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.
4
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, HTTP 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## Configuration13Hooks are user-defined shell commands, HTTP endpoints, 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, HTTP hooks, and MCP tool hooks. If you're setting up hooks for the first time, start with the [guide](/en/hooks-guide) instead.
14
15## Hook lifecycle
16
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, input arrives on stdin. For HTTP hooks, it arrives as the POST request body. 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:
10 18
11Claude Code hooks are configured in your [settings files](/en/settings):19<div style={{maxWidth: "500px", margin: "0 auto"}}>
20 <Frame>
21 <img src="https://mintcdn.com/claude-code/rsuu-ovdPNos9Dnn/images/hooks-lifecycle.svg?fit=max&auto=format&n=rsuu-ovdPNos9Dnn&q=85&s=ce5f1225339bbccdfbb52e99205db912" alt="Hook lifecycle diagram showing the sequence of hooks from SessionStart through the agentic loop to SessionEnd, with WorktreeCreate and WorktreeRemove as standalone setup and teardown events" data-og-width="520" width="520" data-og-height="1020" height="1020" data-path="images/hooks-lifecycle.svg" data-optimize="true" data-opv="3" srcset="https://mintcdn.com/claude-code/rsuu-ovdPNos9Dnn/images/hooks-lifecycle.svg?w=280&fit=max&auto=format&n=rsuu-ovdPNos9Dnn&q=85&s=7c7143c65492c1beb6bc66f5d206ba15 280w, https://mintcdn.com/claude-code/rsuu-ovdPNos9Dnn/images/hooks-lifecycle.svg?w=560&fit=max&auto=format&n=rsuu-ovdPNos9Dnn&q=85&s=dafaebf8f789f94edbf6bd66853c69df 560w, https://mintcdn.com/claude-code/rsuu-ovdPNos9Dnn/images/hooks-lifecycle.svg?w=840&fit=max&auto=format&n=rsuu-ovdPNos9Dnn&q=85&s=2caa51d2d95596f1f80b92e3f5f534fa 840w, https://mintcdn.com/claude-code/rsuu-ovdPNos9Dnn/images/hooks-lifecycle.svg?w=1100&fit=max&auto=format&n=rsuu-ovdPNos9Dnn&q=85&s=614def559f34f9b0c1dec93739d96b64 1100w, https://mintcdn.com/claude-code/rsuu-ovdPNos9Dnn/images/hooks-lifecycle.svg?w=1650&fit=max&auto=format&n=rsuu-ovdPNos9Dnn&q=85&s=ca45b85fdd8b2da81c69d12c453230cb 1650w, https://mintcdn.com/claude-code/rsuu-ovdPNos9Dnn/images/hooks-lifecycle.svg?w=2500&fit=max&auto=format&n=rsuu-ovdPNos9Dnn&q=85&s=7fd92d6b9713493f59962c9f295c9d2f 2500w" />
22 </Frame>
23</div>
12 24
13* `~/.claude/settings.json` - User settings25The 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.
14* `.claude/settings.json` - Project settings
15* `.claude/settings.local.json` - Local project settings (not committed)
16* Enterprise managed policy settings
17 26
18### Structure27| 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| `ConfigChange` | When a configuration file changes during a session |
42| `WorktreeCreate` | When a worktree is being created via `--worktree` or `isolation: "worktree"`. Replaces default git behavior |
43| `WorktreeRemove` | When a worktree is being removed, either at session exit or when a subagent finishes |
44| `PreCompact` | Before context compaction |
45| `SessionEnd` | When a session terminates |
19 46
20Hooks are organized by matchers, where each matcher can have multiple hooks:47### How a hook resolves
48
49To 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:
21 50
22```json theme={null}51```json theme={null}
23{52{
24 "hooks": {53 "hooks": {
25 "EventName": [54 "PreToolUse": [
26 {55 {
27 "matcher": "ToolPattern",56 "matcher": "Bash",
28 "hooks": [57 "hooks": [
29 {58 {
30 "type": "command",59 "type": "command",
31 "command": "your-command-here"60 "command": ".claude/hooks/block-rm.sh"
32 }61 }
33 ]62 ]
34 }63 }
37}66}
38```67```
39 68
40* **matcher**: Pattern to match tool names, case-sensitive (only applicable for69The script reads the JSON input from stdin, extracts the command, and returns a `permissionDecision` of `"deny"` if it contains `rm -rf`:
41 `PreToolUse`, `PermissionRequest`, and `PostToolUse`)
42 * Simple strings match exactly: `Write` matches only the Write tool
43 * Supports regex: `Edit|Write` or `Notebook.*`
44 * Use `*` to match all tools. You can also use empty string (`""`) or leave
45 `matcher` blank.
46* **hooks**: Array of hooks to execute when the pattern matches
47 * `type`: Hook execution type - `"command"` for bash commands or `"prompt"` for LLM-based evaluation
48 * `command`: (For `type: "command"`) The bash command to execute (can use `$CLAUDE_PROJECT_DIR` environment variable)
49 * `prompt`: (For `type: "prompt"`) The prompt to send to the LLM for evaluation
50 * `timeout`: (Optional) How long a hook should run, in seconds, before canceling that specific hook
51
52For events like `UserPromptSubmit`, `Stop`, and `SubagentStop`
53that don't use matchers, you can omit the matcher field:
54 70
55```json theme={null}71```bash theme={null}
56{72#!/bin/bash
57 "hooks": {73# .claude/hooks/block-rm.sh
58 "UserPromptSubmit": [74COMMAND=$(jq -r '.tool_input.command')
59 {75
60 "hooks": [76if echo "$COMMAND" | grep -q 'rm -rf'; then
61 {77 jq -n '{
62 "type": "command",78 hookSpecificOutput: {
63 "command": "/path/to/prompt-validator.py"79 hookEventName: "PreToolUse",
80 permissionDecision: "deny",
81 permissionDecisionReason: "Destructive command blocked by hook"
64 }82 }
65 ]83 }'
84else
85 exit 0 # allow the command
86fi
87```
88
89Now suppose Claude Code decides to run `Bash "rm -rf /tmp/build"`. Here's what happens:
90
91<Frame>
92 <img src="https://mintcdn.com/claude-code/TBPmHzr19mDCuhZi/images/hook-resolution.svg?fit=max&auto=format&n=TBPmHzr19mDCuhZi&q=85&s=5bb890134390ecd0581477cf41ef730b" 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/TBPmHzr19mDCuhZi/images/hook-resolution.svg?w=280&fit=max&auto=format&n=TBPmHzr19mDCuhZi&q=85&s=5dcaecd24c260b8a90365d74e2c1fcda 280w, https://mintcdn.com/claude-code/TBPmHzr19mDCuhZi/images/hook-resolution.svg?w=560&fit=max&auto=format&n=TBPmHzr19mDCuhZi&q=85&s=c03d91c279f01d92e58ddd70fdbe66f2 560w, https://mintcdn.com/claude-code/TBPmHzr19mDCuhZi/images/hook-resolution.svg?w=840&fit=max&auto=format&n=TBPmHzr19mDCuhZi&q=85&s=1be57a4819cbb949a5ea9d08a05c9ecd 840w, https://mintcdn.com/claude-code/TBPmHzr19mDCuhZi/images/hook-resolution.svg?w=1100&fit=max&auto=format&n=TBPmHzr19mDCuhZi&q=85&s=0e9dd1807dc7a5c56011d0889b0d5208 1100w, https://mintcdn.com/claude-code/TBPmHzr19mDCuhZi/images/hook-resolution.svg?w=1650&fit=max&auto=format&n=TBPmHzr19mDCuhZi&q=85&s=69496ac02e70fabfece087ba31a1dcfc 1650w, https://mintcdn.com/claude-code/TBPmHzr19mDCuhZi/images/hook-resolution.svg?w=2500&fit=max&auto=format&n=TBPmHzr19mDCuhZi&q=85&s=a012346cb46a33b86580348802055267 2500w" />
93</Frame>
94
95<Steps>
96 <Step title="Event fires">
97 The `PreToolUse` event fires. Claude Code sends the tool input as JSON on stdin to the hook:
98
99 ```json theme={null}
100 { "tool_name": "Bash", "tool_input": { "command": "rm -rf /tmp/build" }, ... }
101 ```
102 </Step>
103
104 <Step title="Matcher checks">
105 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.
106 </Step>
107
108 <Step title="Hook handler runs">
109 The script extracts `"rm -rf /tmp/build"` from the input and finds `rm -rf`, so it prints a decision to stdout:
110
111 ```json theme={null}
112 {
113 "hookSpecificOutput": {
114 "hookEventName": "PreToolUse",
115 "permissionDecision": "deny",
116 "permissionDecisionReason": "Destructive command blocked by hook"
66 }117 }
67 ]
68 }118 }
69}119 ```
70```120
121 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.
122 </Step>
71 123
72### Project-Specific Hook Scripts124 <Step title="Claude Code acts on the result">
125 Claude Code reads the JSON decision, blocks the tool call, and shows Claude the reason.
126 </Step>
127</Steps>
73 128
74You can use the environment variable `CLAUDE_PROJECT_DIR` (only available when129The [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.
75Claude Code spawns the hook command) to reference scripts stored in your project,130
76ensuring they work regardless of Claude's current directory:131## Configuration
132
133Hooks are defined in JSON settings files. The configuration has three levels of nesting:
134
1351. Choose a [hook event](#hook-events) to respond to, like `PreToolUse` or `Stop`
1362. Add a [matcher group](#matcher-patterns) to filter when it fires, like "only for the Bash tool"
1373. Define one or more [hook handlers](#hook-handler-fields) to run when matched
138
139See [How a hook resolves](#how-a-hook-resolves) above for a complete walkthrough with an annotated example.
140
141<Note>
142 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, HTTP endpoint, prompt, or agent that runs. "Hook" on its own refers to the general feature.
143</Note>
144
145### Hook locations
146
147Where you define a hook determines its scope:
148
149| Location | Scope | Shareable |
150| :--------------------------------------------------------- | :---------------------------- | :--------------------------------- |
151| `~/.claude/settings.json` | All your projects | No, local to your machine |
152| `.claude/settings.json` | Single project | Yes, can be committed to the repo |
153| `.claude/settings.local.json` | Single project | No, gitignored |
154| Managed policy settings | Organization-wide | Yes, admin-controlled |
155| [Plugin](/en/plugins) `hooks/hooks.json` | When plugin is enabled | Yes, bundled with the plugin |
156| [Skill](/en/skills) or [agent](/en/sub-agents) frontmatter | While the component is active | Yes, defined in the component file |
157
158For 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).
159
160### Matcher patterns
161
162The `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:
163
164| Event | What the matcher filters | Example matcher values |
165| :---------------------------------------------------------------------------------------------- | :------------------------ | :--------------------------------------------------------------------------------- |
166| `PreToolUse`, `PostToolUse`, `PostToolUseFailure`, `PermissionRequest` | tool name | `Bash`, `Edit\|Write`, `mcp__.*` |
167| `SessionStart` | how the session started | `startup`, `resume`, `clear`, `compact` |
168| `SessionEnd` | why the session ended | `clear`, `logout`, `prompt_input_exit`, `bypass_permissions_disabled`, `other` |
169| `Notification` | notification type | `permission_prompt`, `idle_prompt`, `auth_success`, `elicitation_dialog` |
170| `SubagentStart` | agent type | `Bash`, `Explore`, `Plan`, or custom agent names |
171| `PreCompact` | what triggered compaction | `manual`, `auto` |
172| `SubagentStop` | agent type | same values as `SubagentStart` |
173| `ConfigChange` | configuration source | `user_settings`, `project_settings`, `local_settings`, `policy_settings`, `skills` |
174| `UserPromptSubmit`, `Stop`, `TeammateIdle`, `TaskCompleted`, `WorktreeCreate`, `WorktreeRemove` | no matcher support | always fires on every occurrence |
175
176The 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.
177
178This example runs a linting script only when Claude writes or edits a file:
77 179
78```json theme={null}180```json theme={null}
79{181{
80 "hooks": {182 "hooks": {
81 "PostToolUse": [183 "PostToolUse": [
82 {184 {
83 "matcher": "Write|Edit",185 "matcher": "Edit|Write",
84 "hooks": [186 "hooks": [
85 {187 {
86 "type": "command",188 "type": "command",
87 "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/check-style.sh"189 "command": "/path/to/lint-check.sh"
88 }190 }
89 ]191 ]
90 }192 }
93}195}
94```196```
95 197
96### Plugin hooks198`UserPromptSubmit`, `Stop`, `TeammateIdle`, `TaskCompleted`, `WorktreeCreate`, and `WorktreeRemove` don't support matchers and always fire on every occurrence. If you add a `matcher` field to these events, it is silently ignored.
199
200#### Match MCP tools
97 201
98[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.202[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.
99 203
100**How plugin hooks work**:204MCP tools follow the naming pattern `mcp__<server>__<tool>`, for example:
101 205
102* 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.206* `mcp__memory__create_entities`: Memory server's create entities tool
103* When a plugin is enabled, its hooks are merged with user and project hooks207* `mcp__filesystem__read_file`: Filesystem server's read file tool
104* Multiple hooks from different sources can respond to the same event208* `mcp__github__search_repositories`: GitHub server's search tool
105* Plugin hooks use the `${CLAUDE_PLUGIN_ROOT}` environment variable to reference plugin files
106 209
107**Example plugin hook configuration**:210Use regex patterns to target specific MCP tools or groups of tools:
211
212* `mcp__memory__.*` matches all tools from the `memory` server
213* `mcp__.*__write.*` matches any tool containing "write" from any server
214
215This example logs all memory server operations and validates write operations from any MCP server:
108 216
109```json theme={null}217```json theme={null}
110{218{
111 "description": "Automatic code formatting",
112 "hooks": {219 "hooks": {
113 "PostToolUse": [220 "PreToolUse": [
114 {221 {
115 "matcher": "Write|Edit",222 "matcher": "mcp__memory__.*",
116 "hooks": [223 "hooks": [
117 {224 {
118 "type": "command",225 "type": "command",
119 "command": "${CLAUDE_PLUGIN_ROOT}/scripts/format.sh",226 "command": "echo 'Memory operation initiated' >> ~/mcp-operations.log"
120 "timeout": 30227 }
228 ]
229 },
230 {
231 "matcher": "mcp__.*__write.*",
232 "hooks": [
233 {
234 "type": "command",
235 "command": "/home/user/scripts/validate-mcp-write.py"
121 }236 }
122 ]237 ]
123 }238 }
126}241}
127```242```
128 243
129<Note>244### Hook handler fields
130 Plugin hooks use the same format as regular hooks with an optional `description` field to explain the hook's purpose.
131</Note>
132 245
133<Note>246Each object in the inner `hooks` array is a hook handler: the shell command, HTTP endpoint, LLM prompt, or agent that runs when the matcher matches. There are four types:
134 Plugin hooks run alongside your custom hooks. If multiple hooks match an event, they all execute in parallel.
135</Note>
136 247
137**Environment variables for plugins**:248* **[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.
249* **[HTTP hooks](#http-hook-fields)** (`type: "http"`): send the event's JSON input as an HTTP POST request to a URL. The endpoint communicates results back through the response body using the same [JSON output format](#json-output) as command hooks.
250* **[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).
251* **[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).
138 252
139* `${CLAUDE_PLUGIN_ROOT}`: Absolute path to the plugin directory253#### Common fields
140* `${CLAUDE_PROJECT_DIR}`: Project root directory (same as for project hooks)
141* All standard environment variables are available
142 254
143See the [plugin components reference](/en/plugins-reference#hooks) for details on creating plugin hooks.255These fields apply to all hook types:
144 256
145## Prompt-Based Hooks257| Field | Required | Description |
258| :-------------- | :------- | :-------------------------------------------------------------------------------------------------------------------------------------------- |
259| `type` | yes | `"command"`, `"http"`, `"prompt"`, or `"agent"` |
260| `timeout` | no | Seconds before canceling. Defaults: 600 for command, 30 for prompt, 60 for agent |
261| `statusMessage` | no | Custom spinner message displayed while the hook runs |
262| `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) |
146 263
147In 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.264#### Command hook fields
148 265
149### How prompt-based hooks work266In addition to the [common fields](#common-fields), command hooks accept these fields:
150 267
151Instead of executing a bash command, prompt-based hooks:268| Field | Required | Description |
269| :-------- | :------- | :------------------------------------------------------------------------------------------------------------------ |
270| `command` | yes | Shell command to execute |
271| `async` | no | If `true`, runs in the background without blocking. See [Run hooks in the background](#run-hooks-in-the-background) |
152 272
1531. Send the hook input and your prompt to a fast LLM (Haiku)273#### HTTP hook fields
1542. The LLM responds with structured JSON containing a decision274
1553. Claude Code processes the decision automatically275In addition to the [common fields](#common-fields), HTTP hooks accept these fields:
276
277| Field | Required | Description |
278| :--------------- | :------- | :----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
279| `url` | yes | URL to send the POST request to |
280| `headers` | no | Additional HTTP headers as key-value pairs. Values support environment variable interpolation using `$VAR_NAME` or `${VAR_NAME}` syntax. Only variables listed in `allowedEnvVars` are resolved |
281| `allowedEnvVars` | no | List of environment variable names that may be interpolated into header values. References to unlisted variables are replaced with empty strings. Required for any env var interpolation to work |
156 282
157### Configuration283Claude Code sends the hook's [JSON input](#hook-input-and-output) as the POST request body with `Content-Type: application/json`. The response body uses the same [JSON output format](#json-output) as command hooks.
284
285Error handling differs from command hooks: non-2xx responses, connection failures, and timeouts all produce non-blocking errors that allow execution to continue. To block a tool call or deny a permission, return a 2xx response with a JSON body containing `decision: "block"` or a `hookSpecificOutput` with `permissionDecision: "deny"`.
286
287This example sends `PreToolUse` events to a local validation service, authenticating with a token from the `MY_TOKEN` environment variable:
158 288
159```json theme={null}289```json theme={null}
160{290{
161 "hooks": {291 "hooks": {
162 "Stop": [292 "PreToolUse": [
163 {293 {
294 "matcher": "Bash",
164 "hooks": [295 "hooks": [
165 {296 {
166 "type": "prompt",297 "type": "http",
167 "prompt": "Evaluate if Claude should stop: $ARGUMENTS. Check if all tasks are complete."298 "url": "http://localhost:8080/hooks/pre-tool-use",
299 "timeout": 30,
300 "headers": {
301 "Authorization": "Bearer $MY_TOKEN"
302 },
303 "allowedEnvVars": ["MY_TOKEN"]
168 }304 }
169 ]305 ]
170 }306 }
173}309}
174```310```
175 311
176**Fields:**312<Note>
177 313 HTTP hooks must be configured by editing settings JSON directly. The `/hooks` interactive menu only supports adding command hooks.
178* `type`: Must be `"prompt"`314</Note>
179* `prompt`: The prompt text to send to the LLM
180 * Use `$ARGUMENTS` as a placeholder for the hook input JSON
181 * If `$ARGUMENTS` is not present, input JSON is appended to the prompt
182* `timeout`: (Optional) Timeout in seconds (default: 30 seconds)
183
184### Response schema
185 315
186The LLM must respond with JSON containing:316#### Prompt and agent hook fields
187 317
188```json theme={null}318In addition to the [common fields](#common-fields), prompt and agent hooks accept these fields:
189{
190 "decision": "approve" | "block",
191 "reason": "Explanation for the decision",
192 "continue": false, // Optional: stops Claude entirely
193 "stopReason": "Message shown to user", // Optional: custom stop message
194 "systemMessage": "Warning or context" // Optional: shown to user
195}
196```
197 319
198**Response fields:**320| Field | Required | Description |
321| :------- | :------- | :------------------------------------------------------------------------------------------ |
322| `prompt` | yes | Prompt text to send to the model. Use `$ARGUMENTS` as a placeholder for the hook input JSON |
323| `model` | no | Model to use for evaluation. Defaults to a fast model |
199 324
200* `decision`: `"approve"` allows the action, `"block"` prevents it325All matching hooks run in parallel, and identical handlers are deduplicated automatically. Command hooks are deduplicated by command string, and HTTP hooks are deduplicated by URL. Handlers run in the current directory with Claude Code's environment. The `$CLAUDE_CODE_REMOTE` environment variable is set to `"true"` in remote web environments and not set in the local CLI.
201* `reason`: Explanation shown to Claude when decision is `"block"`
202* `continue`: (Optional) If `false`, stops Claude's execution entirely
203* `stopReason`: (Optional) Message shown when `continue` is false
204* `systemMessage`: (Optional) Additional message shown to the user
205 326
206### Supported hook events327### Reference scripts by path
207 328
208Prompt-based hooks work with any hook event, but are most useful for:329Use environment variables to reference hook scripts relative to the project or plugin root, regardless of the working directory when the hook runs:
209 330
210* **Stop**: Intelligently decide if Claude should continue working331* `$CLAUDE_PROJECT_DIR`: the project root. Wrap in quotes to handle paths with spaces.
211* **SubagentStop**: Evaluate if a subagent has completed its task332* `${CLAUDE_PLUGIN_ROOT}`: the plugin's root directory, for scripts bundled with a [plugin](/en/plugins).
212* **UserPromptSubmit**: Validate user prompts with LLM assistance
213* **PreToolUse**: Make context-aware permission decisions
214* **PermissionRequest**: Intelligently allow or deny permission dialogs
215 333
216### Example: Intelligent Stop hook334<Tabs>
335 <Tab title="Project scripts">
336 This example uses `$CLAUDE_PROJECT_DIR` to run a style checker from the project's `.claude/hooks/` directory after any `Write` or `Edit` tool call:
217 337
218```json theme={null}338 ```json theme={null}
219{339 {
220 "hooks": {340 "hooks": {
221 "Stop": [341 "PostToolUse": [
222 {342 {
343 "matcher": "Write|Edit",
223 "hooks": [344 "hooks": [
224 {345 {
225 "type": "prompt",346 "type": "command",
226 "prompt": "You are evaluating whether Claude should stop working. Context: $ARGUMENTS\n\nAnalyze the conversation and determine if:\n1. All user-requested tasks are complete\n2. Any errors need to be addressed\n3. Follow-up work is needed\n\nRespond with JSON: {\"decision\": \"approve\" or \"block\", \"reason\": \"your explanation\"}",347 "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/check-style.sh"
227 "timeout": 30
228 }348 }
229 ]349 ]
230 }350 }
231 ]351 ]
232 }352 }
233}353 }
234```354 ```
355 </Tab>
235 356
236### Example: SubagentStop with custom logic357 <Tab title="Plugin scripts">
358 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.
237 359
238```json theme={null}360 This example runs a formatting script bundled with the plugin:
239{361
362 ```json theme={null}
363 {
364 "description": "Automatic code formatting",
240 "hooks": {365 "hooks": {
241 "SubagentStop": [366 "PostToolUse": [
242 {367 {
368 "matcher": "Write|Edit",
243 "hooks": [369 "hooks": [
244 {370 {
245 "type": "prompt",371 "type": "command",
246 "prompt": "Evaluate if this subagent should stop. Input: $ARGUMENTS\n\nCheck if:\n- The subagent completed its assigned task\n- Any errors occurred that need fixing\n- Additional context gathering is needed\n\nReturn: {\"decision\": \"approve\" or \"block\", \"reason\": \"explanation\"}"372 "command": "${CLAUDE_PLUGIN_ROOT}/scripts/format.sh",
373 "timeout": 30
247 }374 }
248 ]375 ]
249 }376 }
250 ]377 ]
251 }378 }
379 }
380 ```
381
382 See the [plugin components reference](/en/plugins-reference#hooks) for details on creating plugin hooks.
383 </Tab>
384</Tabs>
385
386### Hooks in skills and agents
387
388In 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.
389
390All hook events are supported. For subagents, `Stop` hooks are automatically converted to `SubagentStop` since that is the event that fires when a subagent completes.
391
392Hooks use the same configuration format as settings-based hooks but are scoped to the component's lifetime and cleaned up when it finishes.
393
394This skill defines a `PreToolUse` hook that runs a security validation script before each `Bash` command:
395
396```yaml theme={null}
397---
398name: secure-operations
399description: Perform operations with security checks
400hooks:
401 PreToolUse:
402 - matcher: "Bash"
403 hooks:
404 - type: command
405 command: "./scripts/security-check.sh"
406---
407```
408
409Agents use the same format in their YAML frontmatter.
410
411### The `/hooks` menu
412
413Type `/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.
414
415Each hook in the menu is labeled with a bracket prefix indicating its source:
416
417* `[User]`: from `~/.claude/settings.json`
418* `[Project]`: from `.claude/settings.json`
419* `[Local]`: from `.claude/settings.local.json`
420* `[Plugin]`: from a plugin's `hooks/hooks.json`, read-only
421
422### Disable or remove hooks
423
424To remove a hook, delete its entry from the settings JSON file, or use the `/hooks` menu and select the hook to delete it.
425
426To 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.
427
428The `disableAllHooks` setting respects the managed settings hierarchy. If an administrator has configured hooks through managed policy settings, `disableAllHooks` set in user, project, or local settings cannot disable those managed hooks. Only `disableAllHooks` set at the managed settings level can disable managed hooks.
429
430Direct 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.
431
432## Hook input and output
433
434Command hooks receive JSON data via stdin and communicate results through exit codes, stdout, and stderr. HTTP hooks receive the same JSON as the POST request body and communicate results through the HTTP response body. 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.
435
436### Common input fields
437
438All hook events receive these fields as JSON, in addition to event-specific fields documented in each [hook event](#hook-events) section. For command hooks, this JSON arrives via stdin. For HTTP hooks, it arrives as the POST request body.
439
440| Field | Description |
441| :---------------- | :----------------------------------------------------------------------------------------------------------------------------------------- |
442| `session_id` | Current session identifier |
443| `transcript_path` | Path to conversation JSON |
444| `cwd` | Current working directory when the hook is invoked |
445| `permission_mode` | Current [permission mode](/en/permissions#permission-modes): `"default"`, `"plan"`, `"acceptEdits"`, `"dontAsk"`, or `"bypassPermissions"` |
446| `hook_event_name` | Name of the event that fired |
447
448For example, a `PreToolUse` hook for a Bash command receives this on stdin:
449
450```json theme={null}
451{
452 "session_id": "abc123",
453 "transcript_path": "/home/user/.claude/projects/.../transcript.jsonl",
454 "cwd": "/home/user/my-project",
455 "permission_mode": "default",
456 "hook_event_name": "PreToolUse",
457 "tool_name": "Bash",
458 "tool_input": {
459 "command": "npm test"
460 }
252}461}
253```462```
254 463
255### Comparison with bash command hooks464The `tool_name` and `tool_input` fields are event-specific. Each [hook event](#hook-events) section documents the additional fields for that event.
256 465
257| Feature | Bash Command Hooks | Prompt-Based Hooks |466### Exit code output
258| --------------------- | ----------------------- | ------------------------------ |
259| **Execution** | Runs bash script | Queries LLM |
260| **Decision logic** | You implement in code | LLM evaluates context |
261| **Setup complexity** | Requires script file | Just configure prompt |
262| **Context awareness** | Limited to script logic | Natural language understanding |
263| **Performance** | Fast (local execution) | Slower (API call) |
264| **Use case** | Deterministic rules | Context-aware decisions |
265 467
266### Best practices468The exit code from your hook command tells Claude Code whether the action should proceed, be blocked, or be ignored.
267 469
268* **Be specific in prompts**: Clearly state what you want the LLM to evaluate470**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.
269* **Include decision criteria**: List the factors the LLM should consider
270* **Test your prompts**: Verify the LLM makes correct decisions for your use cases
271* **Set appropriate timeouts**: Default is 30 seconds, adjust if needed
272* **Use for complex decisions**: Bash hooks are better for simple, deterministic rules
273 471
274See the [plugin components reference](/en/plugins-reference#hooks) for details on creating plugin hooks.472**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.
275 473
276## Hook Events474**Any other exit code** is a non-blocking error. stderr is shown in verbose mode (`Ctrl+O`) and execution continues.
277 475
278### PreToolUse476For example, a hook command script that blocks dangerous Bash commands:
477
478```bash theme={null}
479#!/bin/bash
480# Reads JSON input from stdin, checks the command
481command=$(jq -r '.tool_input.command' < /dev/stdin)
279 482
280Runs after Claude creates tool parameters and before processing the tool call.483if [[ "$command" == rm* ]]; then
484 echo "Blocked: rm commands are not allowed" >&2
485 exit 2 # Blocking error: tool call is prevented
486fi
281 487
282**Common matchers:**488exit 0 # Success: tool call proceeds
489```
283 490
284* `Task` - Subagent tasks (see [subagents documentation](/en/sub-agents))491#### Exit code 2 behavior per event
285* `Bash` - Shell commands
286* `Glob` - File pattern matching
287* `Grep` - Content search
288* `Read` - File reading
289* `Edit` - File editing
290* `Write` - File writing
291* `WebFetch`, `WebSearch` - Web operations
292 492
293Use [PreToolUse decision control](#pretooluse-decision-control) to allow, deny, or ask for permission to use the tool.493Exit 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.
294 494
295### PermissionRequest495| Hook event | Can block? | What happens on exit 2 |
496| :------------------- | :--------- | :---------------------------------------------------------------------------- |
497| `PreToolUse` | Yes | Blocks the tool call |
498| `PermissionRequest` | Yes | Denies the permission |
499| `UserPromptSubmit` | Yes | Blocks prompt processing and erases the prompt |
500| `Stop` | Yes | Prevents Claude from stopping, continues the conversation |
501| `SubagentStop` | Yes | Prevents the subagent from stopping |
502| `TeammateIdle` | Yes | Prevents the teammate from going idle (teammate continues working) |
503| `TaskCompleted` | Yes | Prevents the task from being marked as completed |
504| `ConfigChange` | Yes | Blocks the configuration change from taking effect (except `policy_settings`) |
505| `PostToolUse` | No | Shows stderr to Claude (tool already ran) |
506| `PostToolUseFailure` | No | Shows stderr to Claude (tool already failed) |
507| `Notification` | No | Shows stderr to user only |
508| `SubagentStart` | No | Shows stderr to user only |
509| `SessionStart` | No | Shows stderr to user only |
510| `SessionEnd` | No | Shows stderr to user only |
511| `PreCompact` | No | Shows stderr to user only |
512| `WorktreeCreate` | Yes | Any non-zero exit code causes worktree creation to fail |
513| `WorktreeRemove` | No | Failures are logged in debug mode only |
296 514
297Runs when the user is shown a permission dialog.515### HTTP response handling
298Use [PermissionRequest decision control](#permissionrequest-decision-control) to allow or deny on behalf of the user.
299 516
300Recognizes the same matcher values as PreToolUse.517HTTP hooks use HTTP status codes and response bodies instead of exit codes and stdout:
301 518
302### PostToolUse519* **2xx with an empty body**: success, equivalent to exit code 0 with no output
520* **2xx with a plain text body**: success, the text is added as context
521* **2xx with a JSON body**: success, parsed using the same [JSON output](#json-output) schema as command hooks
522* **Non-2xx status**: non-blocking error, execution continues
523* **Connection failure or timeout**: non-blocking error, execution continues
303 524
304Runs immediately after a tool completes successfully.525Unlike command hooks, HTTP hooks cannot signal a blocking error through status codes alone. To block a tool call or deny a permission, return a 2xx response with a JSON body containing the appropriate decision fields.
305 526
306Recognizes the same matcher values as PreToolUse.527### JSON output
307 528
308### Notification529Exit 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.
530
531<Note>
532 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.
533</Note>
309 534
310Runs when Claude Code sends notifications. Supports matchers to filter by notification type.535Your 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.
311 536
312**Common matchers:**537The JSON object supports three kinds of fields:
313 538
314* `permission_prompt` - Permission requests from Claude Code539* **Universal fields** like `continue` work across all events. These are listed in the table below.
315* `idle_prompt` - When Claude is waiting for user input (after 60+ seconds of idle time)540* **Top-level `decision` and `reason`** are used by some events to block or provide feedback.
316* `auth_success` - Authentication success notifications541* **`hookSpecificOutput`** is a nested object for events that need richer control. It requires a `hookEventName` field set to the event name.
317* `elicitation_dialog` - When Claude Code needs input for MCP tool elicitation
318 542
319You can use matchers to run different hooks for different notification types, or omit the matcher to run hooks for all notifications.543| Field | Default | Description |
544| :--------------- | :------ | :------------------------------------------------------------------------------------------------------------------------- |
545| `continue` | `true` | If `false`, Claude stops processing entirely after the hook runs. Takes precedence over any event-specific decision fields |
546| `stopReason` | none | Message shown to the user when `continue` is `false`. Not shown to Claude |
547| `suppressOutput` | `false` | If `true`, hides stdout from verbose mode output |
548| `systemMessage` | none | Warning message shown to the user |
320 549
321**Example: Different notifications for different types**550To stop Claude entirely regardless of event type:
322 551
323```json theme={null}552```json theme={null}
324{553{ "continue": false, "stopReason": "Build failed, fix errors before continuing" }
325 "hooks": {554```
326 "Notification": [555
327 {556#### Decision control
328 "matcher": "permission_prompt",557
329 "hooks": [558Not 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:
559
560| Events | Decision pattern | Key fields |
561| :---------------------------------------------------------------------------------- | :------------------- | :-------------------------------------------------------------------------- |
562| UserPromptSubmit, PostToolUse, PostToolUseFailure, Stop, SubagentStop, ConfigChange | Top-level `decision` | `decision: "block"`, `reason` |
563| TeammateIdle, TaskCompleted | Exit code only | Exit code 2 blocks the action, stderr is fed back as feedback |
564| PreToolUse | `hookSpecificOutput` | `permissionDecision` (allow/deny/ask), `permissionDecisionReason` |
565| PermissionRequest | `hookSpecificOutput` | `decision.behavior` (allow/deny) |
566| WorktreeCreate | stdout path | Hook prints absolute path to created worktree. Non-zero exit fails creation |
567| WorktreeRemove, Notification, SessionEnd, PreCompact | None | No decision control. Used for side effects like logging or cleanup |
568
569Here are examples of each pattern in action:
570
571<Tabs>
572 <Tab title="Top-level decision">
573 Used by `UserPromptSubmit`, `PostToolUse`, `PostToolUseFailure`, `Stop`, `SubagentStop`, and `ConfigChange`. The only value is `"block"`. To allow the action to proceed, omit `decision` from your JSON, or exit 0 without any JSON at all:
574
575 ```json theme={null}
330 {576 {
331 "type": "command",577 "decision": "block",
332 "command": "/path/to/permission-alert.sh"578 "reason": "Test suite must pass before proceeding"
333 }579 }
334 ]580 ```
335 },581 </Tab>
582
583 <Tab title="PreToolUse">
584 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.
585
586 ```json theme={null}
336 {587 {
337 "matcher": "idle_prompt",588 "hookSpecificOutput": {
338 "hooks": [589 "hookEventName": "PreToolUse",
590 "permissionDecision": "deny",
591 "permissionDecisionReason": "Database writes are not allowed"
592 }
593 }
594 ```
595 </Tab>
596
597 <Tab title="PermissionRequest">
598 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.
599
600 ```json theme={null}
339 {601 {
340 "type": "command",602 "hookSpecificOutput": {
341 "command": "/path/to/idle-notification.sh"603 "hookEventName": "PermissionRequest",
604 "decision": {
605 "behavior": "allow",
606 "updatedInput": {
607 "command": "npm run lint"
342 }608 }
343 ]
344 }609 }
345 ]
346 }610 }
347}611 }
348```612 ```
613 </Tab>
614</Tabs>
349 615
350### UserPromptSubmit616For 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).
351 617
352Runs when the user submits a prompt, before Claude processes it. This allows you618## Hook events
353to add additional context based on the prompt/conversation, validate prompts, or
354block certain types of prompts.
355 619
356### Stop620Each 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.
357 621
358Runs when the main Claude Code agent has finished responding. Does not run if622### SessionStart
359the stoppage occurred due to a user interrupt.
360 623
361### SubagentStop624Runs 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.
362 625
363Runs when a Claude Code subagent (Task tool call) has finished responding.626SessionStart runs on every session, so keep these hooks fast.
364 627
365### PreCompact628The matcher value corresponds to how the session was initiated:
366 629
367Runs before Claude Code is about to run a compact operation.630| Matcher | When it fires |
631| :-------- | :------------------------------------- |
632| `startup` | New session |
633| `resume` | `--resume`, `--continue`, or `/resume` |
634| `clear` | `/clear` |
635| `compact` | Auto or manual compaction |
368 636
369**Matchers:**637#### SessionStart input
370 638
371* `manual` - Invoked from `/compact`639In 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.
372* `auto` - Invoked from auto-compact (due to full context window)
373 640
374### SessionStart641```json theme={null}
642{
643 "session_id": "abc123",
644 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
645 "cwd": "/Users/...",
646 "permission_mode": "default",
647 "hook_event_name": "SessionStart",
648 "source": "startup",
649 "model": "claude-sonnet-4-6"
650}
651```
375 652
376Runs when Claude Code starts a new session or resumes an existing session (which653#### SessionStart decision control
377currently does start a new session under the hood). Useful for loading in
378development context like existing issues or recent changes to your codebase, installing dependencies, or setting up environment variables.
379 654
380**Matchers:**655Any 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:
381 656
382* `startup` - Invoked from startup657| Field | Description |
383* `resume` - Invoked from `--resume`, `--continue`, or `/resume`658| :------------------ | :------------------------------------------------------------------------ |
384* `clear` - Invoked from `/clear`659| `additionalContext` | String added to Claude's context. Multiple hooks' values are concatenated |
385* `compact` - Invoked from auto or manual compact.660
661```json theme={null}
662{
663 "hookSpecificOutput": {
664 "hookEventName": "SessionStart",
665 "additionalContext": "My additional context here"
666 }
667}
668```
386 669
387#### Persisting environment variables670#### Persist environment variables
388 671
389SessionStart 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.672SessionStart 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.
390 673
391**Example: Setting individual environment variables**674To set individual environment variables, write `export` statements to `CLAUDE_ENV_FILE`. Use append (`>>`) to preserve variables set by other hooks:
392 675
393```bash theme={null}676```bash theme={null}
394#!/bin/bash677#!/bin/bash
395 678
396if [ -n "$CLAUDE_ENV_FILE" ]; then679if [ -n "$CLAUDE_ENV_FILE" ]; then
397 echo 'export NODE_ENV=production' >> "$CLAUDE_ENV_FILE"680 echo 'export NODE_ENV=production' >> "$CLAUDE_ENV_FILE"
398 echo 'export API_KEY=your-api-key' >> "$CLAUDE_ENV_FILE"681 echo 'export DEBUG_LOG=true' >> "$CLAUDE_ENV_FILE"
399 echo 'export PATH="$PATH:./node_modules/.bin"' >> "$CLAUDE_ENV_FILE"682 echo 'export PATH="$PATH:./node_modules/.bin"' >> "$CLAUDE_ENV_FILE"
400fi683fi
401 684
402exit 0685exit 0
403```686```
404 687
405**Example: Persisting all environment changes from the hook**688To capture all environment changes from setup commands, compare the exported variables before and after:
406
407When your setup modifies the environment (e.g., `nvm use`), capture and persist all changes by diffing the environment:
408 689
409```bash theme={null}690```bash theme={null}
410#!/bin/bash691#!/bin/bash
423exit 0704exit 0
424```705```
425 706
426Any variables written to this file will be available in all subsequent bash commands that Claude Code executes during the session.707Any variables written to this file will be available in all subsequent Bash commands that Claude Code executes during the session.
427 708
428<Note>709<Note>
429 `CLAUDE_ENV_FILE` is only available for SessionStart hooks. Other hook types do not have access to this variable.710 `CLAUDE_ENV_FILE` is available for SessionStart hooks. Other hook types do not have access to this variable.
430</Note>711</Note>
431 712
432### SessionEnd713### UserPromptSubmit
433
434Runs when a Claude Code session ends. Useful for cleanup tasks, logging session
435statistics, or saving session state.
436
437The `reason` field in the hook input will be one of:
438 714
439* `clear` - Session cleared with /clear command715Runs when the user submits a prompt, before Claude processes it. This allows you
440* `logout` - User logged out716to add additional context based on the prompt/conversation, validate prompts, or
441* `prompt_input_exit` - User exited while prompt input was visible717block certain types of prompts.
442* `other` - Other exit reasons
443 718
444## Hook Input719#### UserPromptSubmit input
445 720
446Hooks receive JSON data via stdin containing session information and721In addition to the [common input fields](#common-input-fields), UserPromptSubmit hooks receive the `prompt` field containing the text the user submitted.
447event-specific data:
448 722
449```typescript theme={null}723```json theme={null}
450{724{
451 // Common fields725 "session_id": "abc123",
452 session_id: string726 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
453 transcript_path: string // Path to conversation JSON727 "cwd": "/Users/...",
454 cwd: string // The current working directory when the hook is invoked728 "permission_mode": "default",
455 permission_mode: string // Current permission mode: "default", "plan", "acceptEdits", or "bypassPermissions"729 "hook_event_name": "UserPromptSubmit",
456 730 "prompt": "Write a function to calculate the factorial of a number"
457 // Event-specific fields731}
458 hook_event_name: string732```
459 ...
460}
461```
462
463### PreToolUse Input
464 733
465The exact schema for `tool_input` depends on the tool.734#### UserPromptSubmit decision control
735
736`UserPromptSubmit` hooks can control whether a user prompt is processed and add context. All [JSON output fields](#json-output) are available.
737
738There are two ways to add context to the conversation on exit code 0:
739
740* **Plain text stdout**: any non-JSON text written to stdout is added as context
741* **JSON with `additionalContext`**: use the JSON format below for more control. The `additionalContext` field is added as context
742
743Plain stdout is shown as hook output in the transcript. The `additionalContext` field is added more discretely.
744
745To block a prompt, return a JSON object with `decision` set to `"block"`:
746
747| Field | Description |
748| :------------------ | :----------------------------------------------------------------------------------------------------------------- |
749| `decision` | `"block"` prevents the prompt from being processed and erases it from context. Omit to allow the prompt to proceed |
750| `reason` | Shown to the user when `decision` is `"block"`. Not added to context |
751| `additionalContext` | String added to Claude's context |
752
753```json theme={null}
754{
755 "decision": "block",
756 "reason": "Explanation for decision",
757 "hookSpecificOutput": {
758 "hookEventName": "UserPromptSubmit",
759 "additionalContext": "My additional context here"
760 }
761}
762```
763
764<Note>
765 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
766 block prompts or want more structured control.
767</Note>
768
769### PreToolUse
770
771Runs 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).
772
773Use [PreToolUse decision control](#pretooluse-decision-control) to allow, deny, or ask for permission to use the tool.
774
775#### PreToolUse input
776
777In 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:
778
779##### Bash
780
781Executes shell commands.
782
783| Field | Type | Example | Description |
784| :------------------ | :------ | :----------------- | :-------------------------------------------- |
785| `command` | string | `"npm test"` | The shell command to execute |
786| `description` | string | `"Run test suite"` | Optional description of what the command does |
787| `timeout` | number | `120000` | Optional timeout in milliseconds |
788| `run_in_background` | boolean | `false` | Whether to run the command in background |
789
790##### Write
791
792Creates or overwrites a file.
793
794| Field | Type | Example | Description |
795| :---------- | :----- | :-------------------- | :--------------------------------- |
796| `file_path` | string | `"/path/to/file.txt"` | Absolute path to the file to write |
797| `content` | string | `"file content"` | Content to write to the file |
798
799##### Edit
800
801Replaces a string in an existing file.
802
803| Field | Type | Example | Description |
804| :------------ | :------ | :-------------------- | :--------------------------------- |
805| `file_path` | string | `"/path/to/file.txt"` | Absolute path to the file to edit |
806| `old_string` | string | `"original text"` | Text to find and replace |
807| `new_string` | string | `"replacement text"` | Replacement text |
808| `replace_all` | boolean | `false` | Whether to replace all occurrences |
809
810##### Read
811
812Reads file contents.
813
814| Field | Type | Example | Description |
815| :---------- | :----- | :-------------------- | :----------------------------------------- |
816| `file_path` | string | `"/path/to/file.txt"` | Absolute path to the file to read |
817| `offset` | number | `10` | Optional line number to start reading from |
818| `limit` | number | `50` | Optional number of lines to read |
819
820##### Glob
821
822Finds files matching a glob pattern.
823
824| Field | Type | Example | Description |
825| :-------- | :----- | :--------------- | :--------------------------------------------------------------------- |
826| `pattern` | string | `"**/*.ts"` | Glob pattern to match files against |
827| `path` | string | `"/path/to/dir"` | Optional directory to search in. Defaults to current working directory |
828
829##### Grep
830
831Searches file contents with regular expressions.
832
833| Field | Type | Example | Description |
834| :------------ | :------ | :--------------- | :------------------------------------------------------------------------------------ |
835| `pattern` | string | `"TODO.*fix"` | Regular expression pattern to search for |
836| `path` | string | `"/path/to/dir"` | Optional file or directory to search in |
837| `glob` | string | `"*.ts"` | Optional glob pattern to filter files |
838| `output_mode` | string | `"content"` | `"content"`, `"files_with_matches"`, or `"count"`. Defaults to `"files_with_matches"` |
839| `-i` | boolean | `true` | Case insensitive search |
840| `multiline` | boolean | `false` | Enable multiline matching |
841
842##### WebFetch
843
844Fetches and processes web content.
845
846| Field | Type | Example | Description |
847| :------- | :----- | :---------------------------- | :----------------------------------- |
848| `url` | string | `"https://example.com/api"` | URL to fetch content from |
849| `prompt` | string | `"Extract the API endpoints"` | Prompt to run on the fetched content |
850
851##### WebSearch
852
853Searches the web.
854
855| Field | Type | Example | Description |
856| :---------------- | :----- | :----------------------------- | :------------------------------------------------ |
857| `query` | string | `"react hooks best practices"` | Search query |
858| `allowed_domains` | array | `["docs.example.com"]` | Optional: only include results from these domains |
859| `blocked_domains` | array | `["spam.example.com"]` | Optional: exclude results from these domains |
860
861##### Task
862
863Spawns a [subagent](/en/sub-agents).
864
865| Field | Type | Example | Description |
866| :-------------- | :----- | :------------------------- | :------------------------------------------- |
867| `prompt` | string | `"Find all API endpoints"` | The task for the agent to perform |
868| `description` | string | `"Find API endpoints"` | Short description of the task |
869| `subagent_type` | string | `"Explore"` | Type of specialized agent to use |
870| `model` | string | `"sonnet"` | Optional model alias to override the default |
871
872#### PreToolUse decision control
873
874`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.
875
876| Field | Description |
877| :------------------------- | :----------------------------------------------------------------------------------------------------------------------------------------------- |
878| `permissionDecision` | `"allow"` bypasses the permission system, `"deny"` prevents the tool call, `"ask"` prompts the user to confirm |
879| `permissionDecisionReason` | For `"allow"` and `"ask"`, shown to the user but not Claude. For `"deny"`, shown to Claude |
880| `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 |
881| `additionalContext` | String added to Claude's context before the tool executes |
882
883```json theme={null}
884{
885 "hookSpecificOutput": {
886 "hookEventName": "PreToolUse",
887 "permissionDecision": "allow",
888 "permissionDecisionReason": "My reason here",
889 "updatedInput": {
890 "field_to_modify": "new value"
891 },
892 "additionalContext": "Current environment: production. Proceed with caution."
893 }
894}
895```
896
897<Note>
898 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.
899</Note>
900
901### PermissionRequest
902
903Runs when the user is shown a permission dialog.
904Use [PermissionRequest decision control](#permissionrequest-decision-control) to allow or deny on behalf of the user.
905
906Matches on tool name, same values as PreToolUse.
907
908#### PermissionRequest input
909
910PermissionRequest 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.
466 911
467```json theme={null}912```json theme={null}
468{913{
470 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",915 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
471 "cwd": "/Users/...",916 "cwd": "/Users/...",
472 "permission_mode": "default",917 "permission_mode": "default",
473 "hook_event_name": "PreToolUse",918 "hook_event_name": "PermissionRequest",
474 "tool_name": "Write",919 "tool_name": "Bash",
475 "tool_input": {920 "tool_input": {
476 "file_path": "/path/to/file.txt",921 "command": "rm -rf node_modules",
477 "content": "file content"922 "description": "Remove node_modules directory"
478 },923 },
479 "tool_use_id": "toolu_01ABC123..."924 "permission_suggestions": [
925 { "type": "toolAlwaysAllow", "tool": "Bash" }
926 ]
927}
928```
929
930#### PermissionRequest decision control
931
932`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:
933
934| Field | Description |
935| :------------------- | :------------------------------------------------------------------------------------------------------------- |
936| `behavior` | `"allow"` grants the permission, `"deny"` denies it |
937| `updatedInput` | For `"allow"` only: modifies the tool's input parameters before execution |
938| `updatedPermissions` | For `"allow"` only: applies permission rule updates, equivalent to the user selecting an "always allow" option |
939| `message` | For `"deny"` only: tells Claude why the permission was denied |
940| `interrupt` | For `"deny"` only: if `true`, stops Claude |
941
942```json theme={null}
943{
944 "hookSpecificOutput": {
945 "hookEventName": "PermissionRequest",
946 "decision": {
947 "behavior": "allow",
948 "updatedInput": {
949 "command": "npm run lint"
950 }
951 }
952 }
480}953}
481```954```
482 955
483### PostToolUse Input956### PostToolUse
957
958Runs immediately after a tool completes successfully.
959
960Matches on tool name, same values as PreToolUse.
484 961
485The exact schema for `tool_input` and `tool_response` depends on the tool.962#### PostToolUse input
963
964`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.
486 965
487```json theme={null}966```json theme={null}
488{967{
504}983}
505```984```
506 985
507### Notification Input986#### PostToolUse decision control
987
988`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:
989
990| Field | Description |
991| :--------------------- | :----------------------------------------------------------------------------------------- |
992| `decision` | `"block"` prompts Claude with the `reason`. Omit to allow the action to proceed |
993| `reason` | Explanation shown to Claude when `decision` is `"block"` |
994| `additionalContext` | Additional context for Claude to consider |
995| `updatedMCPToolOutput` | For [MCP tools](#match-mcp-tools) only: replaces the tool's output with the provided value |
996
997```json theme={null}
998{
999 "decision": "block",
1000 "reason": "Explanation for decision",
1001 "hookSpecificOutput": {
1002 "hookEventName": "PostToolUse",
1003 "additionalContext": "Additional information for Claude"
1004 }
1005}
1006```
1007
1008### PostToolUseFailure
1009
1010Runs 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.
1011
1012Matches on tool name, same values as PreToolUse.
1013
1014#### PostToolUseFailure input
1015
1016PostToolUseFailure hooks receive the same `tool_name` and `tool_input` fields as PostToolUse, along with error information as top-level fields:
1017
1018```json theme={null}
1019{
1020 "session_id": "abc123",
1021 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
1022 "cwd": "/Users/...",
1023 "permission_mode": "default",
1024 "hook_event_name": "PostToolUseFailure",
1025 "tool_name": "Bash",
1026 "tool_input": {
1027 "command": "npm test",
1028 "description": "Run test suite"
1029 },
1030 "tool_use_id": "toolu_01ABC123...",
1031 "error": "Command exited with non-zero status code 1",
1032 "is_interrupt": false
1033}
1034```
1035
1036| Field | Description |
1037| :------------- | :------------------------------------------------------------------------------ |
1038| `error` | String describing what went wrong |
1039| `is_interrupt` | Optional boolean indicating whether the failure was caused by user interruption |
1040
1041#### PostToolUseFailure decision control
1042
1043`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:
1044
1045| Field | Description |
1046| :------------------ | :------------------------------------------------------------ |
1047| `additionalContext` | Additional context for Claude to consider alongside the error |
1048
1049```json theme={null}
1050{
1051 "hookSpecificOutput": {
1052 "hookEventName": "PostToolUseFailure",
1053 "additionalContext": "Additional information about the failure for Claude"
1054 }
1055}
1056```
1057
1058### Notification
1059
1060Runs 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.
1061
1062Use 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:
1063
1064```json theme={null}
1065{
1066 "hooks": {
1067 "Notification": [
1068 {
1069 "matcher": "permission_prompt",
1070 "hooks": [
1071 {
1072 "type": "command",
1073 "command": "/path/to/permission-alert.sh"
1074 }
1075 ]
1076 },
1077 {
1078 "matcher": "idle_prompt",
1079 "hooks": [
1080 {
1081 "type": "command",
1082 "command": "/path/to/idle-notification.sh"
1083 }
1084 ]
1085 }
1086 ]
1087 }
1088}
1089```
1090
1091#### Notification input
1092
1093In 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.
508 1094
509```json theme={null}1095```json theme={null}
510{1096{
514 "permission_mode": "default",1100 "permission_mode": "default",
515 "hook_event_name": "Notification",1101 "hook_event_name": "Notification",
516 "message": "Claude needs your permission to use Bash",1102 "message": "Claude needs your permission to use Bash",
1103 "title": "Permission needed",
517 "notification_type": "permission_prompt"1104 "notification_type": "permission_prompt"
518}1105}
519```1106```
520 1107
521### UserPromptSubmit Input1108Notification 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:
1109
1110| Field | Description |
1111| :------------------ | :------------------------------- |
1112| `additionalContext` | String added to Claude's context |
1113
1114### SubagentStart
1115
1116Runs 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/`).
1117
1118#### SubagentStart input
1119
1120In 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).
522 1121
523```json theme={null}1122```json theme={null}
524{1123{
526 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1125 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
527 "cwd": "/Users/...",1126 "cwd": "/Users/...",
528 "permission_mode": "default",1127 "permission_mode": "default",
529 "hook_event_name": "UserPromptSubmit",1128 "hook_event_name": "SubagentStart",
530 "prompt": "Write a function to calculate the factorial of a number"1129 "agent_id": "agent-abc123",
1130 "agent_type": "Explore"
531}1131}
532```1132```
533 1133
534### Stop and SubagentStop Input1134SubagentStart 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:
535 1135
536`stop_hook_active` is true when Claude Code is already continuing as a result of1136| Field | Description |
537a stop hook. Check this value or process the transcript to prevent Claude Code1137| :------------------ | :------------------------------------- |
538from running indefinitely.1138| `additionalContext` | String added to the subagent's context |
1139
1140```json theme={null}
1141{
1142 "hookSpecificOutput": {
1143 "hookEventName": "SubagentStart",
1144 "additionalContext": "Follow security guidelines for this task"
1145 }
1146}
1147```
1148
1149### SubagentStop
1150
1151Runs when a Claude Code subagent has finished responding. Matches on agent type, same values as SubagentStart.
1152
1153#### SubagentStop input
1154
1155In addition to the [common input fields](#common-input-fields), SubagentStop hooks receive `stop_hook_active`, `agent_id`, `agent_type`, `agent_transcript_path`, and `last_assistant_message`. 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. The `last_assistant_message` field contains the text content of the subagent's final response, so hooks can access it without parsing the transcript file.
539 1156
540```json theme={null}1157```json theme={null}
541{1158{
542 "session_id": "abc123",1159 "session_id": "abc123",
543 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1160 "transcript_path": "~/.claude/projects/.../abc123.jsonl",
1161 "cwd": "/Users/...",
544 "permission_mode": "default",1162 "permission_mode": "default",
545 "hook_event_name": "Stop",1163 "hook_event_name": "SubagentStop",
546 "stop_hook_active": true1164 "stop_hook_active": false,
1165 "agent_id": "def456",
1166 "agent_type": "Explore",
1167 "agent_transcript_path": "~/.claude/projects/.../abc123/subagents/agent-def456.jsonl",
1168 "last_assistant_message": "Analysis complete. Found 3 potential issues..."
547}1169}
548```1170```
549 1171
550### PreCompact Input1172SubagentStop hooks use the same decision control format as [Stop hooks](#stop-decision-control).
551 1173
552For `manual`, `custom_instructions` comes from what the user passes into1174### Stop
553`/compact`. For `auto`, `custom_instructions` is empty.1175
1176Runs when the main Claude Code agent has finished responding. Does not run if
1177the stoppage occurred due to a user interrupt.
1178
1179#### Stop input
1180
1181In addition to the [common input fields](#common-input-fields), Stop hooks receive `stop_hook_active` and `last_assistant_message`. The `stop_hook_active` 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. The `last_assistant_message` field contains the text content of Claude's final response, so hooks can access it without parsing the transcript file.
554 1182
555```json theme={null}1183```json theme={null}
556{1184{
557 "session_id": "abc123",1185 "session_id": "abc123",
558 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1186 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
1187 "cwd": "/Users/...",
559 "permission_mode": "default",1188 "permission_mode": "default",
560 "hook_event_name": "PreCompact",1189 "hook_event_name": "Stop",
561 "trigger": "manual",1190 "stop_hook_active": true,
562 "custom_instructions": ""1191 "last_assistant_message": "I've completed the refactoring. Here's a summary..."
1192}
1193```
1194
1195#### Stop decision control
1196
1197`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:
1198
1199| Field | Description |
1200| :--------- | :------------------------------------------------------------------------- |
1201| `decision` | `"block"` prevents Claude from stopping. Omit to allow Claude to stop |
1202| `reason` | Required when `decision` is `"block"`. Tells Claude why it should continue |
1203
1204```json theme={null}
1205{
1206 "decision": "block",
1207 "reason": "Must be provided when Claude is blocked from stopping"
563}1208}
564```1209```
565 1210
566### SessionStart Input1211### TeammateIdle
1212
1213Runs 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.
1214
1215When 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.
1216
1217#### TeammateIdle input
1218
1219In addition to the [common input fields](#common-input-fields), TeammateIdle hooks receive `teammate_name` and `team_name`.
567 1220
568```json theme={null}1221```json theme={null}
569{1222{
570 "session_id": "abc123",1223 "session_id": "abc123",
571 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1224 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
1225 "cwd": "/Users/...",
572 "permission_mode": "default",1226 "permission_mode": "default",
573 "hook_event_name": "SessionStart",1227 "hook_event_name": "TeammateIdle",
574 "source": "startup"1228 "teammate_name": "researcher",
1229 "team_name": "my-project"
575}1230}
576```1231```
577 1232
578### SessionEnd Input1233| Field | Description |
1234| :-------------- | :-------------------------------------------- |
1235| `teammate_name` | Name of the teammate that is about to go idle |
1236| `team_name` | Name of the team |
1237
1238#### TeammateIdle decision control
1239
1240TeammateIdle hooks use exit codes only, not JSON decision control. This example checks that a build artifact exists before allowing a teammate to go idle:
1241
1242```bash theme={null}
1243#!/bin/bash
1244
1245if [ ! -f "./dist/output.js" ]; then
1246 echo "Build artifact missing. Run the build before stopping." >&2
1247 exit 2
1248fi
1249
1250exit 0
1251```
1252
1253### TaskCompleted
1254
1255Runs 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.
1256
1257When 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.
1258
1259#### TaskCompleted input
1260
1261In 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`.
579 1262
580```json theme={null}1263```json theme={null}
581{1264{
582 "session_id": "abc123",1265 "session_id": "abc123",
583 "transcript_path": "~/.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",1266 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
584 "cwd": "/Users/...",1267 "cwd": "/Users/...",
585 "permission_mode": "default",1268 "permission_mode": "default",
586 "hook_event_name": "SessionEnd",1269 "hook_event_name": "TaskCompleted",
587 "reason": "exit"1270 "task_id": "task-001",
1271 "task_subject": "Implement user authentication",
1272 "task_description": "Add login and signup endpoints",
1273 "teammate_name": "implementer",
1274 "team_name": "my-project"
588}1275}
589```1276```
590 1277
591## Hook Output1278| Field | Description |
1279| :----------------- | :------------------------------------------------------ |
1280| `task_id` | Identifier of the task being completed |
1281| `task_subject` | Title of the task |
1282| `task_description` | Detailed description of the task. May be absent |
1283| `teammate_name` | Name of the teammate completing the task. May be absent |
1284| `team_name` | Name of the team. May be absent |
592 1285
593There are two mutually-exclusive ways for hooks to return output back to Claude Code. The output1286#### TaskCompleted decision control
594communicates whether to block and any feedback that should be shown to Claude
595and the user.
596 1287
597### Simple: Exit Code1288TaskCompleted hooks use exit codes only, not JSON decision control. This example runs tests and blocks task completion if they fail:
598 1289
599Hooks communicate status through exit codes, stdout, and stderr:1290```bash theme={null}
600 1291#!/bin/bash
601* **Exit code 0**: Success. `stdout` is shown to the user in verbose mode1292INPUT=$(cat)
602 (ctrl+o), except for `UserPromptSubmit` and `SessionStart`, where stdout is1293TASK_SUBJECT=$(echo "$INPUT" | jq -r '.task_subject')
603 added to the context. JSON output in `stdout` is parsed for structured control
604 (see [Advanced: JSON Output](#advanced-json-output)).
605* **Exit code 2**: Blocking error. Only `stderr` is used as the error message
606 and fed back to Claude. The format is `[command]: {stderr}`. JSON in `stdout`
607 is **not** processed for exit code 2. See per-hook-event behavior below.
608* **Other exit codes**: Non-blocking error. `stderr` is shown to the user in verbose mode (ctrl+o) with
609 format `Failed with non-blocking status code: {stderr}`. If `stderr` is empty,
610 it shows `No stderr output`. Execution continues.
611 1294
612<Warning>1295# Run the test suite
613 Reminder: Claude Code does not see stdout if the exit code is 0, except for1296if ! npm test 2>&1; then
614 the `UserPromptSubmit` hook where stdout is injected as context.1297 echo "Tests not passing. Fix failing tests before completing: $TASK_SUBJECT" >&2
615</Warning>1298 exit 2
1299fi
616 1300
617#### Exit Code 2 Behavior1301exit 0
1302```
618 1303
619| Hook Event | Behavior |1304### ConfigChange
620| ------------------- | ------------------------------------------------------------------ |
621| `PreToolUse` | Blocks the tool call, shows stderr to Claude |
622| `PermissionRequest` | Denies the permission, shows stderr to Claude |
623| `PostToolUse` | Shows stderr to Claude (tool already ran) |
624| `Notification` | N/A, shows stderr to user only |
625| `UserPromptSubmit` | Blocks prompt processing, erases prompt, shows stderr to user only |
626| `Stop` | Blocks stoppage, shows stderr to Claude |
627| `SubagentStop` | Blocks stoppage, shows stderr to Claude subagent |
628| `PreCompact` | N/A, shows stderr to user only |
629| `SessionStart` | N/A, shows stderr to user only |
630| `SessionEnd` | N/A, shows stderr to user only |
631 1305
632### Advanced: JSON Output1306Runs when a configuration file changes during a session. Use this to audit settings changes, enforce security policies, or block unauthorized modifications to configuration files.
633 1307
634Hooks can return structured JSON in `stdout` for more sophisticated control.1308ConfigChange hooks fire for changes to settings files, managed policy settings, and skill files. The `source` field in the input tells you which type of configuration changed, and the optional `file_path` field provides the path to the changed file.
635 1309
636<Warning>1310The matcher filters on the configuration source:
637 JSON output is only processed when the hook exits with code 0. If your hook
638 exits with code 2 (blocking error), `stderr` text is used directly—any JSON in `stdout`
639 is ignored. For other non-zero exit codes, only `stderr` is shown to the user in verbose mode (ctrl+o).
640</Warning>
641 1311
642#### Common JSON Fields1312| Matcher | When it fires |
1313| :----------------- | :---------------------------------------- |
1314| `user_settings` | `~/.claude/settings.json` changes |
1315| `project_settings` | `.claude/settings.json` changes |
1316| `local_settings` | `.claude/settings.local.json` changes |
1317| `policy_settings` | Managed policy settings change |
1318| `skills` | A skill file in `.claude/skills/` changes |
643 1319
644All hook types can include these optional fields:1320This example logs all configuration changes for security auditing:
645 1321
646```json theme={null}1322```json theme={null}
647{1323{
648 "continue": true, // Whether Claude should continue after hook execution (default: true)1324 "hooks": {
649 "stopReason": "string", // Message shown when continue is false1325 "ConfigChange": [
1326 {
1327 "hooks": [
1328 {
1329 "type": "command",
1330 "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/audit-config-change.sh"
1331 }
1332 ]
1333 }
1334 ]
1335 }
1336}
1337```
1338
1339#### ConfigChange input
650 1340
651 "suppressOutput": true, // Hide stdout from transcript mode (default: false)1341In addition to the [common input fields](#common-input-fields), ConfigChange hooks receive `source` and optionally `file_path`. The `source` field indicates which configuration type changed, and `file_path` provides the path to the specific file that was modified.
652 "systemMessage": "string" // Optional warning message shown to the user1342
1343```json theme={null}
1344{
1345 "session_id": "abc123",
1346 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
1347 "cwd": "/Users/...",
1348 "permission_mode": "default",
1349 "hook_event_name": "ConfigChange",
1350 "source": "project_settings",
1351 "file_path": "/Users/.../my-project/.claude/settings.json"
653}1352}
654```1353```
655 1354
656If `continue` is false, Claude stops processing after the hooks run.1355#### ConfigChange decision control
657 1356
658* For `PreToolUse`, this is different from `"permissionDecision": "deny"`, which1357ConfigChange hooks can block configuration changes from taking effect. Use exit code 2 or a JSON `decision` to prevent the change. When blocked, the new settings are not applied to the running session.
659 only blocks a specific tool call and provides automatic feedback to Claude.
660* For `PostToolUse`, this is different from `"decision": "block"`, which
661 provides automated feedback to Claude.
662* For `UserPromptSubmit`, this prevents the prompt from being processed.
663* For `Stop` and `SubagentStop`, this takes precedence over any
664 `"decision": "block"` output.
665* In all cases, `"continue" = false` takes precedence over any
666 `"decision": "block"` output.
667 1358
668`stopReason` accompanies `continue` with a reason shown to the user, not shown1359| Field | Description |
669to Claude.1360| :--------- | :--------------------------------------------------------------------------------------- |
1361| `decision` | `"block"` prevents the configuration change from being applied. Omit to allow the change |
1362| `reason` | Explanation shown to the user when `decision` is `"block"` |
670 1363
671#### `PreToolUse` Decision Control1364```json theme={null}
1365{
1366 "decision": "block",
1367 "reason": "Configuration changes to project settings require admin approval"
1368}
1369```
672 1370
673`PreToolUse` hooks can control whether a tool call proceeds.1371`policy_settings` changes cannot be blocked. Hooks still fire for `policy_settings` sources, so you can use them for audit logging, but any blocking decision is ignored. This ensures enterprise-managed settings always take effect.
674 1372
675* `"allow"` bypasses the permission system. `permissionDecisionReason` is shown1373### WorktreeCreate
676 to the user but not to Claude.
677* `"deny"` prevents the tool call from executing. `permissionDecisionReason` is
678 shown to Claude.
679* `"ask"` asks the user to confirm the tool call in the UI.
680 `permissionDecisionReason` is shown to the user but not to Claude.
681 1374
682Additionally, hooks can modify tool inputs before execution using `updatedInput`:1375When you run `claude --worktree` or a [subagent uses `isolation: "worktree"`](/en/sub-agents#choose-the-subagent-scope), Claude Code creates an isolated working copy using `git worktree`. If you configure a WorktreeCreate hook, it replaces the default git behavior, letting you use a different version control system like SVN, Perforce, or Mercurial.
683 1376
684* `updatedInput` allows you to modify the tool's input parameters before the tool executes.1377The hook must print the absolute path to the created worktree directory on stdout. Claude Code uses this path as the working directory for the isolated session.
685* This is most useful with `"permissionDecision": "allow"` to modify and approve tool calls.1378
1379This example creates an SVN working copy and prints the path for Claude Code to use. Replace the repository URL with your own:
686 1380
687```json theme={null}1381```json theme={null}
688{1382{
689 "hookSpecificOutput": {1383 "hooks": {
690 "hookEventName": "PreToolUse",1384 "WorktreeCreate": [
691 "permissionDecision": "allow"1385 {
692 "permissionDecisionReason": "My reason here",1386 "hooks": [
693 "updatedInput": {1387 {
694 "field_to_modify": "new value"1388 "type": "command",
1389 "command": "bash -c 'NAME=$(jq -r .name); DIR=\"$HOME/.claude/worktrees/$NAME\"; svn checkout https://svn.example.com/repo/trunk \"$DIR\" >&2 && echo \"$DIR\"'"
695 }1390 }
1391 ]
1392 }
1393 ]
696 }1394 }
697}1395}
698```1396```
699 1397
700<Note>1398The hook reads the worktree `name` from the JSON input on stdin, checks out a fresh copy into a new directory, and prints the directory path. The `echo` on the last line is what Claude Code reads as the worktree path. Redirect any other output to stderr so it doesn't interfere with the path.
701 The `decision` and `reason` fields are deprecated for PreToolUse hooks.1399
702 Use `hookSpecificOutput.permissionDecision` and1400#### WorktreeCreate input
703 `hookSpecificOutput.permissionDecisionReason` instead. The deprecated fields1401
704 `"approve"` and `"block"` map to `"allow"` and `"deny"` respectively.1402In addition to the [common input fields](#common-input-fields), WorktreeCreate hooks receive the `name` field. This is a slug identifier for the new worktree, either specified by the user or auto-generated (for example, `bold-oak-a3f2`).
705</Note>1403
1404```json theme={null}
1405{
1406 "session_id": "abc123",
1407 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
1408 "cwd": "/Users/...",
1409 "hook_event_name": "WorktreeCreate",
1410 "name": "feature-auth"
1411}
1412```
1413
1414#### WorktreeCreate output
1415
1416The hook must print the absolute path to the created worktree directory on stdout. If the hook fails or produces no output, worktree creation fails with an error.
1417
1418WorktreeCreate hooks do not use the standard allow/block decision model. Instead, the hook's success or failure determines the outcome. Only `type: "command"` hooks are supported.
706 1419
707#### `PermissionRequest` Decision Control1420### WorktreeRemove
708 1421
709`PermissionRequest` hooks can allow or deny permission requests shown to the user.1422The cleanup counterpart to [WorktreeCreate](#worktreecreate). This hook fires when a worktree is being removed, either when you exit a `--worktree` session and choose to remove it, or when a subagent with `isolation: "worktree"` finishes. For git-based worktrees, Claude handles cleanup automatically with `git worktree remove`. If you configured a WorktreeCreate hook for a non-git version control system, pair it with a WorktreeRemove hook to handle cleanup. Without one, the worktree directory is left on disk.
710 1423
711* For `"behavior": "allow"` you can also optionally pass in an `"updatedInput"` that modifies the tool's input parameters before the tool executes.1424Claude Code passes the path that WorktreeCreate printed on stdout as `worktree_path` in the hook input. This example reads that path and removes the directory:
712* 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.
713 1425
714```json theme={null}1426```json theme={null}
715{1427{
716 "hookSpecificOutput": {1428 "hooks": {
717 "hookEventName": "PermissionRequest",1429 "WorktreeRemove": [
718 "decision": {1430 {
719 "behavior": "allow",1431 "hooks": [
720 "updatedInput": {1432 {
721 "command": "npm run lint"1433 "type": "command",
1434 "command": "bash -c 'jq -r .worktree_path | xargs rm -rf'"
722 }1435 }
1436 ]
723 }1437 }
1438 ]
724 }1439 }
725}1440}
726```1441```
727 1442
728#### `PostToolUse` Decision Control1443#### WorktreeRemove input
729
730`PostToolUse` hooks can provide feedback to Claude after tool execution.
731 1444
732* `"block"` automatically prompts Claude with `reason`.1445In addition to the [common input fields](#common-input-fields), WorktreeRemove hooks receive the `worktree_path` field, which is the absolute path to the worktree being removed.
733* `undefined` does nothing. `reason` is ignored.
734* `"hookSpecificOutput.additionalContext"` adds context for Claude to consider.
735 1446
736```json theme={null}1447```json theme={null}
737{1448{
738 "decision": "block" | undefined,1449 "session_id": "abc123",
739 "reason": "Explanation for decision",1450 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
740 "hookSpecificOutput": {1451 "cwd": "/Users/...",
741 "hookEventName": "PostToolUse",1452 "hook_event_name": "WorktreeRemove",
742 "additionalContext": "Additional information for Claude"1453 "worktree_path": "/Users/.../my-project/.claude/worktrees/feature-auth"
743 }
744}1454}
745```1455```
746 1456
747#### `UserPromptSubmit` Decision Control1457WorktreeRemove hooks have no decision control. They cannot block worktree removal but can perform cleanup tasks like removing version control state or archiving changes. Hook failures are logged in debug mode only. Only `type: "command"` hooks are supported.
748
749`UserPromptSubmit` hooks can control whether a user prompt is processed and add context.
750 1458
751**Adding context (exit code 0):**1459### PreCompact
752There are two ways to add context to the conversation:
753 1460
7541. **Plain text stdout** (simpler): Any non-JSON text written to stdout is added1461Runs before Claude Code is about to run a compact operation.
755 as context. This is the easiest way to inject information.
756 1462
7572. **JSON with `additionalContext`** (structured): Use the JSON format below for1463The matcher value indicates whether compaction was triggered manually or automatically:
758 more control. The `additionalContext` field is added as context.
759 1464
760Both methods work with exit code 0. Plain stdout is shown as hook output in1465| Matcher | When it fires |
761the transcript; `additionalContext` is added more discretely.1466| :------- | :------------------------------------------- |
1467| `manual` | `/compact` |
1468| `auto` | Auto-compact when the context window is full |
762 1469
763**Blocking prompts:**1470#### PreCompact input
764 1471
765* `"decision": "block"` prevents the prompt from being processed. The submitted1472In 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.
766 prompt is erased from context. `"reason"` is shown to the user but not added
767 to context.
768* `"decision": undefined` (or omitted) allows the prompt to proceed normally.
769 1473
770```json theme={null}1474```json theme={null}
771{1475{
772 "decision": "block" | undefined,1476 "session_id": "abc123",
773 "reason": "Explanation for decision",1477 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
774 "hookSpecificOutput": {1478 "cwd": "/Users/...",
775 "hookEventName": "UserPromptSubmit",1479 "permission_mode": "default",
776 "additionalContext": "My additional context here"1480 "hook_event_name": "PreCompact",
777 }1481 "trigger": "manual",
1482 "custom_instructions": ""
778}1483}
779```1484```
780 1485
781<Note>1486### SessionEnd
782 The JSON format is not required for simple use cases. To add context, you can
783 just print plain text to stdout with exit code 0. Use JSON when you need to
784 block prompts or want more structured control.
785</Note>
786 1487
787#### `Stop`/`SubagentStop` Decision Control1488Runs when a Claude Code session ends. Useful for cleanup tasks, logging session
1489statistics, or saving session state. Supports matchers to filter by exit reason.
788 1490
789`Stop` and `SubagentStop` hooks can control whether Claude must continue.1491The `reason` field in the hook input indicates why the session ended:
790 1492
791* `"block"` prevents Claude from stopping. You must populate `reason` for Claude1493| Reason | Description |
792 to know how to proceed.1494| :---------------------------- | :----------------------------------------- |
793* `undefined` allows Claude to stop. `reason` is ignored.1495| `clear` | Session cleared with `/clear` command |
1496| `logout` | User logged out |
1497| `prompt_input_exit` | User exited while prompt input was visible |
1498| `bypass_permissions_disabled` | Bypass permissions mode was disabled |
1499| `other` | Other exit reasons |
1500
1501#### SessionEnd input
1502
1503In 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.
794 1504
795```json theme={null}1505```json theme={null}
796{1506{
797 "decision": "block" | undefined,1507 "session_id": "abc123",
798 "reason": "Must be provided when Claude is blocked from stopping"1508 "transcript_path": "/Users/.../.claude/projects/.../00893aaf-19fa-41d2-8238-13269b9b3ca0.jsonl",
1509 "cwd": "/Users/...",
1510 "permission_mode": "default",
1511 "hook_event_name": "SessionEnd",
1512 "reason": "other"
799}1513}
800```1514```
801 1515
802#### `SessionStart` Decision Control1516SessionEnd hooks have no decision control. They cannot block session termination but can perform cleanup tasks.
1517
1518## Prompt-based hooks
1519
1520In 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, and agent hooks (`type: "agent"`) that spawn an agentic verifier with tool access. Not all events support every hook type.
1521
1522Events that support all three hook types (`command`, `prompt`, and `agent`):
803 1523
804`SessionStart` hooks allow you to load in context at the start of a session.1524* `PermissionRequest`
1525* `PostToolUse`
1526* `PostToolUseFailure`
1527* `PreToolUse`
1528* `Stop`
1529* `SubagentStop`
1530* `TaskCompleted`
1531* `UserPromptSubmit`
1532
1533Events that only support `type: "command"` hooks:
1534
1535* `ConfigChange`
1536* `Notification`
1537* `PreCompact`
1538* `SessionEnd`
1539* `SessionStart`
1540* `SubagentStart`
1541* `TeammateIdle`
1542* `WorktreeCreate`
1543* `WorktreeRemove`
1544
1545### How prompt-based hooks work
1546
1547Instead of executing a Bash command, prompt-based hooks:
1548
15491. Send the hook input and your prompt to a Claude model, Haiku by default
15502. The LLM responds with structured JSON containing a decision
15513. Claude Code processes the decision automatically
805 1552
806* `"hookSpecificOutput.additionalContext"` adds the string to the context.1553### Prompt hook configuration
807* Multiple hooks' `additionalContext` values are concatenated.1554
1555Set `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.
1556
1557This `Stop` hook asks the LLM to evaluate whether all tasks are complete before allowing Claude to finish:
808 1558
809```json theme={null}1559```json theme={null}
810{1560{
811 "hookSpecificOutput": {1561 "hooks": {
812 "hookEventName": "SessionStart",1562 "Stop": [
813 "additionalContext": "My additional context here"1563 {
1564 "hooks": [
1565 {
1566 "type": "prompt",
1567 "prompt": "Evaluate if Claude should stop: $ARGUMENTS. Check if all tasks are complete."
1568 }
1569 ]
1570 }
1571 ]
814 }1572 }
815}1573}
816```1574```
817 1575
818#### `SessionEnd` Decision Control1576| Field | Required | Description |
819 1577| :-------- | :------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
820`SessionEnd` hooks run when a session ends. They cannot block session termination1578| `type` | yes | Must be `"prompt"` |
821but can perform cleanup tasks.1579| `prompt` | yes | The prompt text to send to the LLM. Use `$ARGUMENTS` as a placeholder for the hook input JSON. If `$ARGUMENTS` is not present, input JSON is appended to the prompt |
822 1580| `model` | no | Model to use for evaluation. Defaults to a fast model |
823#### Exit Code Example: Bash Command Validation1581| `timeout` | no | Timeout in seconds. Default: 30 |
824
825```python theme={null}
826#!/usr/bin/env python3
827import json
828import re
829import sys
830
831# Define validation rules as a list of (regex pattern, message) tuples
832VALIDATION_RULES = [
833 (
834 r"\bgrep\b(?!.*\|)",
835 "Use 'rg' (ripgrep) instead of 'grep' for better performance and features",
836 ),
837 (
838 r"\bfind\s+\S+\s+-name\b",
839 "Use 'rg --files | rg pattern' or 'rg --files -g pattern' instead of 'find -name' for better performance",
840 ),
841]
842
843
844def validate_command(command: str) -> list[str]:
845 issues = []
846 for pattern, message in VALIDATION_RULES:
847 if re.search(pattern, command):
848 issues.append(message)
849 return issues
850
851
852try:
853 input_data = json.load(sys.stdin)
854except json.JSONDecodeError as e:
855 print(f"Error: Invalid JSON input: {e}", file=sys.stderr)
856 sys.exit(1)
857
858tool_name = input_data.get("tool_name", "")
859tool_input = input_data.get("tool_input", {})
860command = tool_input.get("command", "")
861
862if tool_name != "Bash" or not command:
863 sys.exit(1)
864
865# Validate the command
866issues = validate_command(command)
867
868if issues:
869 for message in issues:
870 print(f"• {message}", file=sys.stderr)
871 # Exit code 2 blocks tool call and shows stderr to Claude
872 sys.exit(2)
873```
874 1582
875#### JSON Output Example: UserPromptSubmit to Add Context and Validation1583### Response schema
876 1584
877<Note>1585The LLM must respond with JSON containing:
878 For `UserPromptSubmit` hooks, you can inject context using either method:
879 1586
880 * **Plain text stdout** with exit code 0: Simplest approach—just print text1587```json theme={null}
881 * **JSON output** with exit code 0: Use `"decision": "block"` to reject prompts,1588{
882 or `additionalContext` for structured context injection1589 "ok": true | false,
1590 "reason": "Explanation for the decision"
1591}
1592```
883 1593
884 Remember: Exit code 2 only uses `stderr` for the error message. To block using1594| Field | Description |
885 JSON (with a custom reason), use `"decision": "block"` with exit code 0.1595| :------- | :--------------------------------------------------------- |
886</Note>1596| `ok` | `true` allows the action, `false` prevents it |
1597| `reason` | Required when `ok` is `false`. Explanation shown to Claude |
887 1598
888```python theme={null}1599### Example: Multi-criteria Stop hook
889#!/usr/bin/env python31600
890import json1601This `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:
891import sys1602
892import re1603```json theme={null}
893import datetime1604{
894 1605 "hooks": {
895# Load input from stdin1606 "Stop": [
896try:1607 {
897 input_data = json.load(sys.stdin)1608 "hooks": [
898except json.JSONDecodeError as e:1609 {
899 print(f"Error: Invalid JSON input: {e}", file=sys.stderr)1610 "type": "prompt",
900 sys.exit(1)1611 "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.",
901 1612 "timeout": 30
902prompt = input_data.get("prompt", "")
903
904# Check for sensitive patterns
905sensitive_patterns = [
906 (r"(?i)\b(password|secret|key|token)\s*[:=]", "Prompt contains potential secrets"),
907]
908
909for pattern, message in sensitive_patterns:
910 if re.search(pattern, prompt):
911 # Use JSON output to block with a specific reason
912 output = {
913 "decision": "block",
914 "reason": f"Security policy violation: {message}. Please rephrase your request without sensitive information."
915 }1613 }
916 print(json.dumps(output))1614 ]
917 sys.exit(0)1615 }
1616 ]
1617 }
1618}
1619```
918 1620
919# Add current time to context1621## Agent-based hooks
920context = f"Current time: {datetime.datetime.now()}"
921print(context)
922 1622
923"""1623Agent-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.
924The following is also equivalent:
925print(json.dumps({
926 "hookSpecificOutput": {
927 "hookEventName": "UserPromptSubmit",
928 "additionalContext": context,
929 },
930}))
931"""
932 1624
933# Allow the prompt to proceed with the additional context1625### How agent hooks work
934sys.exit(0)
935```
936 1626
937#### JSON Output Example: PreToolUse with Approval1627When an agent hook fires:
938
939```python theme={null}
940#!/usr/bin/env python3
941import json
942import sys
943
944# Load input from stdin
945try:
946 input_data = json.load(sys.stdin)
947except json.JSONDecodeError as e:
948 print(f"Error: Invalid JSON input: {e}", file=sys.stderr)
949 sys.exit(1)
950
951tool_name = input_data.get("tool_name", "")
952tool_input = input_data.get("tool_input", {})
953
954# Example: Auto-approve file reads for documentation files
955if tool_name == "Read":
956 file_path = tool_input.get("file_path", "")
957 if file_path.endswith((".md", ".mdx", ".txt", ".json")):
958 # Use JSON output to auto-approve the tool call
959 output = {
960 "decision": "approve",
961 "reason": "Documentation file auto-approved",
962 "suppressOutput": True # Don't show in verbose mode
963 }
964 print(json.dumps(output))
965 sys.exit(0)
966
967# For other cases, let the normal permission flow proceed
968sys.exit(0)
969```
970 1628
971## Working with MCP Tools16291. Claude Code spawns a subagent with your prompt and the hook's JSON input
16302. The subagent can use tools like Read, Grep, and Glob to investigate
16313. After up to 50 turns, the subagent returns a structured `{ "ok": true/false }` decision
16324. Claude Code processes the decision the same way as a prompt hook
972 1633
973Claude Code hooks work seamlessly with1634Agent hooks are useful when verification requires inspecting actual files or test output, not just evaluating the hook input data alone.
974[Model Context Protocol (MCP) tools](/en/mcp). When MCP servers
975provide tools, they appear with a special naming pattern that you can match in
976your hooks.
977 1635
978### MCP Tool Naming1636### Agent hook configuration
979 1637
980MCP tools follow the pattern `mcp__<server>__<tool>`, for example:1638Set `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:
981 1639
982* `mcp__memory__create_entities` - Memory server's create entities tool1640| Field | Required | Description |
983* `mcp__filesystem__read_file` - Filesystem server's read file tool1641| :-------- | :------- | :------------------------------------------------------------------------------------------ |
984* `mcp__github__search_repositories` - GitHub server's search tool1642| `type` | yes | Must be `"agent"` |
1643| `prompt` | yes | Prompt describing what to verify. Use `$ARGUMENTS` as a placeholder for the hook input JSON |
1644| `model` | no | Model to use. Defaults to a fast model |
1645| `timeout` | no | Timeout in seconds. Default: 60 |
985 1646
986### Configuring Hooks for MCP Tools1647The response schema is the same as prompt hooks: `{ "ok": true }` to allow or `{ "ok": false, "reason": "..." }` to block.
987 1648
988You can target specific MCP tools or entire MCP servers:1649This `Stop` hook verifies that all unit tests pass before allowing Claude to finish:
989 1650
990```json theme={null}1651```json theme={null}
991{1652{
992 "hooks": {1653 "hooks": {
993 "PreToolUse": [1654 "Stop": [
994 {1655 {
995 "matcher": "mcp__memory__.*",
996 "hooks": [1656 "hooks": [
997 {1657 {
998 "type": "command",1658 "type": "agent",
999 "command": "echo 'Memory operation initiated' >> ~/mcp-operations.log"1659 "prompt": "Verify that all unit tests pass. Run the test suite and check the results. $ARGUMENTS",
1660 "timeout": 120
1000 }1661 }
1001 ]1662 ]
1002 },1663 }
1664 ]
1665 }
1666}
1667```
1668
1669## Run hooks in the background
1670
1671By 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.
1672
1673### Configure an async hook
1674
1675Add `"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.
1676
1677This 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:
1678
1679```json theme={null}
1680{
1681 "hooks": {
1682 "PostToolUse": [
1003 {1683 {
1004 "matcher": "mcp__.*__write.*",1684 "matcher": "Write",
1005 "hooks": [1685 "hooks": [
1006 {1686 {
1007 "type": "command",1687 "type": "command",
1008 "command": "/home/user/scripts/validate-mcp-write.py"1688 "command": "/path/to/run-tests.sh",
1689 "async": true,
1690 "timeout": 120
1009 }1691 }
1010 ]1692 ]
1011 }1693 }
1014}1696}
1015```1697```
1016 1698
1017## Examples1699The `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.
1018 1700
1019<Tip>1701### How async hooks execute
1020 For practical examples including code formatting, notifications, and file protection, see [More Examples](/en/hooks-guide#more-examples) in the get started guide.
1021</Tip>
1022
1023## Security Considerations
1024
1025### Disclaimer
1026 1702
1027**USE AT YOUR OWN RISK**: Claude Code hooks execute arbitrary shell commands on1703When 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.
1028your system automatically. By using hooks, you acknowledge that:
1029 1704
1030* You are solely responsible for the commands you configure1705After 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.
1031* Hooks can modify, delete, or access any files your user account can access
1032* Malicious or poorly written hooks can cause data loss or system damage
1033* Anthropic provides no warranty and assumes no liability for any damages
1034 resulting from hook usage
1035* You should thoroughly test hooks in a safe environment before production use
1036 1706
1037Always review and understand any hook commands before adding them to your1707### Example: run tests after file changes
1038configuration.
1039 1708
1040### Security Best Practices1709This 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`:
1041 1710
1042Here are some key practices for writing more secure hooks:1711```bash theme={null}
1712#!/bin/bash
1713# run-tests-async.sh
1043 1714
10441. **Validate and sanitize inputs** - Never trust input data blindly1715# Read hook input from stdin
10452. **Always quote shell variables** - Use `"$VAR"` not `$VAR`1716INPUT=$(cat)
10463. **Block path traversal** - Check for `..` in file paths1717FILE_PATH=$(echo "$INPUT" | jq -r '.tool_input.file_path // empty')
10474. **Use absolute paths** - Specify full paths for scripts (use
1048 "\$CLAUDE\_PROJECT\_DIR" for the project path)
10495. **Skip sensitive files** - Avoid `.env`, `.git/`, keys, etc.
1050 1718
1051### Configuration Safety1719# Only run tests for source files
1720if [[ "$FILE_PATH" != *.ts && "$FILE_PATH" != *.js ]]; then
1721 exit 0
1722fi
1052 1723
1053Direct edits to hooks in settings files don't take effect immediately. Claude1724# Run tests and report results via systemMessage
1054Code:1725RESULT=$(npm test 2>&1)
1726EXIT_CODE=$?
1055 1727
10561. Captures a snapshot of hooks at startup1728if [ $EXIT_CODE -eq 0 ]; then
10572. Uses this snapshot throughout the session1729 echo "{\"systemMessage\": \"Tests passed after editing $FILE_PATH\"}"
10583. Warns if hooks are modified externally1730else
10594. Requires review in `/hooks` menu for changes to apply1731 echo "{\"systemMessage\": \"Tests failed after editing $FILE_PATH: $RESULT\"}"
1732fi
1733```
1060 1734
1061This prevents malicious hook modifications from affecting your current session.1735Then add this configuration to `.claude/settings.json` in your project root. The `async: true` flag lets Claude keep working while tests run:
1062 1736
1063## Hook Execution Details1737```json theme={null}
1738{
1739 "hooks": {
1740 "PostToolUse": [
1741 {
1742 "matcher": "Write|Edit",
1743 "hooks": [
1744 {
1745 "type": "command",
1746 "command": "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/run-tests-async.sh",
1747 "async": true,
1748 "timeout": 300
1749 }
1750 ]
1751 }
1752 ]
1753 }
1754}
1755```
1064 1756
1065* **Timeout**: 60-second execution limit by default, configurable per command.1757### Limitations
1066 * A timeout for an individual command does not affect the other commands.
1067* **Parallelization**: All matching hooks run in parallel
1068* **Deduplication**: Multiple identical hook commands are deduplicated automatically
1069* **Environment**: Runs in current directory with Claude Code's environment
1070 * The `CLAUDE_PROJECT_DIR` environment variable is available and contains the
1071 absolute path to the project root directory (where Claude Code was started)
1072 * 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.
1073* **Input**: JSON via stdin
1074* **Output**:
1075 * PreToolUse/PermissionRequest/PostToolUse/Stop/SubagentStop: Progress shown in verbose mode (ctrl+o)
1076 * Notification/SessionEnd: Logged to debug only (`--debug`)
1077 * UserPromptSubmit/SessionStart: stdout added as context for Claude
1078 1758
1079## Debugging1759Async hooks have several constraints compared to synchronous hooks:
1080 1760
1081### Basic Troubleshooting1761* Only `type: "command"` hooks support `async`. Prompt-based hooks cannot run asynchronously.
1762* Async hooks cannot block tool calls or return decisions. By the time the hook completes, the triggering action has already proceeded.
1763* Hook output is delivered on the next conversation turn. If the session is idle, the response waits until the next user interaction.
1764* Each execution creates a separate background process. There is no deduplication across multiple firings of the same async hook.
1082 1765
1083If your hooks aren't working:1766## Security considerations
1084 1767
10851. **Check configuration** - Run `/hooks` to see if your hook is registered1768### Disclaimer
10862. **Verify syntax** - Ensure your JSON settings are valid
10873. **Test commands** - Run hook commands manually first
10884. **Check permissions** - Make sure scripts are executable
10895. **Review logs** - Use `claude --debug` to see hook execution details
1090 1769
1091Common issues:1770Command hooks run with your system user's full permissions.
1092 1771
1093* **Quotes not escaped** - Use `\"` inside JSON strings1772<Warning>
1094* **Wrong matcher** - Check tool names match exactly (case-sensitive)1773 Command 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.
1095* **Command not found** - Use full paths for scripts1774</Warning>
1096 1775
1097### Advanced Debugging1776### Security best practices
1098 1777
1099For complex hook issues:1778Keep these practices in mind when writing hooks:
1100 1779
11011. **Inspect hook execution** - Use `claude --debug` to see detailed hook1780* **Validate and sanitize inputs**: never trust input data blindly
1102 execution1781* **Always quote shell variables**: use `"$VAR"` not `$VAR`
11032. **Validate JSON schemas** - Test hook input/output with external tools1782* **Block path traversal**: check for `..` in file paths
11043. **Check environment variables** - Verify Claude Code's environment is correct1783* **Use absolute paths**: specify full paths for scripts, using `"$CLAUDE_PROJECT_DIR"` for the project root
11054. **Test edge cases** - Try hooks with unusual file paths or inputs1784* **Skip sensitive files**: avoid `.env`, `.git/`, keys, etc.
11065. **Monitor system resources** - Check for resource exhaustion during hook
1107 execution
11086. **Use structured logging** - Implement logging in your hook scripts
1109 1785
1110### Debug Output Example1786## Debug hooks
1111 1787
1112Use `claude --debug` to see hook execution details:1788Run `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.
1113 1789
1114```1790```text theme={null}
1115[DEBUG] Executing hooks for PostToolUse:Write1791[DEBUG] Executing hooks for PostToolUse:Write
1116[DEBUG] Getting matching hook commands for PostToolUse with query: Write1792[DEBUG] Getting matching hook commands for PostToolUse with query: Write
1117[DEBUG] Found 1 hook matchers in settings1793[DEBUG] Found 1 hook matchers in settings
1118[DEBUG] Matched 1 hooks for query "Write"1794[DEBUG] Matched 1 hooks for query "Write"
1119[DEBUG] Found 1 hook commands to execute1795[DEBUG] Found 1 hook commands to execute
1120[DEBUG] Executing hook command: <Your command> with timeout 60000ms1796[DEBUG] Executing hook command: <Your command> with timeout 600000ms
1121[DEBUG] Hook command completed with status 0: <Your stdout>1797[DEBUG] Hook command completed with status 0: <Your stdout>
1122```1798```
1123 1799
1124Progress messages appear in verbose mode (ctrl+o) showing:1800For 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.
1125
1126* Which hook is running
1127* Command being executed
1128* Success/failure status
1129* Output or error messages
1130
1131
1132
1133> To find navigation and other pages in this documentation, fetch the llms.txt file at: https://code.claude.com/docs/llms.txt
