resume-previous-conversations).

108

109### Resume remote sessions from Claude.ai

110

111If you use [Claude Code on the web](/en/claude-code-on-the-web), you can resume those remote sessions directly in VS Code. This requires signing in with **Claude.ai Subscription**, not Anthropic Console.

112

113<Steps>

114 <Step title="Open Past Conversations">

115 Click the **Past Conversations** dropdown at the top of the Claude Code panel.

116 </Step>

117

118 <Step title="Select the Remote tab">

119 The dialog shows two tabs: Local and Remote. Click **Remote** to see sessions from claude.ai.

120 </Step>

121

122 <Step title="Select a session to resume">

123 Browse or search your remote sessions. Click any session to download it and continue the conversation locally.

124 </Step>

125</Steps>

126

127<Note>

128 Only web sessions started with a GitHub repository appear in the Remote tab. Resuming loads the conversation history locally; changes are not synced back to claude.ai.

129</Note>

130

68## Customize your workflow131## Customize your workflow

69 132

~~70Once you're up and running, you can reposition the Claude panel or switch to terminal mode.~~133Once you're up and running, you can reposition the Claude panel, run multiple sessions, or switch to terminal mode.

71 134

~~72### Change the layout~~135### Choose where Claude lives

73 136

74You can drag the Claude panel to reposition it anywhere in VS Code. Grab the panel's tab or title bar and drag it to:137You can drag the Claude panel to reposition it anywhere in VS Code. Grab the panel's tab or title bar and drag it to:

75 138

~~76* **Secondary sidebar** (default): The right side of the window~~139* **Secondary sidebar**: The right side of the window. Keeps Claude visible while you code.

77* **Primary sidebar**: The left sidebar with icons for Explorer, Search, etc.140* **Primary sidebar**: The left sidebar with icons for Explorer, Search, etc.

~~78* **Editor area**: Opens Claude as a tab alongside your files~~141* **Editor area**: Opens Claude as a tab alongside your files. Useful for side tasks.

79 142

~~80<Note>~~143<Tip>

81 The Spark icon only appears in the Activity Bar (left sidebar icons) when the Claude panel is docked to the left. Since Claude defaults to the right side, use the Editor Toolbar icon to open Claude.144 Use the sidebar for your main Claude session and open additional tabs for side tasks. Claude remembers your preferred location. Note that the Spark icon only appears in the Activity Bar when the Claude panel is docked to the left. Since Claude defaults to the right side, use the Editor Toolbar icon to open Claude.

~~82</Note>~~145</Tip>

146

147### Run multiple conversations

148

149Use **Open in New Tab** or **Open in New Window** from the Command Palette to start additional conversations. Each conversation maintains its own history and context, allowing you to work on different tasks in parallel.

150

151When using tabs, a small colored dot on the spark icon indicates status: blue means a permission request is pending, orange means Claude finished while the tab was hidden.

83 152

84### Switch to terminal mode153### Switch to terminal mode

85 154

87 156

88You can also open VS Code settings (`Cmd+,` on Mac or `Ctrl+,` on Windows/Linux), go to Extensions → Claude Code, and check **Use Terminal**.157You can also open VS Code settings (`Cmd+,` on Mac or `Ctrl+,` on Windows/Linux), go to Extensions → Claude Code, and check **Use Terminal**.

89 158

159## Manage plugins

160

161The VS Code extension includes a graphical interface for installing and managing [plugins](/en/plugins). Type `/plugins` in the prompt box to open the **Manage plugins** interface.

162

163### Install plugins

164

165The plugin dialog shows two tabs: **Plugins** and **Marketplaces**.

166

167In the Plugins tab:

168

169* **Installed plugins** appear at the top with toggle switches to enable or disable them

170* **Available plugins** from your configured marketplaces appear below

171* Search to filter plugins by name or description

172* Click **Install** on any available plugin

173

174When you install a plugin, choose the installation scope:

175

176* **Install for you**: Available in all your projects (user scope)

177* **Install for this project**: Shared with project collaborators (project scope)

178* **Install locally**: Only for you, only in this repository (local scope)

179

180### Manage marketplaces

181

182Switch to the **Marketplaces** tab to add or remove plugin sources:

183

184* Enter a GitHub repo, URL, or local path to add a new marketplace

185* Click the refresh icon to update a marketplace's plugin list

186* Click the trash icon to remove a marketplace

187

188After making changes, a banner prompts you to restart Claude Code to apply the updates.

189

190<Note>

191 Plugin management in VS Code uses the same CLI commands under the hood. Plugins and marketplaces you configure in the extension are also available in the CLI, and vice versa.

192</Note>

193

194For more about the plugin system, see [Plugins](/en/plugins) and [Plugin marketplaces](/en/plugin-marketplaces).

195

196## Automate browser tasks with Chrome

197

198Connect Claude to your Chrome browser to test web apps, debug with console logs, and automate browser workflows without leaving VS Code. This requires the [Claude in Chrome extension](https://chromewebstore.google.com/detail/claude/fcoeoabgfenejglbffodgkkbkcdhcgfn) version 1.0.36 or higher.

199

200Type `@browser` in the prompt box followed by what you want Claude to do:

201

202```text theme={null}

203@browser go to localhost:3000 and check the console for errors

204```

205

206You can also open the attachment menu to select specific browser tools like opening a new tab or reading page content.

207

208Claude opens new tabs for browser tasks and shares your browser's login state, so it can access any site you're already signed into.

209

210For setup instructions, the full list of capabilities, and troubleshooting, see [Use Claude Code with Chrome](/en/chrome).

211

90## VS Code commands and shortcuts212## VS Code commands and shortcuts

91 213

~~92Open the Command Palette (`Cmd+Shift+P` on Mac or `Ctrl+Shift+P` on Windows/Linux) and type "Claude Code" to see all available VS Code commands for the Claude Code extension:~~214Open the Command Palette (`Cmd+Shift+P` on Mac or `Ctrl+Shift+P` on Windows/Linux) and type "Claude Code" to see all available VS Code commands for the Claude Code extension.

215

216Some shortcuts depend on which panel is "focused" (receiving keyboard input). When your cursor is in a code file, the editor is focused. When your cursor is in Claude's prompt box, Claude is focused. Use `Cmd+Esc` / `Ctrl+Esc` to toggle between them.

93 217

94<Note>218<Note>

95 These are VS Code commands for controlling the extension. For Claude Code slash commands (like `/help` or `/compact`), not all CLI commands are available in the extension yet. See [VS Code extension vs. Claude Code CLI](#vs-code-extension-vs-claude-code-cli) for details.219 These are VS Code commands for controlling the extension. Not all built-in Claude Code commands are available in the extension. See [VS Code extension vs. Claude Code CLI](#vs-code-extension-vs-claude-code-cli) for details.

96</Note>220</Note>

97 221

~~99| -------------------------- | -------------------------------------------------------- | ---------------------------------------------------------------------------------- |~~223| -------------------------- | -------------------------------------------------------- | ------------------------------------------------------------------------------------ |

101| Open in Side Bar | — | Open Claude in the left sidebar |225| Open in Side Bar | - | Open Claude in the left sidebar |

102| Open in Terminal | — | Open Claude in terminal mode |226| Open in Terminal | - | Open Claude in terminal mode |

104| Open in New Window | — | Open a new conversation in a separate window |228| Open in New Window | - | Open a new conversation in a separate window |

107| Show Logs | — | View extension debug logs |231| Show Logs | - | View extension debug logs |

108| Logout | — | Sign out of your Anthropic account |232| Logout | - | Sign out of your Anthropic account |

~~109~~

110Use **Open in New Tab** or **Open in New Window** to run multiple conversations simultaneously. Each tab or window maintains its own conversation history and context.

111 233

112## Configure settings234## Configure settings

113 235

114The extension has two types of settings:236The extension has two types of settings:

115 237

116* **Extension settings**: Open with `Cmd+,` (Mac) or `Ctrl+,` (Windows/Linux), then go to Extensions → Claude Code.238* **Extension settings** in VS Code: Control the extension's behavior within VS Code. Open with `Cmd+,` (Mac) or `Ctrl+,` (Windows/Linux), then go to Extensions → Claude Code. You can also type `/` and select **General Config** to open settings.

~~117~~ 239* **Claude Code settings** in `~/.claude/settings.json`: Shared between the extension and CLI. Use for allowed commands, environment variables, hooks, and MCP servers. See [Settings](/en/settings) for details.

118 | Setting | Description |240

119 | ---------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |241<Tip>

120 | Selected Model | Default model for new conversations. Change per-session with `/model`. |242 Add `"$schema": "https://json.schemastore.org/claude-code-settings.json"` to your `settings.json` to get autocomplete and inline validation for all available settings directly in VS Code.

121 | Use Terminal | Launch Claude in terminal mode instead of graphical panel |243</Tip>

122 | Initial Permission Mode | Controls approval prompts for file edits and commands. Defaults to `default` (ask before each action). |244

123 | Preferred Location | Default location: sidebar (right) or panel (new tab) |245### Extension settings

124 | Autosave | Auto-save files before Claude reads or writes them |246

126 | Enable New Conversation Shortcut | Enable Cmd/Ctrl+N to start a new conversation |248| --------------------------------- | --------- | ----------------------------------------------------------------------------------------------------- |

128 | Environment Variables | Set environment variables for the Claude process. **Not recommended**—use [Claude Code settings](/en/settings) instead so configuration is shared between extension and CLI. |250| `useTerminal` | `false` | Launch Claude in terminal mode instead of graphical panel |

~~132~~ 254| `useCtrlEnterToSend` | `false` | Use Ctrl/Cmd+Enter instead of Enter to send prompts |

133* **Claude Code settings** (`~/.claude/settings.json`): These settings are shared between the VS Code extension and the CLI. Use this file for allowed commands and directories, environment variables, hooks, and MCP servers. See the [settings documentation](/en/settings) for details.255| `enableNewConversationShortcut` | `true` | Enable Cmd/Ctrl+N to start a new conversation |

256| `hideOnboarding` | `false` | Hide the onboarding checklist (graduation cap icon) |

257| `respectGitIgnore` | `true` | Exclude .gitignore patterns from file searches |

258| `environmentVariables` | `[]` | Set environment variables for the Claude process. Use Claude Code settings instead for shared config. |

259| `disableLoginPrompt` | `false` | Skip authentication prompts (for third-party provider setups) |

260| `allowDangerouslySkipPermissions` | `false` | Bypass all permission prompts. **Use with extreme caution.** |

261| `claudeProcessWrapper` | - | Executable path used to launch the Claude process |

262

263## VS Code extension vs. Claude Code CLI

264

265Claude Code is available as both a VS Code extension (graphical panel) and a CLI (command-line interface in the terminal). Some features are only available in the CLI. If you need a CLI-only feature, run `claude` in VS Code's integrated terminal.

266

267| Feature | CLI | VS Code Extension |

268| ------------------- | --------------------------------------------- | ---------------------------------------- |

269| Commands and skills | [All](/en/interactive-mode#built-in-commands) | Subset (type `/` to see available) |

270| MCP server config | Yes | No (configure via CLI, use in extension) |

271| Checkpoints | Yes | Yes |

272| `!` bash shortcut | Yes | No |

273| Tab completion | Yes | No |

274

275### Rewind with checkpoints

276

277The VS Code extension supports checkpoints, which track Claude's file edits and let you rewind to a previous state. Hover over any message to reveal the rewind button, then choose from three options:

278

279* **Fork conversation from here**: start a new conversation branch from this message while keeping all code changes intact

280* **Rewind code to here**: revert file changes back to this point in the conversation while keeping the full conversation history

281* **Fork conversation and rewind code**: start a new conversation branch and revert file changes to this point

282

283For full details on how checkpoints work and their limitations, see [Checkpointing](/en/checkpointing).

284

285### Run CLI in VS Code

286

287To use the CLI while staying in VS Code, open the integrated terminal (`` Ctrl+` `` on Windows/Linux or `` Cmd+` `` on Mac) and run `claude`. The CLI automatically integrates with your IDE for features like diff viewing and diagnostic sharing.

288

289If using an external terminal, run `/ide` inside Claude Code to connect it to VS Code.

290

291### Switch between extension and CLI

292

293The extension and CLI share the same conversation history. To continue an extension conversation in the CLI, run `claude --resume` in the terminal. This opens an interactive picker where you can search for and select your conversation.

294

295### Include terminal output in prompts

296

297Reference terminal output in your prompts using `@terminal:name` where `name` is the terminal's title. This lets Claude see command output, error messages, or logs without copy-pasting.

298

299### Monitor background processes

300

301When Claude runs long-running commands, the extension shows progress in the status bar. However, visibility for background tasks is limited compared to the CLI. For better visibility, have Claude output the command so you can run it in VS Code's integrated terminal.

302

303### Connect to external tools with MCP

304

305MCP (Model Context Protocol) servers give Claude access to external tools, databases, and APIs. Configure them via CLI, then use them in both extension and CLI.

306

307To add an MCP server, open the integrated terminal (`` Ctrl+` `` or `` Cmd+` ``) and run:

308

309```bash theme={null}

310claude mcp add --transport http github https://api.githubcopilot.com/mcp/

311```

312

313Once configured, ask Claude to use the tools (e.g., "Review PR #456"). Some servers require authentication: run `claude` in the terminal, then type `/mcp` to authenticate. See the [MCP documentation](/en/mcp) for available servers.

314

315## Work with git

316

317Claude Code integrates with git to help with version control workflows directly in VS Code. Ask Claude to commit changes, create pull requests, or work across branches.

318

319### Create commits and pull requests

320

321Claude can stage changes, write commit messages, and create pull requests based on your work:

322

323```

324> commit my changes with a descriptive message

325> create a pr for this feature

326> summarize the changes I've made to the auth module

327```

328

329When creating pull requests, Claude generates descriptions based on the actual code changes and can add context about testing or implementation decisions.

330

331### Use git worktrees for parallel tasks

332

333Git worktrees allow multiple Claude Code sessions to work on separate branches simultaneously, each with isolated files:

334

335```bash theme={null}

336# Create a worktree for a new feature

337git worktree add ../project-feature-a -b feature-a

338

339# Run Claude Code in each worktree

340cd ../project-feature-a && claude

341```

342

343Each worktree maintains independent file state while sharing git history. This prevents Claude instances from interfering with each other when working on different tasks.

344

345For detailed git workflows including PR reviews and branch management, see [Common workflows](/en/common-workflows#create-pull-requests).

134 346

135## Use third-party providers347## Use third-party providers

136 348

154 </Step>366 </Step>

155</Steps>367</Steps>

156 368

157## VS Code extension vs. Claude Code CLI369## Security and privacy

~~158~~

159The extension doesn't yet have full feature parity with the CLI. If you need CLI-only features, you can run `claude` directly in VS Code's integrated terminal.

~~160~~

161| Feature | CLI | VS Code Extension |

162| ----------------- | ------------------------------ | ---------------------------------------- |

163| Slash commands | [Full set](/en/slash-commands) | Subset (type `/` to see available) |

164| MCP server config | Yes | No (configure via CLI, use in extension) |

165| Checkpoints | Yes | Coming soon |

166| `!` bash shortcut | Yes | No |

167| Tab completion | Yes | No |

~~168~~

169### Run CLI in VS Code

~~170~~

171To use the CLI while staying in VS Code, open the integrated terminal (`` Ctrl+` `` on Windows/Linux or `` Cmd+` `` on Mac) and run `claude`. The CLI automatically integrates with your IDE for features like diff viewing and diagnostic sharing.

~~172~~

173If using an external terminal, run `/ide` inside Claude Code to connect it to VS Code.

~~174~~

175### Switch between extension and CLI

~~176~~

177The extension and CLI share the same conversation history. To continue an extension conversation in the CLI, run `claude --resume` in the terminal. This opens an interactive picker where you can search for and select your conversation.

~~178~~

179## Security considerations

180 370

181With auto-edit permissions enabled, Claude Code can modify VS Code configuration files (like `settings.json` or `tasks.json`) that VS Code may execute automatically. This could potentially bypass Claude Code's normal permission prompts.371Your code stays private. Claude Code processes your code to provide assistance but does not use it to train models. For details on data handling and how to opt out of logging, see [Data and privacy](/en/data-usage).

182 372

183To reduce risk when working with untrusted code:373With auto-edit permissions enabled, Claude Code can modify VS Code configuration files (like `settings.json` or `tasks.json`) that VS Code may execute automatically. To reduce risk when working with untrusted code:

184 374

185* Enable [VS Code Restricted Mode](https://code.visualstudio.com/docs/editor/workspace-trust#_restricted-mode) for untrusted workspaces375* Enable [VS Code Restricted Mode](https://code.visualstudio.com/docs/editor/workspace-trust#_restricted-mode) for untrusted workspaces

186* Use manual approval mode instead of auto-accept for edits376* Use manual approval mode instead of auto-accept for edits

192 382

193* Ensure you have a compatible version of VS Code (1.98.0 or later)383* Ensure you have a compatible version of VS Code (1.98.0 or later)

194* Check that VS Code has permission to install extensions384* Check that VS Code has permission to install extensions

195* Try installing directly from the Marketplace website385* Try installing directly from the [VS Code Marketplace](https://marketplace.visualstudio.com/items?itemName=anthropic.claude-code)

196 386

197### Spark icon not visible387### Spark icon not visible

198 388

199The Spark icon appears in the **Editor Toolbar** (top-right of editor) when you have a file open. If you don't see it:389The Spark icon appears in the **Editor Toolbar** (top-right of editor) when you have a file open. If you don't see it:

200 390

2011. **Open a file**: The icon requires a file to be open—having just a folder open isn't enough3911. **Open a file**: The icon requires a file to be open. Having just a folder open isn't enough.

2022. **Check VS Code version**: Requires 1.98.0 or higher (Help → About)3922. **Check VS Code version**: Requires 1.98.0 or higher (Help → About)

2033. **Restart VS Code**: Run "Developer: Reload Window" from the Command Palette3933. **Restart VS Code**: Run "Developer: Reload Window" from the Command Palette

2044. **Disable conflicting extensions**: Temporarily disable other AI extensions (Cline, Continue, etc.)3944. **Disable conflicting extensions**: Temporarily disable other AI extensions (Cline, Continue, etc.)

2055. **Check workspace trust**: The extension doesn't work in Restricted Mode3955. **Check workspace trust**: The extension doesn't work in Restricted Mode

206 396

207Alternatively, click "✱ Claude Code" in the **Status Bar** (bottom-right corner)—this works even without a file open. You can also use the **Command Palette** (`Cmd+Shift+P` / `Ctrl+Shift+P`) and type "Claude Code".397Alternatively, click "✱ Claude Code" in the **Status Bar** (bottom-right corner). This works even without a file open. You can also use the **Command Palette** (`Cmd+Shift+P` / `Ctrl+Shift+P`) and type "Claude Code".

208 398

209### Claude Code never responds399### Claude Code never responds

210 400

2131. **Check your internet connection**: Ensure you have a stable internet connection4031. **Check your internet connection**: Ensure you have a stable internet connection

2142. **Start a new conversation**: Try starting a fresh conversation to see if the issue persists4042. **Start a new conversation**: Try starting a fresh conversation to see if the issue persists

2153. **Try the CLI**: Run `claude` from the terminal to see if you get more detailed error messages4053. **Try the CLI**: Run `claude` from the terminal to see if you get more detailed error messages

2164. **File a bug report**: If the problem continues, [file an issue on GitHub](https://github.com/anthropics/claude-code/issues) with details about the error

217 406

218### Standalone CLI not connecting to IDE407If problems persist, [file an issue on GitHub](https://github.com/anthropics/claude-code/issues) with details about the error.

~~219~~

220* Ensure you're running Claude Code from VS Code's integrated terminal (not an external terminal)

221* Ensure the CLI for your IDE variant is installed:

222 * VS Code: `code` command should be available

223 * Cursor: `cursor` command should be available

224 * Windsurf: `windsurf` command should be available

225 * VSCodium: `codium` command should be available

226* If the command isn't available, install it from the Command Palette → "Shell Command: Install 'code' command in PATH"

227 408

228## Uninstall the extension409## Uninstall the extension

229 410

248* [Explore common workflows](/en/common-workflows) to get the most out of Claude Code429* [Explore common workflows](/en/common-workflows) to get the most out of Claude Code

249* [Set up MCP servers](/en/mcp) to extend Claude's capabilities with external tools. Configure servers using the CLI, then use them in the extension.430* [Set up MCP servers](/en/mcp) to extend Claude's capabilities with external tools. Configure servers using the CLI, then use them in the extension.

250* [Configure Claude Code settings](/en/settings) to customize allowed commands, hooks, and more. These settings are shared between the extension and CLI.431* [Configure Claude Code settings](/en/settings) to customize allowed commands, hooks, and more. These settings are shared between the extension and CLI.

~~251~~

~~252~~

~~253~~

254> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt

Anthropic - Claude Code Docs Changes 2026-01-07 00:04 UTC → 2026-02-12 00:06 UTC