OpenAI - Codex Docs Changes 2026-03-05 06:22 UTC → 2026-04-23 12:28 UTC

1# Agent approvals & security

2 

3Codex helps protect your code and data and reduces the risk of misuse.

4 

5This page covers how to operate Codex safely, including sandboxing, approvals,

6 and network access. If you are looking for Codex Security, the product for

7 scanning connected GitHub repositories, see [Codex Security](https://developers.openai.com/codex/security).

8 

9By default, the agent runs with network access turned off. Locally, Codex uses an OS-enforced sandbox that limits what it can touch (typically to the current workspace), plus an approval policy that controls when it must stop and ask you before acting.

10 

11For a high-level explanation of how sandboxing works across the Codex app, IDE

12extension, and CLI, see [sandboxing](https://developers.openai.com/codex/concepts/sandboxing).

13For a broader enterprise security overview, see the [Codex security white paper](https://trust.openai.com/?itemUid=382f924d-54f3-43a8-a9df-c39e6c959958&source=click).

14 

15## Sandbox and approvals

16 

17Codex security controls come from two layers that work together:

18 

19- **Sandbox mode**: What Codex can do technically (for example, where it can write and whether it can reach the network) when it executes model-generated commands.

20- **Approval policy**: When Codex must ask you before it executes an action (for example, leaving the sandbox, using the network, or running commands outside a trusted set).

21 

22Codex uses different sandbox modes depending on where you run it:

23 

24- **Codex cloud**: Runs in isolated OpenAI-managed containers, preventing access to your host system or unrelated data. Uses a two-phase runtime model: setup runs before the agent phase and can access the network to install specified dependencies, then the agent phase runs offline by default unless you enable internet access for that environment. Secrets configured for cloud environments are available only during setup and are removed before the agent phase starts.

25- **Codex CLI / IDE extension**: OS-level mechanisms enforce sandbox policies. Defaults include no network access and write permissions limited to the active workspace. You can configure the sandbox, approval policy, and network settings based on your risk tolerance.

26 

27In the `Auto` preset (for example, `--full-auto`), Codex can read files, make edits, and run commands in the working directory automatically.

28 

29Codex asks for approval to edit files outside the workspace or to run commands that require network access. If you want to chat or plan without making changes, switch to `read-only` mode with the `/permissions` command.

30 

31Codex can also elicit approval for app (connector) tool calls that advertise side effects, even when the action isn't a shell command or file change. Destructive app/MCP tool calls always require approval when the tool advertises a destructive annotation, even if it also advertises other hints (for example, read-only hints).

32 

33## Network access [Elevated Risk](https://help.openai.com/articles/20001061)

34 

35For Codex cloud, see [agent internet access](https://developers.openai.com/codex/cloud/internet-access) to enable full internet access or a domain allow list.

36 

37For the Codex app, CLI, or IDE Extension, the default `workspace-write` sandbox mode keeps network access turned off unless you enable it in your configuration:

38 

39```toml

40[sandbox_workspace_write]

41network_access = true

42```

43 

44You can also control the [web search tool](https://platform.openai.com/docs/guides/tools-web-search) without granting full network access to spawned commands. Codex defaults to using a web search cache to access results. The cache is an OpenAI-maintained index of web results, so cached mode returns pre-indexed results instead of fetching live pages. This reduces exposure to prompt injection from arbitrary live content, but you should still treat web results as untrusted. If you are using `--yolo` or another [full access sandbox setting](#common-sandbox-and-approval-combinations), web search defaults to live results. Use `--search` or set `web_search = "live"` to allow live browsing, or set it to `"disabled"` to turn the tool off:

45 

46```toml

47web_search = "cached" # default

48# web_search = "disabled"

49# web_search = "live" # same as --search

50```

51 

52Use caution when enabling network access or web search in Codex. Prompt injection can cause the agent to fetch and follow untrusted instructions.

53 

54## Defaults and recommendations

55 

56- On launch, Codex detects whether the folder is version-controlled and recommends:

57 - Version-controlled folders: `Auto` (workspace write + on-request approvals)

58 - Non-version-controlled folders: `read-only`

59- Depending on your setup, Codex may also start in `read-only` until you explicitly trust the working directory (for example, via an onboarding prompt or `/permissions`).

60- The workspace includes the current directory and temporary directories like `/tmp`. Use the `/status` command to see which directories are in the workspace.

61- To accept the defaults, run `codex`.

62- You can set these explicitly:

63 - `codex --sandbox workspace-write --ask-for-approval on-request`

64 - `codex --sandbox read-only --ask-for-approval on-request`

65 

66### Protected paths in writable roots

67 

68In the default `workspace-write` sandbox policy, writable roots still include protected paths:

69 

70- `<writable_root>/.git` is protected as read-only whether it appears as a directory or file.

71- If `<writable_root>/.git` is a pointer file (`gitdir: ...`), the resolved Git directory path is also protected as read-only.

72- `<writable_root>/.agents` is protected as read-only when it exists as a directory.

73- `<writable_root>/.codex` is protected as read-only when it exists as a directory.

74- Protection is recursive, so everything under those paths is read-only.

75 

76### Deny reads with filesystem profiles

77 

78Named permission profiles can also deny reads for exact paths or glob patterns.

79This is useful when a workspace should stay writable but specific sensitive

80files, such as local environment files, must stay unreadable:

81 

82```toml

83default_permissions = "workspace"

84 

85[permissions.workspace.filesystem]

86":project_roots" = { "." = "write", "**/*.env" = "none" }

87glob_scan_max_depth = 3

88```

89 

90Use `"none"` for paths or globs that Codex shouldn't read. The sandbox policy

91evaluates globs for local macOS and Linux command execution. On platforms that

92pre-expand glob matches before the sandbox starts, set `glob_scan_max_depth` for

93unbounded `**` patterns, or list explicit depths such as `*.env`, `*/*.env`, and

94`*/*/*.env`.

95 

96### Run without approval prompts

97 

98You can disable approval prompts with `--ask-for-approval never` or `-a never` (shorthand).

99 

100This option works with all `--sandbox` modes, so you still control Codex's level of autonomy. Codex makes a best effort within the constraints you set.

101 

102If you need Codex to read files, make edits, and run commands with network access without approval prompts, use `--sandbox danger-full-access` (or the `--dangerously-bypass-approvals-and-sandbox` flag). Use caution before doing so.

103 

104For a middle ground, `approval_policy = { granular = { ... } }` lets you keep specific approval prompt categories interactive while automatically rejecting others. The granular policy covers sandbox approvals, execpolicy-rule prompts, MCP prompts, `request_permissions` prompts, and skill-script approvals.

105 

106Set `approvals_reviewer = "guardian_subagent"` to route eligible approval reviews through the Guardian reviewer subagent instead of prompting the user directly. Admin requirements can constrain this with `allowed_approvals_reviewers`.

107 

108### Common sandbox and approval combinations

109 

110| Intent | Flags | Effect |

111| ----------------------------------------------------------------- | ------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------ |

112| Auto (preset) | *no flags needed* or `--full-auto` | Codex can read files, make edits, and run commands in the workspace. Codex requires approval to edit outside the workspace or to access network. |

113| Safe read-only browsing | `--sandbox read-only --ask-for-approval on-request` | Codex can read files and answer questions. Codex requires approval to make edits, run commands, or access network. |

114| Read-only non-interactive (CI) | `--sandbox read-only --ask-for-approval never` | Codex can only read files; never asks for approval. |

115| Automatically edit but ask for approval to run untrusted commands | `--sandbox workspace-write --ask-for-approval untrusted` | Codex can read and edit files but asks for approval before running untrusted commands. |

116| Dangerous full access | `--dangerously-bypass-approvals-and-sandbox` (alias: `--yolo`) | [Elevated Risk](https://help.openai.com/articles/20001061) No sandbox; no approvals *(not recommended)* |

117 

118`--full-auto` is a convenience alias for `--sandbox workspace-write --ask-for-approval on-request`.

119 

120With `--ask-for-approval untrusted`, Codex runs only known-safe read operations automatically. Commands that can mutate state or trigger external execution paths (for example, destructive Git operations or Git output/config-override flags) require approval.

121 

122#### Configuration in `config.toml`

123 

124For the broader configuration workflow, see [Config basics](https://developers.openai.com/codex/config-basic), [Advanced Config](https://developers.openai.com/codex/config-advanced#approval-policies-and-sandbox-modes), and the [Configuration Reference](https://developers.openai.com/codex/config-reference).

125 

126```toml

127# Always ask for approval mode

128approval_policy = "untrusted"

129sandbox_mode = "read-only"

130allow_login_shell = false # optional hardening: disallow login shells for shell-based tools

131 

132# Optional: Allow network in workspace-write mode

133[sandbox_workspace_write]

134network_access = true

135 

136# Optional: granular approval policy

137# approval_policy = { granular = {

138# sandbox_approval = true,

139# rules = true,

140# mcp_elicitations = true,

141# request_permissions = false,

142# skill_approval = false

143# } }

144```

145 

146You can also save presets as profiles, then select them with `codex --profile <name>`:

147 

148```toml

149[profiles.full_auto]

150approval_policy = "on-request"

151sandbox_mode = "workspace-write"

152 

153[profiles.readonly_quiet]

154approval_policy = "never"

155sandbox_mode = "read-only"

156```

157 

158### Test the sandbox locally

159 

160To see what happens when a command runs under the Codex sandbox, use these Codex CLI commands:

161 

162```bash

163# macOS

164codex sandbox macos [--full-auto] [--log-denials] [COMMAND]...

165# Linux

166codex sandbox linux [--full-auto] [COMMAND]...

167```

168 

169The `sandbox` command is also available as `codex debug`, and the platform helpers have aliases (for example `codex sandbox seatbelt` and `codex sandbox landlock`).

170 

171## OS-level sandbox

172 

173Codex enforces the sandbox differently depending on your OS:

174 

175- **macOS** uses Seatbelt policies and runs commands using `sandbox-exec` with a profile (`-p`) that corresponds to the `--sandbox` mode you selected. When restricted read access enables platform defaults, Codex appends a curated macOS platform policy (instead of broadly allowing `/System`) to preserve common tool compatibility.

176- **Linux** uses `bwrap` plus `seccomp` by default.

177- **Windows** uses the Linux sandbox implementation when running in [Windows Subsystem for Linux 2 (WSL2)](https://developers.openai.com/codex/windows#windows-subsystem-for-linux). WSL1 was supported through Codex `0.114`; starting in `0.115`, the Linux sandbox moved to `bwrap`, so WSL1 is no longer supported. When running natively on Windows, Codex uses a [Windows sandbox](https://developers.openai.com/codex/windows#windows-sandbox) implementation.

178 

179If you use the Codex IDE extension on Windows, it supports WSL2 directly. Set the following in your VS Code settings to keep the agent inside WSL2 whenever it's available:

180 

181```json

182{

183 "chatgpt.runCodexInWindowsSubsystemForLinux": true

184}

185```

186 

187This ensures the IDE extension inherits Linux sandbox semantics for commands, approvals, and filesystem access even when the host OS is Windows. Learn more in the [Windows setup guide](https://developers.openai.com/codex/windows).

188 

189When running natively on Windows, configure the native sandbox mode in `config.toml`:

190 

191```toml

192[windows]

193sandbox = "unelevated" # or "elevated"

194# sandbox_private_desktop = true # default; set false only for compatibility

195```

196 

197See the [Windows setup guide](https://developers.openai.com/codex/windows#windows-sandbox) for details.

198 

199When you run Linux in a containerized environment such as Docker, the sandbox may not work if the host or container configuration blocks the namespace, setuid `bwrap`, or `seccomp` operations that Codex needs.

200 

201In that case, configure your Docker container to provide the isolation you need, then run `codex` with `--sandbox danger-full-access` (or the `--dangerously-bypass-approvals-and-sandbox` flag) inside the container.

202 

203### Run Codex in Dev Containers

204 

205If your host cannot run the Linux sandbox directly, or if your organization already standardizes on containerized development, run Codex with Dev Containers and let Docker provide the outer isolation boundary. This works with Visual Studio Code Dev Containers and compatible tools.

206 

207Use the [Codex secure devcontainer example](https://github.com/openai/codex/tree/main/.devcontainer) as a reference implementation. The example installs Codex, common development tools, `bubblewrap`, and firewall-based outbound controls.

208 

209Devcontainers provide substantial protection, but they do not prevent every

210 attack. If you run Codex with `--sandbox danger-full-access` or

211 `--dangerously-bypass-approvals-and-sandbox` inside the container, a malicious

212 project can exfiltrate anything available inside the devcontainer, including

213 Codex credentials. Use this pattern only with trusted repositories, and

214 monitor Codex activity as you would in any other elevated environment.

215 

216The reference implementation includes:

217 

218- an Ubuntu 24.04 base image with Codex and common development tools installed;

219- an allowlist-driven firewall profile for outbound access;

220- VS Code settings and extension recommendations for reopening the workspace in a container;

221- persistent mounts for command history and Codex configuration;

222- `bubblewrap`, so Codex can still use its Linux sandbox when the container grants the needed capabilities.

223 

224To try it:

225 

2261. Install Visual Studio Code and the [Dev Containers extension](https://marketplace.visualstudio.com/items?itemName=ms-vscode-remote.remote-containers).

2272. Copy the Codex example `.devcontainer` setup into your repository, or start from the Codex repository directly.

2283. In VS Code, run **Dev Containers: Open Folder in Container…** and select `.devcontainer/devcontainer.secure.json`.

2294. After the container starts, open a terminal and run `codex`.

230 

231You can also start the container from the CLI:

232 

233```bash

234devcontainer up --workspace-folder . --config .devcontainer/devcontainer.secure.json

235```

236 

237The example has three main pieces:

238 

239- `.devcontainer/devcontainer.secure.json` controls container settings, capabilities, mounts, environment variables, and VS Code extensions.

240- `.devcontainer/Dockerfile.secure` defines the Ubuntu-based image and installed tools.

241- `.devcontainer/init-firewall.sh` applies the outbound network policy.

242 

243The reference firewall is intentionally a starting point. If you depend on domain allowlisting for isolation, implement DNS rebinding and DNS refresh protections that fit your environment, such as TTL-aware refreshes or a DNS-aware firewall.

244 

245Inside the container, choose one of these modes:

246 

247- Keep Codex's Linux sandbox enabled if the Dev Container profile grants the capabilities needed for `bwrap` to create the inner sandbox.

248- If the container is your intended security boundary, run Codex with `--sandbox danger-full-access` inside the container so Codex does not try to create a second sandbox layer.

249 

250## Version control

251 

252Codex works best with a version control workflow:

253 

254- Work on a feature branch and keep `git status` clean before delegating. This keeps Codex patches easier to isolate and revert.

255- Prefer patch-based workflows (for example, `git diff`/`git apply`) over editing tracked files directly. Commit frequently so you can roll back in small increments.

256- Treat Codex suggestions like any other PR: run targeted verification, review diffs, and document decisions in commit messages for auditing.

257 

258## Monitoring and telemetry

259 

260Codex supports opt-in monitoring via OpenTelemetry (OTel) to help teams audit usage, investigate issues, and meet compliance requirements without weakening local security defaults. Telemetry is off by default; enable it explicitly in your configuration.

261 

262### Overview

263 

264- Codex turns off OTel export by default to keep local runs self-contained.

265- When enabled, Codex emits structured log events covering conversations, API requests, SSE/WebSocket stream activity, user prompts (redacted by default), tool approval decisions, and tool results.

266- Codex tags exported events with `service.name` (originator), CLI version, and an environment label to separate dev/staging/prod traffic.

267 

268### Enable OTel (opt-in)

269 

270Add an `[otel]` block to your Codex configuration (typically `~/.codex/config.toml`), choosing an exporter and whether to log prompt text.

271 

272```toml

273[otel]

274environment = "staging" # dev | staging | prod

275exporter = "none" # none | otlp-http | otlp-grpc

276log_user_prompt = false # redact prompt text unless policy allows

277```

278 

279- `exporter = "none"` leaves instrumentation active but doesn't send data anywhere.

280- To send events to your own collector, pick one of:

281 

282```toml

283[otel]

284exporter = { otlp-http = {

285 endpoint = "https://otel.example.com/v1/logs",

286 protocol = "binary",

287 headers = { "x-otlp-api-key" = "${OTLP_TOKEN}" }

288}}

289```

290 

291```toml

292[otel]

293exporter = { otlp-grpc = {

294 endpoint = "https://otel.example.com:4317",

295 headers = { "x-otlp-meta" = "abc123" }

296}}

297```

298 

299Codex batches events and flushes them on shutdown. Codex exports only telemetry produced by its OTel module.

300 

301### Event categories

302 

303Representative event types include:

304 

305- `codex.conversation_starts` (model, reasoning settings, sandbox/approval policy)

306- `codex.api_request` (attempt, status/success, duration, and error details)

307- `codex.sse_event` (stream event kind, success/failure, duration, plus token counts on `response.completed`)

308- `codex.websocket_request` and `codex.websocket_event` (request duration plus per-message kind/success/error)

309- `codex.user_prompt` (length; content redacted unless explicitly enabled)

310- `codex.tool_decision` (approved/denied, source: configuration vs. user)

311- `codex.tool_result` (duration, success, output snippet)

312 

313Associated OTel metrics (counter plus duration histogram pairs) include `codex.api_request`, `codex.sse_event`, `codex.websocket.request`, `codex.websocket.event`, and `codex.tool.call` (with corresponding `.duration_ms` instruments).

314 

315For the full event catalog and configuration reference, see the [Codex configuration documentation on GitHub](https://github.com/openai/codex/blob/main/docs/config.md#otel).

316 

317### Security and privacy guidance

318 

319- Keep `log_user_prompt = false` unless policy explicitly permits storing prompt contents. Prompts can include source code and sensitive data.

320- Route telemetry only to collectors you control; apply retention limits and access controls aligned with your compliance requirements.

321- Treat tool arguments and outputs as sensitive. Favor redaction at the collector or SIEM when possible.

322- Review local data retention settings (for example, `history.persistence` / `history.max_bytes`) if you don't want Codex to save session transcripts under `CODEX_HOME`. See [Advanced Config](https://developers.openai.com/codex/config-advanced#history-persistence) and [Configuration Reference](https://developers.openai.com/codex/config-reference).

323- If you run the CLI with network access turned off, OTel export can't reach your collector. To export, allow network access in `workspace-write` mode for the OTel endpoint, or export from Codex cloud with the collector domain on your approved list.

324- Review events periodically for approval/sandbox changes and unexpected tool executions.

325 

326OTel is optional and designed to complement, not replace, the sandbox and approval protections described above.

327 

328## Managed configuration

329 

330Enterprise admins can configure Codex security settings for their workspace in [Managed configuration](https://developers.openai.com/codex/enterprise/managed-configuration). See that page for setup and policy details.

10 10 

11## Getting started11## Getting started

12 12 

13The Codex app is available on macOS (Apple Silicon).13The Codex app is available on macOS and Windows.

14 

15Most Codex app features are available on both platforms. Platform-specific

16exceptions are noted in the relevant docs.

14 17 

151. Download and install the Codex app181. Download and install the Codex app

16 19 

17 Download the Codex app for Windows or macOS.20 Download the Codex app for macOS or Windows. Choose the Intel build if you're using an Intel-based Mac.

21 

22 [Download for macOS (Apple Silicon)](https://persistent.oaistatic.com/codex-app-prod/Codex.dmg)[Download for macOS (Intel)](https://persistent.oaistatic.com/codex-app-prod/Codex-latest-x64.dmg)

23 

24 Need a different operating system?

18 25 

19 [Download for macOS](https://persistent.oaistatic.com/codex-app-prod/Codex.dmg)26 [Download for Windows](https://get.microsoft.com/installer/download/9PLM9XGG6VKS?cid=website_cta_psi)

20 27 

21 [Get notified for Linux](https://openai.com/form/codex-app/)28 [Get notified for Linux](https://openai.com/form/codex-app/)

222. Open Codex and sign in292. Open Codex and sign in

40- Build a classic Snake game in this repo.47- Build a classic Snake game in this repo.

41- Find and fix bugs in my codebase with minimal, high-confidence changes.48- Find and fix bugs in my codebase with minimal, high-confidence changes.

42 49 

43 If you need more inspiration, check out the [explore section](https://developers.openai.com/codex/explore).50 If you need more inspiration, explore [Codex use cases](https://developers.openai.com/codex/use-cases).

51 If you're new to Codex, read the [best practices guide](https://developers.openai.com/codex/learn/best-practices).

44 52 

45---53---

46 54 

48 56 

49[### Multitask across projects57[### Multitask across projects

50 58 

51Run multiple tasks in parallel and switch quickly between them.](https://developers.openai.com/codex/app/features#multitask-across-projects)[### Built-in Git tools59Run project threads side by side and switch between them quickly.](https://developers.openai.com/codex/app/features#multitask-across-projects)[### Worktrees

60 

61Keep parallel code changes isolated with built-in Git worktree support.](https://developers.openai.com/codex/app/worktrees)[### Computer use

62 

63Let Codex use macOS apps for GUI tasks, browser flows, and native app testing.](https://developers.openai.com/codex/app/computer-use)[### Review and ship changes

64 

65Inspect diffs, address PR feedback, stage files, commit, and push.](https://developers.openai.com/codex/app/review)[### Terminal and actions

52 66 

53Review diffs, comment inline, stage or revert chunks, and commit without leaving the app.](https://developers.openai.com/codex/app/features#built-in-git-tools)[### Worktrees for parallel tasks67Run commands in each thread and launch repeatable project actions.](https://developers.openai.com/codex/app/features#integrated-terminal)[### In-app browser

54 68 

55Isolate changes of multiple Codex threads using built-in Git worktree support.](https://developers.openai.com/codex/app/worktrees)[### Skills support69Open unauthenticated local or public pages and comment on rendered output.](https://developers.openai.com/codex/app/browser)[### Image generation

56 70 

57Give your Codex agent additional capabilities and reuse skills across App, CLI, and IDE Extension.](https://developers.openai.com/codex/app/features#skills-support)[### Automations71Generate or edit images in a thread while you work on the surrounding code and assets.](https://developers.openai.com/codex/app/features#image-generation)[### Automations

58 72 

59Pair skills with automations to automate recurring tasks in the background. Codex adds findings to the inbox, or automatically archives runs if there’s nothing to report.](https://developers.openai.com/codex/app/automations)[### Built-in terminal73Schedule recurring tasks, or wake up the same thread for ongoing checks.](https://developers.openai.com/codex/app/automations)[### Skills

60 74 

61Open a terminal per thread to test your changes, run dev servers, scripts, and custom commands.](https://developers.openai.com/codex/app/features#integrated-terminal)[### Local environments75Reuse instructions and workflows across the app, CLI, and IDE Extension.](https://developers.openai.com/codex/app/features#skills-support)[### Sidebar and artifacts

62 76 

63Define worktree setup scripts and common project actions for easy access.](https://developers.openai.com/codex/app/local-environments)[### Sync with the IDE extension77Follow plans, sources, task summaries, and generated file previews.](https://developers.openai.com/codex/app/features#richer-outputs-and-artifacts)[### Plugins

64 78 

65Share Auto Context and active threads across app and IDE sessions.](https://developers.openai.com/codex/app/features#sync-with-the-ide-extension)[### MCP support79Connect apps, skills, and MCP servers to extend what Codex can do.](https://developers.openai.com/codex/plugins)[### IDE Extension sync

66 80 

67Connect your Codex agent to additional services using MCP.](https://developers.openai.com/codex/app/features#mcp-support)81Share Auto Context and active threads across app and IDE sessions.](https://developers.openai.com/codex/app/features#sync-with-the-ide-extension)

68 82 

69---83---

70 84 

12Supported transports:12Supported transports:

13 13 

14- `stdio` (`--listen stdio://`, default): newline-delimited JSON (JSONL).14- `stdio` (`--listen stdio://`, default): newline-delimited JSON (JSONL).

15- `websocket` (`--listen ws://IP:PORT`, experimental): one JSON-RPC message per WebSocket text frame.15- `websocket` (`--listen ws://IP:PORT`, experimental and unsupported): one JSON-RPC message per WebSocket text frame.

16- `off` (`--listen off`): don't expose a local transport.

17 

18When you run with `--listen ws://IP:PORT`, the same listener also serves basic HTTP health probes:

19 

20- `GET /readyz` returns `200 OK` once the listener accepts new connections.

21- `GET /healthz` returns `200 OK` when the request doesn't include an `Origin` header.

22- Requests with an `Origin` header are rejected with `403 Forbidden`.

23 

24WebSocket transport is experimental and unsupported. Loopback listeners such as `ws://127.0.0.1:PORT` are appropriate for localhost and SSH port-forwarding workflows. Non-loopback WebSocket listeners currently allow unauthenticated connections by default during rollout, so configure WebSocket auth before exposing one remotely.

25 

26Supported WebSocket auth flags:

27 

28- `--ws-auth capability-token --ws-token-file /absolute/path`

29- `--ws-auth capability-token --ws-token-sha256 HEX`

30- `--ws-auth signed-bearer-token --ws-shared-secret-file /absolute/path`

31 

32For signed bearer tokens, you can also set `--ws-issuer`, `--ws-audience`, and `--ws-max-clock-skew-seconds`. Clients present the credential as `Authorization: Bearer <token>` during the WebSocket handshake, and app-server enforces auth before JSON-RPC `initialize`.

33 

34Prefer `--ws-token-file` over passing raw bearer tokens on the command line. Use `--ws-token-sha256` only when the client keeps the raw high-entropy token in a separate local secret store; the hash is only a verifier, and clients still need the original token.

16 35 

17In WebSocket mode, app-server uses bounded queues. When request ingress is full, the server rejects new requests with JSON-RPC error code `-32001` and message `"Server overloaded; retry later."` Clients should retry with an exponentially increasing delay and jitter.36In WebSocket mode, app-server uses bounded queues. When request ingress is full, the server rejects new requests with JSON-RPC error code `-32001` and message `"Server overloaded; retry later."` Clients should retry with an exponentially increasing delay and jitter.

18 37 

21Requests include `method`, `params`, and `id`:40Requests include `method`, `params`, and `id`:

22 41 

23```json42```json

24{ "method": "thread/start", "id": 10, "params": { "model": "gpt-5.1-codex" } }43{ "method": "thread/start", "id": 10, "params": { "model": "gpt-5.4" } }

25```44```

26 45 

27Responses echo the `id` with either `result` or `error`:46Responses echo the `id` with either `result` or `error`:

99 },118 },

100});119});

101send({ method: "initialized", params: {} });120send({ method: "initialized", params: {} });

102send({ method: "thread/start", id: 1, params: { model: "gpt-5.1-codex" } });121send({ method: "thread/start", id: 1, params: { model: "gpt-5.4" } });

103```122```

104 123 

105## Core primitives124## Core primitives

123 142 

124Clients must send a single `initialize` request per transport connection before invoking any other method on that connection, then acknowledge with an `initialized` notification. Requests sent before initialization receive a `Not initialized` error, and repeated `initialize` calls on the same connection return `Already initialized`.143Clients must send a single `initialize` request per transport connection before invoking any other method on that connection, then acknowledge with an `initialized` notification. Requests sent before initialization receive a `Not initialized` error, and repeated `initialize` calls on the same connection return `Already initialized`.

125 144 

126The server returns the user agent string it will present to upstream services. Set `clientInfo` to identify your integration.145The server returns the user agent string it will present to upstream services plus `platformFamily` and `platformOs` values that describe the runtime target. Set `clientInfo` to identify your integration.

127 146 

128`initialize.params.capabilities` also supports per-connection notification opt-out via `optOutNotificationMethods`, which is a list of exact method names to suppress for that connection. Matching is exact (no wildcards/prefixes). Unknown method names are accepted and ignored.147`initialize.params.capabilities` also supports per-connection notification opt-out via `optOutNotificationMethods`, which is a list of exact method names to suppress for that connection. Matching is exact (no wildcards/prefixes). Unknown method names are accepted and ignored.

129 148 

159 },178 },

160 "capabilities": {179 "capabilities": {

161 "experimentalApi": true,180 "experimentalApi": true,

162 "optOutNotificationMethods": [181 "optOutNotificationMethods": ["thread/started", "item/agentMessage/delta"]

163 "codex/event/session_configured",

164 "item/agentMessage/delta"

165 ]

166 }182 }

167 }183 }

168}184}

202- `thread/resume` - reopen an existing thread by id so later `turn/start` calls append to it.218- `thread/resume` - reopen an existing thread by id so later `turn/start` calls append to it.

203- `thread/fork` - fork a thread into a new thread id by copying stored history; emits `thread/started` for the new thread.219- `thread/fork` - fork a thread into a new thread id by copying stored history; emits `thread/started` for the new thread.

204- `thread/read` - read a stored thread by id without resuming it; set `includeTurns` to return full turn history. Returned `thread` objects include runtime `status`.220- `thread/read` - read a stored thread by id without resuming it; set `includeTurns` to return full turn history. Returned `thread` objects include runtime `status`.

205- `thread/list` - page through stored thread logs; supports cursor-based pagination plus `modelProviders`, `sourceKinds`, `archived`, and `cwd` filters. Returned `thread` objects include runtime `status`.221- `thread/list` - page through stored thread logs; supports cursor-based pagination plus `modelProviders`, `sourceKinds`, `archived`, `cwd`, and `searchTerm` filters. Returned `thread` objects include runtime `status`.

222- `thread/turns/list` - page through a stored thread's turn history without resuming it.

206- `thread/loaded/list` - list the thread ids currently loaded in memory.223- `thread/loaded/list` - list the thread ids currently loaded in memory.

224- `thread/name/set` - set or update a thread's user-facing name for a loaded thread or a persisted rollout; emits `thread/name/updated`.

225- `thread/metadata/update` - patch SQLite-backed stored thread metadata; currently supports persisted `gitInfo`.

207- `thread/archive` - move a thread's log file into the archived directory; returns `{}` on success and emits `thread/archived`.226- `thread/archive` - move a thread's log file into the archived directory; returns `{}` on success and emits `thread/archived`.

208- `thread/unsubscribe` - unsubscribe this connection from thread turn/item events. If this was the last subscriber, the server unloads the thread and emits `thread/closed`.227- `thread/unsubscribe` - unsubscribe this connection from thread turn/item events. If this was the last subscriber, the server unloads the thread after a no-subscriber inactivity grace period and emits `thread/closed`.

209- `thread/unarchive` - restore an archived thread rollout back into the active sessions directory; returns the restored `thread` and emits `thread/unarchived`.228- `thread/unarchive` - restore an archived thread rollout back into the active sessions directory; returns the restored `thread` and emits `thread/unarchived`.

210- `thread/status/changed` - notification emitted when a loaded thread's runtime `status` changes.229- `thread/status/changed` - notification emitted when a loaded thread's runtime `status` changes.

211- `thread/compact/start` - trigger conversation history compaction for a thread; returns `{}` immediately while progress streams via `turn/*` and `item/*` notifications.230- `thread/compact/start` - trigger conversation history compaction for a thread; returns `{}` immediately while progress streams via `turn/*` and `item/*` notifications.

231- `thread/shellCommand` - run a user-initiated shell command against a thread. This runs outside the sandbox with full access and doesn't inherit the thread sandbox policy.

232- `thread/backgroundTerminals/clean` - stop all running background terminals for a thread (experimental; requires `capabilities.experimentalApi`).

212- `thread/rollback` - drop the last N turns from the in-memory context and persist a rollback marker; returns the updated `thread`.233- `thread/rollback` - drop the last N turns from the in-memory context and persist a rollback marker; returns the updated `thread`.

213- `turn/start` - add user input to a thread and begin Codex generation; responds with the initial `turn` and streams events. For `collaborationMode`, `settings.developer_instructions: null` means "use built-in instructions for the selected mode."234- `turn/start` - add user input to a thread and begin Codex generation; responds with the initial `turn` and streams events. For `collaborationMode`, `settings.developer_instructions: null` means "use built-in instructions for the selected mode."

235- `thread/inject_items` - append raw Responses API items to a loaded thread's model-visible history without starting a user turn.

214- `turn/steer` - append user input to the active in-flight turn for a thread; returns the accepted `turnId`.236- `turn/steer` - append user input to the active in-flight turn for a thread; returns the accepted `turnId`.

215- `turn/interrupt` - request cancellation of an in-flight turn; success is `{}` and the turn ends with `status: "interrupted"`.237- `turn/interrupt` - request cancellation of an in-flight turn; success is `{}` and the turn ends with `status: "interrupted"`.

216- `review/start` - kick off the Codex reviewer for a thread; emits `enteredReviewMode` and `exitedReviewMode` items.238- `review/start` - kick off the Codex reviewer for a thread; emits `enteredReviewMode` and `exitedReviewMode` items.

217- `command/exec` - run a single command under the server sandbox without starting a thread/turn.239- `command/exec` - run a single command under the server sandbox without starting a thread/turn.

240- `command/exec/write` - write `stdin` bytes to a running `command/exec` session or close `stdin`.

241- `command/exec/resize` - resize a running PTY-backed `command/exec` session.

242- `command/exec/terminate` - stop a running `command/exec` session.

243- `command/exec/outputDelta` (notify) - emitted for base64-encoded stdout/stderr chunks from a streaming `command/exec` session.

218- `model/list` - list available models (set `includeHidden: true` to include entries with `hidden: true`) with effort options, optional `upgrade`, and `inputModalities`.244- `model/list` - list available models (set `includeHidden: true` to include entries with `hidden: true`) with effort options, optional `upgrade`, and `inputModalities`.

219- `experimentalFeature/list` - list feature flags with lifecycle stage metadata and cursor pagination.245- `experimentalFeature/list` - list feature flags with lifecycle stage metadata and cursor pagination.

246- `experimentalFeature/enablement/set` - patch in-memory runtime enablement for supported feature keys such as `apps` and `plugins`.

220- `collaborationMode/list` - list collaboration mode presets (experimental, no pagination).247- `collaborationMode/list` - list collaboration mode presets (experimental, no pagination).

221- `skills/list` - list skills for one or more `cwd` values (supports `forceReload` and optional `perCwdExtraUserRoots`).248- `skills/list` - list skills for one or more `cwd` values (supports `forceReload` and optional `perCwdExtraUserRoots`).

249- `skills/changed` (notify) - emitted when watched local skill files change.

250- `marketplace/add` - add a remote plugin marketplace and persist it into the user's marketplace config.

251- `plugin/list` - list discovered plugin marketplaces and plugin state, including install/auth policy metadata, marketplace load errors, featured plugin ids, and local, Git, or remote plugin source metadata.

252- `plugin/read` - read one plugin by marketplace path or remote marketplace name and plugin name, including bundled skills, apps, and MCP server names when those details are available.

253- `plugin/install` - install a plugin from a marketplace path or remote marketplace name.

254- `plugin/uninstall` - uninstall an installed plugin.

222- `app/list` - list available apps (connectors) with pagination plus accessibility/enabled metadata.255- `app/list` - list available apps (connectors) with pagination plus accessibility/enabled metadata.

223- `skills/config/write` - enable or disable skills by path.256- `skills/config/write` - enable or disable skills by path.

224- `mcpServer/oauth/login` - start an OAuth login for a configured MCP server; returns an authorization URL and emits `mcpServer/oauthLogin/completed` on completion.257- `mcpServer/oauth/login` - start an OAuth login for a configured MCP server; returns an authorization URL and emits `mcpServer/oauthLogin/completed` on completion.

225- `tool/requestUserInput` - prompt the user with 1-3 short questions for a tool call (experimental); questions can set `isOther` for a free-form option.258- `tool/requestUserInput` - prompt the user with 1-3 short questions for a tool call (experimental); questions can set `isOther` for a free-form option.

226- `config/mcpServer/reload` - reload MCP server configuration from disk and queue a refresh for loaded threads.259- `config/mcpServer/reload` - reload MCP server configuration from disk and queue a refresh for loaded threads.

227- `mcpServerStatus/list` - list MCP servers, tools, resources, and auth status (cursor + limit pagination).260- `mcpServerStatus/list` - list MCP servers, tools, resources, and auth status (cursor + limit pagination). Use `detail: "full"` for full data or `detail: "toolsAndAuthOnly"` to omit resources.

261- `mcpServer/resource/read` - read a single MCP resource through an initialized MCP server.

262- `mcpServer/tool/call` - call a tool on a thread's configured MCP server.

263- `mcpServer/startupStatus/updated` (notify) - emitted when a configured MCP server's startup status changes for a loaded thread.

228- `windowsSandbox/setupStart` - start Windows sandbox setup for `elevated` or `unelevated` mode; returns quickly and later emits `windowsSandbox/setupCompleted`.264- `windowsSandbox/setupStart` - start Windows sandbox setup for `elevated` or `unelevated` mode; returns quickly and later emits `windowsSandbox/setupCompleted`.

229- `feedback/upload` - submit a feedback report (classification + optional reason/logs + conversation id, plus optional `extraLogFiles` attachments).265- `feedback/upload` - submit a feedback report (classification + optional reason/logs + conversation id, plus optional `extraLogFiles` attachments).

230- `config/read` - fetch the effective configuration on disk after resolving configuration layering.266- `config/read` - fetch the effective configuration on disk after resolving configuration layering.

231- `externalAgentConfig/detect` - detect migratable external-agent artifacts with `includeHome` and optional `cwds`; each detected item includes `cwd` (`null` for home).267- `externalAgentConfig/detect` - detect external-agent artifacts that can be migrated with `includeHome` and optional `cwds`; each detected item includes `cwd` (`null` for home).

232- `externalAgentConfig/import` - apply selected external-agent migration items by passing explicit `migrationItems` with `cwd` (`null` for home).268- `externalAgentConfig/import` - apply selected external-agent migration items by passing explicit `migrationItems` with `cwd` (`null` for home); plugin imports emit `externalAgentConfig/import/completed`.

233- `config/value/write` - write a single configuration key/value to the user's `config.toml` on disk.269- `config/value/write` - write a single configuration key/value to the user's `config.toml` on disk.

234- `config/batchWrite` - apply configuration edits atomically to the user's `config.toml` on disk.270- `config/batchWrite` - apply configuration edits atomically to the user's `config.toml` on disk.

235- `configRequirements/read` - fetch requirements from `requirements.toml` and/or MDM, including allow-lists and residency requirements (or `null` if you haven’t set any up).271- `configRequirements/read` - fetch requirements from `requirements.toml` and/or MDM, including allow-lists, pinned `featureRequirements`, and residency/network requirements (or `null` if you haven't set any up).

272- `fs/readFile`, `fs/writeFile`, `fs/createDirectory`, `fs/getMetadata`, `fs/readDirectory`, `fs/remove`, `fs/copy`, `fs/watch`, `fs/unwatch`, and `fs/changed` (notify) - operate on absolute filesystem paths through the app-server v2 filesystem API.

273 

274Plugin summaries include a `source` union. Local plugins return

275`{ "type": "local", "path": ... }`, Git-backed marketplace entries return

276`{ "type": "git", "url": ..., "path": ..., "refName": ..., "sha": ... }`,

277and remote catalog entries return `{ "type": "remote" }`. For remote-only

278catalog entries, `PluginMarketplaceEntry.path` can be `null`; pass

279`remoteMarketplaceName` instead of `marketplacePath` when reading or installing

280those plugins.

236 281 

237## Models282## Models

238 283 

244{ "method": "model/list", "id": 6, "params": { "limit": 20, "includeHidden": false } }289{ "method": "model/list", "id": 6, "params": { "limit": 20, "includeHidden": false } }

245{ "id": 6, "result": {290{ "id": 6, "result": {

246 "data": [{291 "data": [{

247 "id": "gpt-5.2-codex",292 "id": "gpt-5.4",

248 "model": "gpt-5.2-codex",293 "model": "gpt-5.4",

249 "upgrade": "gpt-5.3-codex",294 "displayName": "GPT-5.4",

250 "displayName": "GPT-5.2 Codex",

251 "hidden": false,295 "hidden": false,

252 "defaultReasoningEffort": "medium",296 "defaultReasoningEffort": "medium",

253 "reasoningEffort": [{297 "supportedReasoningEfforts": [{

254 "effort": "low",298 "reasoningEffort": "low",

255 "description": "Lower latency"299 "description": "Lower latency"

256 }],300 }],

257 "inputModalities": ["text", "image"],301 "inputModalities": ["text", "image"],

264 308 

265Each model entry can include:309Each model entry can include:

266 310 

267- `reasoningEffort` - supported effort options for the model.311- `supportedReasoningEfforts` - supported effort options for the model.

268- `defaultReasoningEffort` - suggested default effort for clients.312- `defaultReasoningEffort` - suggested default effort for clients.

269- `upgrade` - optional recommended upgrade model id for migration prompts in clients.313- `upgrade` - optional recommended upgrade model id for migration prompts in clients.

314- `upgradeInfo` - optional upgrade metadata for migration prompts in clients.

270- `hidden` - whether the model is hidden from the default picker list.315- `hidden` - whether the model is hidden from the default picker list.

271- `inputModalities` - supported input types for the model (for example `text`, `image`).316- `inputModalities` - supported input types for the model (for example `text`, `image`).

272- `supportsPersonality` - whether the model supports personality-specific instructions such as `/personality`.317- `supportsPersonality` - whether the model supports personality-specific instructions such as `/personality`.

301## Threads346## Threads

302 347 

303- `thread/read` reads a stored thread without subscribing to it; set `includeTurns` to include turns.348- `thread/read` reads a stored thread without subscribing to it; set `includeTurns` to include turns.

304- `thread/list` supports cursor pagination plus `modelProviders`, `sourceKinds`, `archived`, and `cwd` filtering.349- `thread/turns/list` pages through a stored thread's turn history without resuming it.

350- `thread/list` supports cursor pagination plus `modelProviders`, `sourceKinds`, `archived`, `cwd`, and `searchTerm` filtering.

305- `thread/loaded/list` returns the thread IDs currently in memory.351- `thread/loaded/list` returns the thread IDs currently in memory.

306- `thread/archive` moves the thread's persisted JSONL log into the archived directory.352- `thread/archive` moves the thread's persisted JSONL log into the archived directory.

307- `thread/unsubscribe` unsubscribes the current connection from a loaded thread and can trigger `thread/closed`.353- `thread/metadata/update` patches stored thread metadata, currently including persisted `gitInfo`.

354- `thread/unsubscribe` unsubscribes the current connection from a loaded thread and can trigger `thread/closed` after an inactivity grace period.

308- `thread/unarchive` restores an archived thread rollout back into the active sessions directory.355- `thread/unarchive` restores an archived thread rollout back into the active sessions directory.

309- `thread/compact/start` triggers compaction and returns `{}` immediately.356- `thread/compact/start` triggers compaction and returns `{}` immediately.

310- `thread/rollback` drops the last N turns from the in-memory context and records a rollback marker in the thread's persisted JSONL log.357- `thread/rollback` drops the last N turns from the in-memory context and records a rollback marker in the thread's persisted JSONL log.

358- `thread/inject_items` appends raw Responses API items to a loaded thread's model-visible history without starting a user turn.

311 359 

312### Start or resume a thread360### Start or resume a thread

313 361 

315 363 

316```json364```json

317{ "method": "thread/start", "id": 10, "params": {365{ "method": "thread/start", "id": 10, "params": {

318 "model": "gpt-5.1-codex",366 "model": "gpt-5.4",

319 "cwd": "/Users/me/project",367 "cwd": "/Users/me/project",

320 "approvalPolicy": "never",368 "approvalPolicy": "never",

321 "sandbox": "workspaceWrite",369 "sandbox": "workspaceWrite",

378 426 

379Unlike `thread/resume`, `thread/read` doesn't load the thread into memory or emit `thread/started`.427Unlike `thread/resume`, `thread/read` doesn't load the thread into memory or emit `thread/started`.

380 428 

429### List thread turns

430 

431Use `thread/turns/list` to page a stored thread's turn history without resuming it. Results default to newest-first so clients can fetch older turns with `nextCursor`. The response also includes `backwardsCursor`; pass it as `cursor` with `sortDirection: "asc"` to fetch turns newer than the first item from the earlier page.

432 

433```json

434{ "method": "thread/turns/list", "id": 20, "params": {

435 "threadId": "thr_123",

436 "limit": 50,

437 "sortDirection": "desc"

438} }

439{ "id": 20, "result": {

440 "data": [],

441 "nextCursor": "older-turns-cursor-or-null",

442 "backwardsCursor": "newer-turns-cursor-or-null"

443} }

444```

445 

381### List threads (with pagination & filters)446### List threads (with pagination & filters)

382 447 

383`thread/list` lets you render a history UI. Results default to newest-first by `createdAt`. Filters apply before pagination. Pass any combination of:448`thread/list` lets you render a history UI. Results default to newest-first by `createdAt`. Filters apply before pagination. Pass any combination of:

389- `sourceKinds` - restrict results to specific thread sources. When omitted or `[]`, the server defaults to interactive sources only: `cli` and `vscode`.454- `sourceKinds` - restrict results to specific thread sources. When omitted or `[]`, the server defaults to interactive sources only: `cli` and `vscode`.

390- `archived` - when `true`, list archived threads only. When `false` or omitted, list non-archived threads (default).455- `archived` - when `true`, list archived threads only. When `false` or omitted, list non-archived threads (default).

391- `cwd` - restrict results to threads whose session current working directory exactly matches this path.456- `cwd` - restrict results to threads whose session current working directory exactly matches this path.

457- `searchTerm` - search stored thread summaries and metadata before pagination.

392 458 

393`sourceKinds` accepts the following values:459`sourceKinds` accepts the following values:

394 460 

422 488 

423When `nextCursor` is `null`, you have reached the final page.489When `nextCursor` is `null`, you have reached the final page.

424 490 

491### Update stored thread metadata

492 

493Use `thread/metadata/update` to patch stored thread metadata without resuming the thread. Today this supports persisted `gitInfo`; omitted fields are left unchanged, and explicit `null` clears a stored value.

494 

495```json

496{ "method": "thread/metadata/update", "id": 21, "params": {

497 "threadId": "thr_123",

498 "gitInfo": { "branch": "feature/sidebar-pr" }

499} }

500{ "id": 21, "result": {

501 "thread": {

502 "id": "thr_123",

503 "gitInfo": { "sha": null, "branch": "feature/sidebar-pr", "originUrl": null }

504 }

505} }

506```

507 

425### Track thread status changes508### Track thread status changes

426 509 

427`thread/status/changed` is emitted whenever a loaded thread's runtime status changes. The payload includes `threadId` and the new `status`.510`thread/status/changed` is emitted whenever a loaded thread's runtime status changes. The payload includes `threadId` and the new `status`.

450`thread/unsubscribe` removes the current connection's subscription to a thread. The response status is one of:533`thread/unsubscribe` removes the current connection's subscription to a thread. The response status is one of:

451 534 

452- `unsubscribed` when the connection was subscribed and is now removed.535- `unsubscribed` when the connection was subscribed and is now removed.

453- `notSubscribed` when the connection was not subscribed to that thread.536- `notSubscribed` when the connection wasn't subscribed to that thread.

454- `notLoaded` when the thread is not loaded.537- `notLoaded` when the thread isn't loaded.

455 538 

456If this was the last subscriber, the server unloads the thread and emits a `thread/status/changed` transition to `notLoaded` plus `thread/closed`.539If this was the last subscriber, the server keeps the thread loaded until it has no subscribers and no thread activity for 30 minutes. When the grace period expires, app-server unloads the thread and emits a `thread/status/changed` transition to `notLoaded` plus `thread/closed`.

457 540 

458```json541```json

459{ "method": "thread/unsubscribe", "id": 22, "params": { "threadId": "thr_123" } }542{ "method": "thread/unsubscribe", "id": 22, "params": { "threadId": "thr_123" } }

460{ "id": 22, "result": { "status": "unsubscribed" } }543{ "id": 22, "result": { "status": "unsubscribed" } }

544```

545 

546If the thread later expires:

547 

548```json

461{ "method": "thread/status/changed", "params": {549{ "method": "thread/status/changed", "params": {

462 "threadId": "thr_123",550 "threadId": "thr_123",

463 "status": { "type": "notLoaded" }551 "status": { "type": "notLoaded" }

498{ "id": 25, "result": {} }586{ "id": 25, "result": {} }

499```587```

500 588 

589### Run a thread shell command

590 

591Use `thread/shellCommand` for user-initiated shell commands that belong to a thread. The request returns immediately with `{}` while progress streams through standard `turn/*` and `item/*` notifications.

592 

593This API runs outside the sandbox with full access and doesn't inherit the thread sandbox policy. Clients should expose it only for explicit user-initiated commands.

594 

595If the thread already has an active turn, the command runs as an auxiliary action on that turn and its formatted output is injected into the turn's message stream. If the thread is idle, app-server starts a standalone turn for the shell command.

596 

597```json

598{ "method": "thread/shellCommand", "id": 26, "params": { "threadId": "thr_b", "command": "git status --short" } }

599{ "id": 26, "result": {} }

600```

601 

602### Clean background terminals

603 

604Use `thread/backgroundTerminals/clean` to stop all running background terminals associated with a thread. This method is experimental and requires `capabilities.experimentalApi = true`.

605 

606```json

607{ "method": "thread/backgroundTerminals/clean", "id": 27, "params": { "threadId": "thr_b" } }

608{ "id": 27, "result": {} }

609```

610 

501### Roll back recent turns611### Roll back recent turns

502 612 

503Use `thread/rollback` to remove the last `numTurns` entries from the in-memory context and persist a rollback marker in the rollout log. The returned `thread` includes `turns` populated after the rollback.613Use `thread/rollback` to remove the last `numTurns` entries from the in-memory context and persist a rollback marker in the rollout log. The returned `thread` includes `turns` populated after the rollback.

504 614 

505```json615```json

506{ "method": "thread/rollback", "id": 26, "params": { "threadId": "thr_b", "numTurns": 1 } }616{ "method": "thread/rollback", "id": 28, "params": { "threadId": "thr_b", "numTurns": 1 } }

507{ "id": 26, "result": { "thread": { "id": "thr_b", "name": "Bug bash notes", "ephemeral": false } } }617{ "id": 28, "result": { "thread": { "id": "thr_b", "name": "Bug bash notes", "ephemeral": false } } }

508```618```

509 619 

510## Turns620## Turns

570 "writableRoots": ["/Users/me/project"],680 "writableRoots": ["/Users/me/project"],

571 "networkAccess": true681 "networkAccess": true

572 },682 },

573 "model": "gpt-5.1-codex",683 "model": "gpt-5.4",

574 "effort": "medium",684 "effort": "medium",

575 "summary": "concise",685 "summary": "concise",

576 "personality": "friendly",686 "personality": "friendly",

584{ "id": 30, "result": { "turn": { "id": "turn_456", "status": "inProgress", "items": [], "error": null } } }694{ "id": 30, "result": { "turn": { "id": "turn_456", "status": "inProgress", "items": [], "error": null } } }

585```695```

586 696 

697### Inject items into a thread

698 

699Use `thread/inject_items` to append prebuilt Responses API items to a loaded thread's prompt history without starting a user turn. These items are persisted to the rollout and included in subsequent model requests.

700 

701```json

702{ "method": "thread/inject_items", "id": 31, "params": {

703 "threadId": "thr_123",

704 "items": [

705 {

706 "type": "message",

707 "role": "assistant",

708 "content": [{ "type": "output_text", "text": "Previously computed context." }]

709 }

710 ]

711} }

712{ "id": 31, "result": {} }

713```

714 

587### Steer an active turn715### Steer an active turn

588 716 

589Use `turn/steer` to append more user input to the active in-flight turn.717Use `turn/steer` to append more user input to the active in-flight turn.

713- The server rejects empty `command` arrays.841- The server rejects empty `command` arrays.

714- `sandboxPolicy` accepts the same shape used by `turn/start` (for example, `dangerFullAccess`, `readOnly`, `workspaceWrite`, `externalSandbox`).842- `sandboxPolicy` accepts the same shape used by `turn/start` (for example, `dangerFullAccess`, `readOnly`, `workspaceWrite`, `externalSandbox`).

715- When omitted, `timeoutMs` falls back to the server default.843- When omitted, `timeoutMs` falls back to the server default.

844- Set `tty: true` for PTY-backed sessions, and use `processId` when you plan to follow up with `command/exec/write`, `command/exec/resize`, or `command/exec/terminate`.

845- Set `streamStdoutStderr: true` to receive `command/exec/outputDelta` notifications while the command is running.

716 846 

717### Read admin requirements (`configRequirements/read`)847### Read admin requirements (`configRequirements/read`)

718 848 

724 "requirements": {854 "requirements": {

725 "allowedApprovalPolicies": ["onRequest", "unlessTrusted"],855 "allowedApprovalPolicies": ["onRequest", "unlessTrusted"],

726 "allowedSandboxModes": ["readOnly", "workspaceWrite"],856 "allowedSandboxModes": ["readOnly", "workspaceWrite"],

857 "featureRequirements": {

858 "personality": true,

859 "unified_exec": false

860 },

727 "network": {861 "network": {

728 "enabled": true,862 "enabled": true,

729 "allowedDomains": ["api.openai.com"],863 "allowedDomains": ["api.openai.com"],

734} }868} }

735```869```

736 870 

737`result.requirements` is `null` when no requirements are configured. When present, the optional `network` object carries managed proxy constraints (domain rules, proxy settings, and unix-socket policy).871`result.requirements` is `null` when no requirements are configured. See the docs on [`requirements.toml`](https://developers.openai.com/codex/config-reference#requirementstoml) for details on supported keys and values.

738 872 

739### Windows sandbox setup (`windowsSandbox/setupStart`)873### Windows sandbox setup (`windowsSandbox/setupStart`)

740 874 

759- `elevated` - run the elevated Windows sandbox setup path.893- `elevated` - run the elevated Windows sandbox setup path.

760- `unelevated` - run the legacy setup/preflight path.894- `unelevated` - run the legacy setup/preflight path.

761 895 

896## Filesystem

897 

898The v2 filesystem APIs operate on absolute paths. Use `fs/watch` when a client needs to invalidate UI state after a file or directory changes.

899 

900```json

901{ "method": "fs/watch", "id": 54, "params": {

902 "watchId": "0195ec6b-1d6f-7c2e-8c7a-56f2c4a8b9d1",

903 "path": "/Users/me/project/.git/HEAD"

904} }

905{ "id": 54, "result": { "path": "/Users/me/project/.git/HEAD" } }

906{ "method": "fs/changed", "params": {

907 "watchId": "0195ec6b-1d6f-7c2e-8c7a-56f2c4a8b9d1",

908 "changedPaths": ["/Users/me/project/.git/HEAD"]

909} }

910{ "method": "fs/unwatch", "id": 55, "params": {

911 "watchId": "0195ec6b-1d6f-7c2e-8c7a-56f2c4a8b9d1"

912} }

913{ "id": 55, "result": {} }

914```

915 

916Watching a file emits `fs/changed` for that file path, including updates delivered by replace or rename operations.

917 

762## Events918## Events

763 919 

764Event notifications are the server-initiated stream for thread lifecycles, turn lifecycles, and the items within them. After you start or resume a thread, keep reading the active transport stream for `thread/started`, `thread/archived`, `thread/unarchived`, `thread/closed`, `thread/status/changed`, `turn/*`, `item/*`, and `serverRequest/resolved` notifications.920Event notifications are the server-initiated stream for thread lifecycles, turn lifecycles, and the items within them. After you start or resume a thread, keep reading the active transport stream for `thread/started`, `thread/archived`, `thread/unarchived`, `thread/closed`, `thread/status/changed`, `turn/*`, `item/*`, and `serverRequest/resolved` notifications.

769 925 

770- Exact-match only: `item/agentMessage/delta` suppresses only that method.926- Exact-match only: `item/agentMessage/delta` suppresses only that method.

771- Unknown method names are ignored.927- Unknown method names are ignored.

772- Applies to both legacy (`codex/event/*`) and v2 (`thread/*`, `turn/*`, `item/*`, etc.) notifications.928- Applies to the current `thread/*`, `turn/*`, `item/*`, and related v2 notifications.

773- Doesn't apply to requests, responses, or errors.929- Doesn't apply to requests, responses, or errors.

774 930 

775### Fuzzy file search events (experimental)931### Fuzzy file search events (experimental)

868 1024 

869When `networkApprovalContext` is present, the prompt is for managed network access (not a general shell-command approval). The current v2 schema exposes the target `host` and `protocol`; clients should render a network-specific prompt and not rely on `command` being a user-meaningful shell command preview.1025When `networkApprovalContext` is present, the prompt is for managed network access (not a general shell-command approval). The current v2 schema exposes the target `host` and `protocol`; clients should render a network-specific prompt and not rely on `command` being a user-meaningful shell command preview.

870 1026 

871Codex deduplicates concurrent network approval prompts by destination (`host`, protocol, and port). The app-server may therefore send one prompt that unblocks multiple queued requests to the same destination, while different ports on the same host are treated separately.1027Codex groups concurrent network approval prompts by destination (`host`, protocol, and port). The app-server may therefore send one prompt that unblocks multiple queued requests to the same destination, while different ports on the same host are treated separately.

872 1028 

873### File change approvals1029### File change approvals

874 1030 

979} }1135} }

980```1136```

981 1137 

1138The server also emits `skills/changed` notifications when watched local skill files change. Treat this as an invalidation signal and rerun `skills/list` with your current params when needed.

1139 

982To enable or disable a skill by path:1140To enable or disable a skill by path:

983 1141 

984```json1142```json

1145 1303 

1146### Detect and import external agent config1304### Detect and import external agent config

1147 1305 

1148Use `externalAgentConfig/detect` to discover migratable external-agent artifacts, then pass the selected entries to `externalAgentConfig/import`.1306Use `externalAgentConfig/detect` to discover external-agent artifacts that can be migrated, then pass the selected entries to `externalAgentConfig/import`.

1149 1307 

1150Detection example:1308Detection example:

1151 1309 

1185{ "id": 64, "result": {} }1343{ "id": 64, "result": {} }

1186```1344```

1187 1345 

1188Supported `itemType` values are `AGENTS_MD`, `CONFIG`, `SKILLS`, and `MCP_SERVER_CONFIG`. Detection returns only items that still have work to do. For example, AGENTS migration is skipped when `AGENTS.md` already exists and is non-empty, and skill imports do not overwrite existing skill directories.1346When a request includes plugin imports, the server emits `externalAgentConfig/import/completed` after the import finishes. This notification may arrive immediately after the response or after background remote imports complete.

1347 

1348Supported `itemType` values are `AGENTS_MD`, `CONFIG`, `SKILLS`, `PLUGINS`,

1349and `MCP_SERVER_CONFIG`. For `PLUGINS` items, `details.plugins` lists each

1350`marketplaceName` and the `pluginNames` Codex can try to migrate. Detection

1351returns only items that still have work to do. For example, Codex skips AGENTS

1352migration when `AGENTS.md` already exists and is non-empty, and skill imports

1353don't overwrite existing skill directories.

1354 

1355When detecting plugins from `.claude/settings.json`, Codex reads configured

1356marketplace sources from `extraKnownMarketplaces`. If `enabledPlugins` contains

1357plugins from `claude-plugins-official` but the marketplace source is missing,

1358Codex infers `anthropics/claude-plugins-official` as the source.

1189 1359 

1190## Auth endpoints1360## Auth endpoints

1191 1361 

1192The JSON-RPC auth/account surface exposes request/response methods plus server-initiated notifications (no `id`). Use these to determine auth state, start or cancel logins, logout, and inspect ChatGPT rate limits.1362The JSON-RPC auth/account surface exposes request/response methods plus server-initiated notifications (no `id`). Use these to determine auth state, start or cancel logins, logout, inspect ChatGPT rate limits, and notify workspace owners about depleted credits or usage limits.

1193 1363 

1194### Authentication modes1364### Authentication modes

1195 1365 

1196Codex supports three authentication modes. `account/updated.authMode` shows the active mode, and `account/read` also reports it.1366Codex supports these authentication modes. `account/updated.authMode` shows the active mode and includes the current ChatGPT `planType` when available. `account/read` also reports account and plan details.

1197 1367 

1198- **API key (`apikey`)** - the caller supplies an OpenAI API key and Codex stores it for API requests.1368- **API key (`apikey`)** - the caller supplies an OpenAI API key with `type: "apiKey"`, and Codex stores it for API requests.

1199- **ChatGPT managed (`chatgpt`)** - Codex owns the ChatGPT OAuth flow, persists tokens, and refreshes them automatically.1369- **ChatGPT managed (`chatgpt`)** - Codex owns the ChatGPT OAuth flow, persists tokens, and refreshes them automatically. Start with `type: "chatgpt"` for the browser flow or `type: "chatgptDeviceCode"` for the device-code flow.

1200- **ChatGPT external tokens (`chatgptAuthTokens`)** - a host app supplies `idToken` and `accessToken` directly. Codex stores these tokens in memory, and the host app must refresh them when asked.1370- **ChatGPT external tokens (`chatgptAuthTokens`)** - experimental and intended for host apps that already own the user's ChatGPT auth lifecycle. The host app supplies an `accessToken`, `chatgptAccountId`, and optional `chatgptPlanType` directly, and must refresh the token when asked.

1201 1371 

1202### API overview1372### API overview

1203 1373 

1204- `account/read` - fetch current account info; optionally refresh tokens.1374- `account/read` - fetch current account info; optionally refresh tokens.

1205- `account/login/start` - begin login (`apiKey`, `chatgpt`, or `chatgptAuthTokens`).1375- `account/login/start` - begin login (`apiKey`, `chatgpt`, `chatgptDeviceCode`, or experimental `chatgptAuthTokens`).

1206- `account/login/completed` (notify) - emitted when a login attempt finishes (success or error).1376- `account/login/completed` (notify) - emitted when a login attempt finishes (success or error).

1207- `account/login/cancel` - cancel a pending ChatGPT login by `loginId`.1377- `account/login/cancel` - cancel a pending managed ChatGPT login by `loginId`.

1208- `account/logout` - sign out; triggers `account/updated`.1378- `account/logout` - sign out; triggers `account/updated`.

1209- `account/updated` (notify) - emitted whenever auth mode changes (`authMode`: `apikey`, `chatgpt`, `chatgptAuthTokens`, or `null`).1379- `account/updated` (notify) - emitted whenever auth mode changes (`authMode`: `apikey`, `chatgpt`, `chatgptAuthTokens`, or `null`) and includes `planType` when available.

1210- `account/chatgptAuthTokens/refresh` (server request) - request fresh externally managed ChatGPT tokens after an authorization error.1380- `account/chatgptAuthTokens/refresh` (server request) - request fresh externally managed ChatGPT tokens after an authorization error.

1211- `account/rateLimits/read` - fetch ChatGPT rate limits.1381- `account/rateLimits/read` - fetch ChatGPT rate limits.

1212- `account/rateLimits/updated` (notify) - emitted whenever a user's ChatGPT rate limits change.1382- `account/rateLimits/updated` (notify) - emitted whenever a user's ChatGPT rate limits change.

1383- `account/sendAddCreditsNudgeEmail` - ask ChatGPT to email a workspace owner about depleted credits or a reached usage limit.

1213- `mcpServer/oauthLogin/completed` (notify) - emitted after a `mcpServer/oauth/login` flow finishes; payload includes `{ name, success, error? }`.1384- `mcpServer/oauthLogin/completed` (notify) - emitted after a `mcpServer/oauth/login` flow finishes; payload includes `{ name, success, error? }`.

1385- `mcpServer/startupStatus/updated` (notify) - emitted when a configured MCP server's startup status changes for a loaded thread; payload includes `{ name, status, error }`.

1214 1386 

1215### 1) Check auth state1387### 1) Check auth state

1216 1388 

1282 ```1454 ```

1283 1455 

1284 ```json1456 ```json

1285 { "method": "account/updated", "params": { "authMode": "apikey" } }1457 {

1458 "method": "account/updated",

1459 "params": { "authMode": "apikey", "planType": null }

1460 }

1286 ```1461 ```

1287 1462 

1288### 3) Log in with ChatGPT (browser flow)1463### 3) Log in with ChatGPT (browser flow)

1314 ```1489 ```

1315 1490 

1316 ```json1491 ```json

1317 { "method": "account/updated", "params": { "authMode": "chatgpt" } }1492 {

1493 "method": "account/updated",

1494 "params": { "authMode": "chatgpt", "planType": "plus" }

1495 }

1318 ```1496 ```

1319 1497 

1320### 3b) Log in with externally managed ChatGPT tokens (`chatgptAuthTokens`)1498### 3b) Log in with ChatGPT (device-code flow)

1499 

1500Use this flow when your client owns the sign-in ceremony or when a browser callback is brittle.

1501 

15021. Start:

1321 1503 

1322Use this mode when a host application owns the user’s ChatGPT auth lifecycle and supplies tokens directly.1504 ```json

1505 {

1506 "method": "account/login/start",

1507 "id": 4,

1508 "params": { "type": "chatgptDeviceCode" }

1509 }

1510 ```

1511 

1512 ```json

1513 {

1514 "id": 4,

1515 "result": {

1516 "type": "chatgptDeviceCode",

1517 "loginId": "<uuid>",

1518 "verificationUrl": "https://auth.openai.com/codex/device",

1519 "userCode": "ABCD-1234"

1520 }

1521 }

1522 ```

15232. Show `verificationUrl` and `userCode` to the user; the frontend owns the UX.

15243. Wait for notifications:

1525 

1526 ```json

1527 {

1528 "method": "account/login/completed",

1529 "params": { "loginId": "<uuid>", "success": true, "error": null }

1530 }

1531 ```

1532 

1533 ```json

1534 {

1535 "method": "account/updated",

1536 "params": { "authMode": "chatgpt", "planType": "plus" }

1537 }

1538 ```

1539 

1540### 3c) Log in with externally managed ChatGPT tokens (`chatgptAuthTokens`)

1541 

1542Use this experimental mode only when a host application owns the user's ChatGPT auth lifecycle and supplies tokens directly. Clients must set `capabilities.experimentalApi = true` during `initialize` before using this login type.

1323 1543 

13241. Send:15441. Send:

1325 1545 

1329 "id": 7,1549 "id": 7,

1330 "params": {1550 "params": {

1331 "type": "chatgptAuthTokens",1551 "type": "chatgptAuthTokens",

1332 "idToken": "<jwt>",1552 "accessToken": "<jwt>",

1333 "accessToken": "<jwt>"1553 "chatgptAccountId": "org-123",

1554 "chatgptPlanType": "business"

1334 }1555 }

1335 }1556 }

1336 ```1557 ```

1351 ```json1572 ```json

1352 {1573 {

1353 "method": "account/updated",1574 "method": "account/updated",

1354 "params": { "authMode": "chatgptAuthTokens" }1575 "params": { "authMode": "chatgptAuthTokens", "planType": "business" }

1355 }1576 }

1356 ```1577 ```

1357 1578 

1363 "id": 8,1584 "id": 8,

1364 "params": { "reason": "unauthorized", "previousAccountId": "org-123" }1585 "params": { "reason": "unauthorized", "previousAccountId": "org-123" }

1365}1586}

1366{ "id": 8, "result": { "idToken": "<jwt>", "accessToken": "<jwt>" } }1587{ "id": 8, "result": { "accessToken": "<jwt>", "chatgptAccountId": "org-123", "chatgptPlanType": "business" } }

1367```1588```

1368 1589 

1369The server retries the original request after a successful refresh response. Requests time out after about 10 seconds.1590The server retries the original request after a successful refresh response. Requests time out after about 10 seconds.

1380```json1601```json

1381{ "method": "account/logout", "id": 5 }1602{ "method": "account/logout", "id": 5 }

1382{ "id": 5, "result": {} }1603{ "id": 5, "result": {} }

1383{ "method": "account/updated", "params": { "authMode": null } }1604{ "method": "account/updated", "params": { "authMode": null, "planType": null } }

1384```1605```

1385 1606 

1386### 6) Rate limits (ChatGPT)1607### 6) Rate limits (ChatGPT)

1392 "limitId": "codex",1613 "limitId": "codex",

1393 "limitName": null,1614 "limitName": null,

1394 "primary": { "usedPercent": 25, "windowDurationMins": 15, "resetsAt": 1730947200 },1615 "primary": { "usedPercent": 25, "windowDurationMins": 15, "resetsAt": 1730947200 },

1395 "secondary": null1616 "secondary": null,

1617 "rateLimitReachedType": null

1396 },1618 },

1397 "rateLimitsByLimitId": {1619 "rateLimitsByLimitId": {

1398 "codex": {1620 "codex": {

1399 "limitId": "codex",1621 "limitId": "codex",

1400 "limitName": null,1622 "limitName": null,

1401 "primary": { "usedPercent": 25, "windowDurationMins": 15, "resetsAt": 1730947200 },1623 "primary": { "usedPercent": 25, "windowDurationMins": 15, "resetsAt": 1730947200 },

1402 "secondary": null1624 "secondary": null,

1625 "rateLimitReachedType": null

1403 },1626 },

1404 "codex_other": {1627 "codex_other": {

1405 "limitId": "codex_other",1628 "limitId": "codex_other",

1406 "limitName": "codex_other",1629 "limitName": "codex_other",

1407 "primary": { "usedPercent": 42, "windowDurationMins": 60, "resetsAt": 1730950800 },1630 "primary": { "usedPercent": 42, "windowDurationMins": 60, "resetsAt": 1730950800 },

1408 "secondary": null1631 "secondary": null,

1632 "rateLimitReachedType": null

1409 }1633 }

1410 }1634 }

1411} }1635} }

1426- `usedPercent` is current usage within the quota window.1650- `usedPercent` is current usage within the quota window.

1427- `windowDurationMins` is the quota window length.1651- `windowDurationMins` is the quota window length.

1428- `resetsAt` is a Unix timestamp (seconds) for the next reset.1652- `resetsAt` is a Unix timestamp (seconds) for the next reset.

1653- `planType` is included when the backend returns the ChatGPT plan associated with a bucket.

1654- `credits` is included when the backend returns remaining workspace credit details.

1655- `rateLimitReachedType` identifies the backend-classified limit state when one has been reached.

1656 

1657### 7) Notify a workspace owner about a limit

1658 

1659Use `account/sendAddCreditsNudgeEmail` to ask ChatGPT to email a workspace owner when credits are depleted or a usage limit has been reached.

1660 

1661```json

1662{ "method": "account/sendAddCreditsNudgeEmail", "id": 7, "params": { "creditType": "credits" } }

1663{ "id": 7, "result": { "status": "sent" } }

1664```

1665 

1666Use `creditType: "credits"` when workspace credits are depleted, or `creditType: "usage_limit"` when the workspace usage limit has been reached. If the owner was already notified recently, the response status is `cooldown_active`.

2 2 

3Automate recurring tasks in the background. Codex adds findings to the inbox, or automatically archives the task if there's nothing to report. You can combine automations with [skills](https://developers.openai.com/codex/skills) for more complex tasks.3Automate recurring tasks in the background. Codex adds findings to the inbox, or automatically archives the task if there's nothing to report. You can combine automations with [skills](https://developers.openai.com/codex/skills) for more complex tasks.

4 4 

5Automations run locally in the Codex app. The app needs to be running, and the5For project-scoped automations, the app needs to be running, and the selected

6selected project needs to be available on disk.6project needs to be available on disk.

7 7 

8In Git repositories, each automation run starts in a new8In Git repositories, you can choose whether an automation runs in your local

9[worktree](https://developers.openai.com/codex/app/worktrees) so it doesn’t interfere with your main9project or on a new [worktree](https://developers.openai.com/codex/app/worktrees). Both options run in the

10checkout. In non-version-controlled projects, automations run directly in the10background. Worktrees keep automation changes separate from unfinished local

11work, while running in your local project can modify files you are still

12working on. In non-version-controlled projects, automations run directly in the

11project directory.13project directory.

12 14 

13![Automation creation form with schedule and prompt fields](/images/codex/app/create-automation-light.webp)15You can also leave the model and reasoning effort on their default settings, or

16choose them explicitly if you want more control over how the automation runs.

17 

18![Automation creation form with schedule and prompt fields](/images/codex/app/codex-automations-light.webp)

14 19 

15## Managing tasks20## Managing tasks

16 21 

17All automations and their runs can be found in the automations pane inside your Codex app sidebar.22Find all automations and their runs in the automations pane inside your Codex app sidebar.

18 23 

19The "Triage" section acts as your inbox. Automation runs with findings show up there, and you can filter your inbox to show all automation runs or only unread ones.24The "Triage" section acts as your inbox. Automation runs with findings show up there, and you can filter your inbox to show all automation runs or only unread ones.

20 25 

21When an automation runs in a Git repository, Codex uses a dedicated background [worktree](https://developers.openai.com/codex/app/features#worktree-support). In non-version-controlled projects, automations run directly in the project directory. Consider using Git to enable running on background worktrees. You can have the same automation run on multiple projects.26Standalone automations start fresh runs on a schedule and report results in

27Triage. Use them when each run should be independent or when one automation

28should run across one or more projects. If you need a custom cadence, choose a

29custom schedule and enter cron syntax.

30 

31For Git repositories, each automation can run either in your local project or

32on a dedicated background [worktree](https://developers.openai.com/codex/app/features#worktree-support). Use

33worktrees when you want to isolate automation changes from unfinished local

34work. Use local mode when you want the automation to work directly in your main

35checkout, keeping in mind that it can change files you are actively editing.

36In non-version-controlled projects, automations run directly in the project

37directory. You can have the same automation run on more than one project.

22 38 

23Automations use your default sandbox settings. In read-only mode, tool calls fail if they require modifying files, network access, or working with apps on your computer. With full access enabled, background automations carry elevated risk. You can adjust sandbox settings in [Settings](https://developers.openai.com/codex/app/settings) and selectively allowlist commands with [rules](https://developers.openai.com/codex/rules).39Automations use your default sandbox settings. In read-only mode, tool calls fail if they require modifying files, network access, or working with apps on your computer. With full access enabled, background automations carry elevated risk. You can adjust sandbox settings in [Settings](https://developers.openai.com/codex/app/settings) and selectively allowlist commands with [rules](https://developers.openai.com/codex/rules).

24 40 

25To keep automations maintainable and shareable across teams, you can use [skills](https://developers.openai.com/codex/skills) to define the action and provide tools and context to Codex. You can explicitly trigger a skill as part of an automation by using `$skill-name` inside your automation.41Automations can use the same plugins and skills available to Codex. To keep

42automations maintainable and shareable across teams, use [skills](https://developers.openai.com/codex/skills)

43to define the action and provide tools and context. You can explicitly trigger a

44skill as part of an automation by using `$skill-name` inside your automation.

45 

46## Ask Codex to create or update automations

47 

48You can create and update automations from a regular Codex thread. Describe the

49task, the schedule, and whether the automation should stay attached to the

50current thread or start fresh runs. Codex can draft the automation prompt, choose

51the right automation type, and update it when the scope or cadence changes.

52 

53For example, ask Codex to remind you in this thread while a deployment finishes,

54or ask it to create a standalone automation that checks a project on a recurring

55schedule.

56 

57Skills can also create or update automations. For example, a skill for

58babysitting a pull request could set up a recurring automation that checks the

59PR status with the GitHub plugin and fixes new review feedback.

60 

61## Thread automations

62 

63Thread automations are heartbeat-style recurring wake-up calls attached to the

64current thread. Use them when you want Codex to keep returning to the same

65conversation on a schedule.

66 

67Use a thread automation when the scheduled work should preserve the thread's

68context instead of starting from a new prompt each time.

69 

70Thread automations can use minute-based intervals for active follow-up loops,

71or daily and weekly schedules when you need a check-in at a specific time.

72 

73Thread automations are useful for:

74 

75- checking a long-running command until it finishes

76- polling Slack, GitHub, or another connected source when the results should

77 stay in the same thread

78- reminding Codex to continue a review loop at a fixed cadence

79- running a skill-driven workflow that uses plugins, such as checking PR status

80 and addressing new feedback

81- keeping a chat focused on an ongoing research or triage task

82 

83Use a standalone or project automation when each run should be independent,

84when it should run across more than one project, or when findings should appear

85as separate automation runs in Triage.

86 

87When you create a thread automation, make the prompt durable. It should

88describe what Codex should do each time the thread wakes up, how to decide

89whether there is anything important to report, and when to stop or ask you for

90input.

26 91 

27## Testing automations safely92## Test automations

28 93 

29Before you schedule an automation, test the prompt manually in a regular thread94Before you schedule an automation, test the prompt manually in a regular thread

30first. This helps you confirm:95first. This helps you confirm:

31 96 

32- The prompt is clear and scoped correctly.97- The prompt is clear and scoped correctly.

33- The selected model and tools behave as expected.98- The selected or default model, reasoning effort, and tools behave as expected.

34- The resulting diff is reviewable.99- The resulting diff is reviewable.

35 100 

36When you start scheduling runs, review the first few outputs closely and adjust101When you start scheduling runs, review the first few outputs and adjust the

37the prompt or cadence as needed.102prompt or cadence as needed.

38 103 

39## Worktree cleanup for automations104## Worktree cleanup for automations

40 105 

41For Git repositories, automations run in worktrees. Frequent schedules can106If you choose worktrees for Git repositories, frequent schedules can create

42create many worktrees over time. Archive automation runs you no longer need,107many worktrees over time. Archive automation runs you no longer need, and avoid

43and avoid pinning runs unless you intend to keep their worktrees.108pinning runs unless you intend to keep their worktrees.

44 109 

45## Permissions and security model110## Permissions and security model

46 111 

47Automations are designed to run unattended and use your default sandbox112Automations run unattended and use your default sandbox settings.

48settings.

49 113 

50- If your sandbox mode is **read-only**, tool calls fail if they require114- If your sandbox mode is **read-only**, tool calls fail if they require

51 modifying files, accessing network, or working with apps on your computer.115 modifying files, accessing network, or working with apps on your computer.

55 on your computer. You can selectively allowlist commands to run outside the119 on your computer. You can selectively allowlist commands to run outside the

56 sandbox using [rules](https://developers.openai.com/codex/rules).120 sandbox using [rules](https://developers.openai.com/codex/rules).

57- If your sandbox mode is **full access**, background automations carry121- If your sandbox mode is **full access**, background automations carry

58 elevated risk, as Codex may modify files, run commands, and access network122 elevated risk, as Codex may change files, run commands, and access network

59 without asking. Consider updating sandbox settings to workspace write, and123 without asking. Consider updating sandbox settings to workspace write, and

60 using [rules](https://developers.openai.com/codex/rules) to selectively define which commands the agent124 using [rules](https://developers.openai.com/codex/rules) to selectively define which commands the agent

61 can run with full access.125 can run with full access.

65[Admin-enforced requirements (`requirements.toml`)](https://developers.openai.com/codex/enterprise/managed-configuration#admin-enforced-requirements-requirementstoml).129[Admin-enforced requirements (`requirements.toml`)](https://developers.openai.com/codex/enterprise/managed-configuration#admin-enforced-requirements-requirementstoml).

66 130 

67Automations use `approval_policy = "never"` when your organization policy131Automations use `approval_policy = "never"` when your organization policy

68allows it. If `approval_policy = "never"` is disallowed by admin requirements,132allows it. If admin requirements disallow `approval_policy = "never"`,

69automations fall back to the approval behavior of your selected mode.133automations fall back to the approval behavior of your selected mode.

70 134 

71## Examples135## Examples

1# In-app browser

2 

3The in-app browser gives you and Codex a shared view of rendered web pages

4inside a thread. Use it when you're building or debugging a web app and want to

5preview pages and attach visual comments.

6 

7Use it for local development servers, file-backed previews, and public pages

8that don't require sign-in. For anything that depends on login state or browser

9extensions, use your regular browser.

10 

11Open the in-app browser from the toolbar, by clicking a URL, by navigating

12manually in the browser, or by pressing <kbd>Cmd</kbd>+<kbd>Shift</kbd>+<kbd>B</kbd>

13(<kbd>Ctrl</kbd>+<kbd>Shift</kbd>+<kbd>B</kbd> on Windows).

14 

15The in-app browser does not support authentication flows, signed-in pages,

16 your regular browser profile, cookies, extensions, or existing tabs. Use it

17 for pages Codex can open without logging in.

18 

19Treat page content as untrusted context. Don't paste secrets into browser flows.

20 

21![Codex app showing a browser comment on a local web app preview](/images/codex/app/in-app-browser-light.webp)

22 

23## Preview a page

24 

251. Start your app's development server in the [integrated terminal](https://developers.openai.com/codex/app/features#integrated-terminal) or with a [local environment action](https://developers.openai.com/codex/app/local-environments#actions).

262. Open an unauthenticated local route, file-backed page, or public page by

27 clicking a URL or navigating manually in the browser.

283. Review the rendered state alongside the code diff.

294. Leave browser comments on the elements or areas that need changes.

305. Ask Codex to address the comments and keep the scope narrow.

31 

32Example feedback:

33 

34```text

35I left comments on the pricing page in the in-app browser. Address the mobile

36layout issues and keep the card structure unchanged.

37```

38 

39## Comment on the page

40 

41When a bug is visible only in the rendered page, use browser comments to give

42Codex precise feedback on the page.

43 

44- Turn on comment mode, select an element or area, and submit a comment.

45- In comment mode, hold <kbd>Shift</kbd> and click to select an area.

46- Hold <kbd>Cmd</kbd> while clicking to send a comment immediately.

47 

48After you leave comments, send a message in the thread asking Codex to address

49them. Comments are most useful when Codex needs to make a precise visual change.

50 

51Good feedback is specific:

52 

53```text

54This button overflows on mobile. Keep the label on one line if it fits,

55otherwise wrap it without changing the card height.

56```

57 

58```text

59This tooltip covers the data point under the cursor. Reposition the tooltip so

60it stays inside the chart bounds.

61```

62 

63## Keep browser tasks scoped

64 

65The in-app browser is for review and iteration. Keep each browser task small

66enough to review in one pass.

67 

68- Name the page, route, or local URL.

69- Name the visual state you care about, such as loading, empty, error, or

70 success.

71- Leave comments on the exact elements or areas that need changes.

72- Review the updated route after Codex changes the code.

73- Ask Codex to start or check the dev server before it uses the browser.

74 

75For repository changes, use the [review pane](https://developers.openai.com/codex/app/review) to inspect the

76changes and leave comments.

36 36 

37You can also explicitly invoke skills by typing `$` in the thread composer. See [Skills](https://developers.openai.com/codex/skills).37You can also explicitly invoke skills by typing `$` in the thread composer. See [Skills](https://developers.openai.com/codex/skills).

38 38 

39Enabled skills also appear in the slash command list (for example, `/imagegen`).39Enabled skills also appear in the slash command list.

40 40 

41### Available slash commands41### Available slash commands

42 42 

48| `/review` | Start code review mode to review uncommitted changes or compare against a base branch. |48| `/review` | Start code review mode to review uncommitted changes or compare against a base branch. |

49| `/status` | Show the thread ID, context usage, and rate limits. |49| `/status` | Show the thread ID, context usage, and rate limits. |

50 50 

51## Deeplinks

52 

53The Codex app registers the `codex://` URL scheme so links can open specific parts of the app directly.

54 

55| Deeplink | Opens | Supported query parameters |

56| --- | --- | --- |

57| `codex://settings` | Settings. | None. |

58| `codex://skills` | Skills. | None. |

59| `codex://automations` | Inbox in automation create mode. | None. |

60| `codex://threads/<thread-id>` | A local thread. `<thread-id>` must be a UUID. | None. |

61| `codex://new` | A new thread. | Optional: `prompt`, `originUrl`, `path`. |

62 

63For new-thread deeplinks:

64 

65- `prompt` sets the initial composer text.

66- `path` must be an absolute path to a local directory and, when valid, makes that directory the active workspace for the new thread.

67- `originUrl` tries to match one of your current workspace roots by Git remote URL. If both `path` and `originUrl` are present, Codex resolves `path` first.

68 

51## See also69## See also

52 70 

53- [Features](https://developers.openai.com/codex/app/features)71- [Features](https://developers.openai.com/codex/app/features)

1# Computer Use

2 

3In the Codex app, computer use is currently available on macOS, except in the

4 European Economic Area, the United Kingdom, and Switzerland at launch. Install

5 the Computer Use plugin, then grant Screen Recording and Accessibility

6 permissions when macOS prompts you.

7 

8With computer use, Codex can see and operate graphical user interfaces on macOS.

9Use it for tasks where command-line tools or structured integrations aren't

10enough, such as checking a desktop app, using a browser, changing app settings,

11working with a data source that isn't available as a plugin, or reproducing a

12bug that only happens in a graphical user interface.

13 

14Because computer use can affect app and system state outside your project

15workspace, use it for scoped tasks and review permission prompts before

16continuing.

17 

18## Set up computer use

19 

20In Codex settings, open **Computer Use** and click **Install** to install the

21Computer Use plugin before you ask Codex to operate desktop apps. When macOS

22prompts for access, grant Screen Recording and Accessibility permissions if you

23want Codex to see and interact with the target app.

24 

25To use computer use, grant:

26 

27- **Screen Recording** permission so Codex can see the target app.

28- **Accessibility** permission so Codex can click, type, and navigate.

29 

30## When to use computer use

31 

32Choose computer use when the task depends on a graphical user interface that's

33hard to verify through files or command output alone.

34 

35Good fits include:

36 

37- Testing a macOS app, an iOS simulator flow, or another desktop app that Codex

38 is building.

39- Performing a task that requires your web browser.

40- Reproducing a bug that only appears in a graphical interface.

41- Changing app settings that require clicking through a UI.

42- Inspecting information in an app or data source that isn't available through a

43 plugin.

44- Running a scoped task in the background while you keep working elsewhere.

45- Executing a workflow that spans more than one app.

46 

47For web apps you are building locally, use the

48[in-app browser](https://developers.openai.com/codex/app/browser) first.

49 

50## Start a computer use task

51 

52Mention `@Computer Use` or `@AppName` in your prompt, or ask Codex to use

53computer use. Describe the exact app, window, or flow Codex should operate.

54 

55```text

56Open the app with computer use, reproduce the onboarding bug, and fix the

57smallest code path that causes it. After each change, run the same UI flow

58again.

59```

60 

61```text

62Open @Chrome and verify the checkout page still works after the latest changes.

63```

64 

65If the target app exposes a dedicated plugin or MCP server, prefer that

66structured integration for data access and repeatable operations. Choose

67computer use when Codex needs to inspect or operate the app visually.

68 

69## Permissions and approvals

70 

71The macOS system permissions for computer use are separate from app approvals in

72Codex. The macOS permissions let Codex see and operate apps. App approvals

73determine which apps you allow Codex to use. File reads, file edits, and shell

74commands still follow the sandbox and approval settings for the thread.

75 

76With computer use, Codex can see and take action only in the apps you allow.

77During a task, Codex asks for your permission before it can use an app on your

78computer. You can choose **Always allow** so Codex can use that app in the future

79without asking again. You can remove apps from the **Always allow** list in the

80**Computer Use** section of Codex settings.

81 

82![Codex app asking for permission to use Calculator with computer use](/images/codex/app/computer-use-approval-light.webp)

83 

84Codex may also ask for permission before taking sensitive or disruptive actions.

85 

86If Codex can't see or control an app, open **System Settings > Privacy &

87Security** and check **Screen Recording** and **Accessibility** for the Codex

88app.

89 

90## Safety guidance

91 

92With computer use, Codex can view screen content, take screenshots, and interact

93with windows, menus, keyboard input, and clipboard state in the target app.

94Treat visible app content, browser pages, screenshots, and files opened in the

95target app as context Codex may process while the task runs.

96 

97Keep tasks narrow and stay present for sensitive flows:

98 

99- Give Codex one clear target app or flow at a time.

100- You can stop the task or take over your computer at any time.

101- Keep sensitive apps closed unless they're required for the task.

102- Avoid tasks that require secrets unless you're present and can approve each

103 step.

104- Review app permission prompts before allowing Codex to use an app.

105- Use **Always allow** only for apps you trust Codex to use automatically in

106 future tasks.

107- Stay present for account, security, privacy, network, payment, or

108 credential-related settings.

109- Cancel the task if Codex starts interacting with the wrong window.

110 

111If Codex uses your browser, it can interact with pages where you're already

112signed in. Review website actions as if you were taking them yourself: web pages

113can contain malicious or misleading content, and sites may treat approved clicks,

114form submissions, and signed-in actions as coming from your account. To keep

115using your browser while Codex works, ask Codex to use a different browser.

116 

117The feature can't automate terminal apps or Codex itself, since automating them

118could bypass Codex security policies. It also can't authenticate as an

119administrator or approve security and privacy permission prompts on your

120computer.

121 

122File edits and shell commands still follow Codex approval and sandbox settings

123where applicable. Changes made through desktop apps may not appear in the review

124pane until they're saved to disk and tracked by the project. Your ChatGPT data

125controls apply to content processed through Codex, including screenshots taken

126by computer use.

3The Codex app is a focused desktop experience for working on Codex threads in parallel,3The Codex app is a focused desktop experience for working on Codex threads in parallel,

4with built-in worktree support, automations, and Git functionality.4with built-in worktree support, automations, and Git functionality.

5 5 

6Most Codex app features are available on both macOS and Windows.

7Platform-specific exceptions are noted below.

8 

6---9---

7 10 

8## Multitask across projects11## Multitask across projects

14session in a specific directory.17session in a specific directory.

15 18 

16If you work in a single repository with two or more apps or packages, split19If you work in a single repository with two or more apps or packages, split

17distinct projects into separate app projects so the [sandbox](https://developers.openai.com/codex/security)20distinct projects into separate app projects so the [sandbox](https://developers.openai.com/codex/agent-approvals-security)

18only includes the files for that project.21only includes the files for that project.

19 22 

20![Codex app showing multiple projects in the sidebar and threads in the main pane](/images/codex/app/multitask-light.webp)23![Codex app showing multiple projects in the sidebar and threads in the main pane](/images/codex/app/multitask-light.webp)

31 34 

32You can also combine skills with [automations](https://developers.openai.com/codex/app/automations) to perform routine tasks35You can also combine skills with [automations](https://developers.openai.com/codex/app/automations) to perform routine tasks

33such as evaluating errors in your telemetry and submitting fixes or creating reports on recent36such as evaluating errors in your telemetry and submitting fixes or creating reports on recent

34codebase changes.37codebase changes. For ongoing work that should stay in one thread, use a

38[thread automation](https://developers.openai.com/codex/app/automations#thread-automations).

35 39 

36![Automation creation form with schedule and prompt fields](/images/codex/app/create-automation-light.webp)40![Automation creation form with schedule and prompt fields](/images/codex/app/create-automation-light.webp)

37 41 

85pressing <kbd>Cmd</kbd>+<kbd>J</kbd>.89pressing <kbd>Cmd</kbd>+<kbd>J</kbd>.

86 90 

87Use the terminal to validate changes, run scripts, and perform Git operations91Use the terminal to validate changes, run scripts, and perform Git operations

88without leaving the app.92without leaving the app. Codex can also read the current terminal output, so

93it can check the status of a running development server or refer back to a

94failed build while it works with you.

89 95 

90Common tasks include:96Common tasks include:

91 97 

128 134 

129![Pop-out window preview in light mode](/images/codex/app/popover-light.webp)135![Pop-out window preview in light mode](/images/codex/app/popover-light.webp)

130 136 

137## In-app browser

138 

139Use the [in-app browser](https://developers.openai.com/codex/app/browser) to preview, review, and comment on

140local development servers, file-backed previews, and public pages that don't

141require sign-in while you iterate on a web app.

142 

143The in-app browser doesn't support authentication flows, signed-in pages, your

144regular browser profile, cookies, extensions, or existing tabs.

145 

146Use browser comments to mark specific elements or areas on a page, then ask

147Codex to address that feedback.

148 

149![Codex app showing a browser comment on a local web app preview](/images/codex/app/in-app-browser-light.webp)

150 

151## Computer use

152 

153[Computer use](https://developers.openai.com/codex/app/computer-use) helps Codex operate a macOS app by

154seeing, clicking, and typing. This is useful for testing desktop apps, checking

155browser or simulator flows, working with data sources that aren't available as

156plugins, changing app settings, and reproducing GUI-only bugs.

157 

158Because computer use can affect app and system state outside your project

159workspace, keep tasks narrow and review permission prompts before continuing.

160 

161The feature isn't available in the European Economic Area, the United Kingdom, or

162Switzerland at launch.

163 

164![Codex app asking for permission to use Calculator with computer use](/images/codex/app/computer-use-approval-light.webp)

165 

166## Work with non-code artifacts