- API v2 threads can now inject dynamic tools at startup and route their calls/responses end-to-end through the server and core tool pipeline. (#9539)
- Added filtering on the thread list in the app server to make large thread sets easier to browse. (#9897)
- Introduced a `thread/unarchive` RPC to restore archived rollouts back into active sessions. (#9843)
- MCP servers can now define OAuth scopes in `config.toml`, reducing the need to pass `--scopes` on each login. (#9647)
- Multi-agent collaboration is more capable and safer, with an explorer role, better collab event mapping, and max-depth guardrails. (#9817, #9818, #9918, #9899)
- Cached `web_search` is now the default client behavior. (#9974)
## Bug Fixes
- Fixed a TUI deadlock/freeze under high streaming throughput by avoiding blocking sends on the main Tokio thread. (#9951)
- The `web_search` tool now handles and displays all action types, and shows in-progress activity instead of appearing stuck. (#9960)
- Reduced high CPU usage in collaboration flows by eliminating busy-waiting on subagents. (#9776)
- Fixed `codex resume --last --json` so prompts parse correctly without conflicting argument errors. (#9475)
- Windows sandbox logging now handles UTF-8 safely, preventing failures when truncating multibyte content. (#8647)
- `request_user_input` is now rejected outside Plan/Pair modes to prevent invalid tool calls. (#9955)
## Documentation
- Updated the contribution guidelines for clearer onboarding and workflow expectations. (#9933)
- Refreshed protocol/MCP docs to reflect `thread/unarchive` and the updated `request_user_input` question shape. (#9843, #9890)
## Chores
- Self-update via Homebrew now uses an explicit cask upgrade command to avoid warnings and ambiguity. (#9823)
- Release packaging now consistently writes the bundle zip to `dist/`. (#9934)
- Updated key dependencies in the Rust workspace (including `axum`, `tracing`, `globset`, and `tokio-test`). (#9880, #9882, #9883, #9884)
- Aligned feature stage naming with public maturity stages and added clearer warnings for underdevelopment features. (#9929, #9954)
## Changelog
Full Changelog: https://github.com/openai/codex/compare/rust-v0.91.0...rust-v0.92.0
- #9850 chore: remove extra newline in println @SohailRaoufi
- #9868 Adjust modes masks @aibrahim-oai
- #9874 Prompt @aibrahim-oai
- #9877 Plan prompt @aibrahim-oai
- #9718 feat(tui) /personality @dylan-hurd-oai
- #9871 chore(core) move model_instructions_template config @dylan-hurd-oai
- #9539 feat: dynamic tools injection @jif-oai
- #9818 feat: rebase multi-agent tui on `config_snapshot` @jif-oai
- #9789 Fix flakey resume test @gt-oai
- #9784 Fix flakey conversation flow test @gt-oai
- #9918 feat: explorer collab @jif-oai
- #9899 feat: disable collab at max depth @jif-oai
- #9916 Fix up config disabled err msg @gt-oai
- #9890 Feat: add isOther to question returned by request user input tool @shijie-oai
- #9817 feat: codex exec mapping of collab tools @jif-oai
- #9919 Fix flakey shell snapshot test @gt-oai
- #9776 fix: attempt to reduce high cpu usage when using collab @eugeneoden
- #9928 prompt @aibrahim-oai
- #9925 chore: update interrupt message @jif-oai
- #9843 Add thread/unarchive to restore archived rollouts @charley-oai
- #9929 Aligned feature stage names with public feature maturity stages @etraut-openai
- #9934 ensure codex bundle zip is created in dist/ @iceweasel-oai
- #9897 [app-server] feat: add filtering on thread list @jif-oai
- #9647 Add MCP server `scopes` config and use it as fallback for OAuth login @blevy-oai
- #9943 plan prompt @aibrahim-oai
- #8647 fix: handle utf-8 in windows sandbox logs @davidgilbertson
- #9891 Add composer config and shared menu surface helpers @aibrahim-oai
- #9880 chore(deps): bump tracing from 0.1.43 to 0.1.44 in /codex-rs @dependabot
- #9882 chore(deps): bump tokio-test from 0.4.4 to 0.4.5 in /codex-rs @dependabot
- #9883 chore(deps): bump axum from 0.8.4 to 0.8.8 in /codex-rs @dependabot
- #9884 chore(deps): bump globset from 0.4.16 to 0.4.18 in /codex-rs @dependabot
- #9901 fix: remove cli tooltip references to custom prompts @mattridley
- #9823 fix: use `brew upgrade --cask codex` to avoid warnings and ambiguity @JBallin
- #9951 fix: try to fix freezes 2 @jif-oai
- #9955 Reject request_user_input outside Plan/Pair @charley-oai
- #9933 Updated contribution guidelines @etraut-openai
- #9892 Reuse ChatComposer in request_user_input overlay @aibrahim-oai
- #9957 NIT larger buffer @jif-oai
- #9953 feat: load interface metadata from SKILL.json @alexsong-oai
- #9954 Warn users on enabling underdevelopment features @aibrahim-oai
- #9961 Use test_codex more @pakrym-oai
- #9960 fix: handle all web_search actions and in progress invocations @sayan-oai
- #9966 plan prompt v7 @aibrahim-oai
- #9968 Improve plan mode prompt @aibrahim-oai
- #9969 prompt final @aibrahim-oai
- #9475 Fix `resume --last` with `--json` option @etraut-openai
- #9970 prompt @aibrahim-oai
- #9975 plan prompt @aibrahim-oai
- #9974 make cached web_search client-side default @sayan-oai
- #9971 tui: wrapping user input questions @aibrahim-oai
- #9977 make plan prompt less detailed @aibrahim-oai
- #9980 Fixing main and make plan mode reasoning effort medium @aibrahim-oai
- #9759 Fix: cap aggregated exec output consistently @Kbediako
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
Fix resume --last prompt parsing by dropping the clap conflict on the
codex resume subcommand so a positional prompt is accepted when --last
is set. This aligns interactive resume behavior with exec-mode logic and
avoids the “--last cannot be used with SESSION_ID” error.
This addresses #6717
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
### Summary
- Parse all `web_search` tool actions (`search`, `find_in_page`,
`open_page`).
- Previously we only parsed + displayed `search`, which made the TUI
appear to pause when the other actions were being used.
- Show in progress `web_search` calls as `Searching the web`
- Previously we only showed completed tool calls
<img width="308" height="149" alt="image"
src="https://github.com/user-attachments/assets/90a4e8ff-b06a-48ff-a282-b57b31121845"
/>
### Tests
Added + updated tests, tested locally
### Follow ups
Update VSCode extension to display these as well
Reuse the shared chat composer for notes and freeform answers in
request_user_input.
- Build the overlay composer with ChatComposerConfig::plain_text.
- Wire paste-burst flushing + menu surface sizing through the bottom
pane.
## Context
Previous work in https://github.com/openai/codex/pull/9560 only rejected
`request_user_input` in Execute and Custom modes. Since then, additional
modes
(e.g., Code) were added, so the guard should be mode-agnostic.
## What changed
- Switch the handler to an allowlist: only Plan and PairProgramming are
allowed
- Return the same error for any other mode (including Code)
- Add a Code-mode rejection test alongside the existing Execute/Custom
tests
## Why
This prevents `request_user_input` from being used in modes where it is
not
intended, even as new modes are introduced.
Fixes a TUI freeze caused by awaiting `mpsc::Sender::send()` that blocks
the tokio thread, stopping the consumption runtime and creating a
deadlock. This could happen if the server was producing enough chunks to
fill the `mpsc` fast enough. To solve this we try on insert using a
`try_send()` (not requiring an `await`) and delegate to a tokio task if
this does not work
This is a temporary solution as it can contain races for delta elements
and a stronger design should come here
Fixes#9822
### Summary
Make the Homebrew upgrade command explicit by using `brew upgrade --cask
codex`.
### Motivation
During the Codex self-update, Homebrew can emit an avoidable warning
because the
name `codex` resolves to a cask:
```
Warning: Formula codex was renamed to homebrew/cask/codex.
````
While the upgrade succeeds, this relies on implicit name resolution and
produces
unnecessary output during the update flow.
### Why `--cask`
* Eliminates warning/noise for users
* Explicitly matches how Codex is distributed via Homebrew
* Avoids reliance on name resolution behavior
* Makes the command more robust if a `codex` formula is ever introduced
### Context
This restores the `--cask` flag that was removed in #6238 after being
considered
“not necessary” during review:
[https://github.com/openai/codex/pull/6238#discussion_r2505947880](https://github.com/openai/codex/pull/6238#discussion_r2505947880).
Co-authored-by: Eric Traut <etraut@openai.com>
Bumps [globset](https://github.com/BurntSushi/ripgrep) from 0.4.16 to
0.4.18.
<details>
<summary>Commits</summary>
<ul>
<li><a
href="0b0e013f5a"><code>0b0e013</code></a>
globset-0.4.18</li>
<li><a
href="cac9870a02"><code>cac9870</code></a>
doc: update date in man page template</li>
<li><a
href="24e88dc15b"><code>24e88dc</code></a>
ignore/types: add <code>ssa</code> type</li>
<li><a
href="5748f81bb1"><code>5748f81</code></a>
printer: use <code>doc_cfg</code> instead of
<code>doc_auto_cfg</code></li>
<li><a
href="d47663b1b4"><code>d47663b</code></a>
searcher: fix regression with <code>--line-buffered</code> flag</li>
<li><a
href="38d630261a"><code>38d6302</code></a>
printer: add Cursor hyperlink alias</li>
<li><a
href="b3dc4b0998"><code>b3dc4b0</code></a>
globset: improve debug log</li>
<li><a
href="ca2e34f37c"><code>ca2e34f</code></a>
grep-0.4.0</li>
<li><a
href="a0d61a063f"><code>a0d61a0</code></a>
grep-printer-0.3.0</li>
<li><a
href="c22fc0f13c"><code>c22fc0f</code></a>
deps: bump to grep-searcher 0.1.15</li>
<li>Additional commits viewable in <a
href="https://github.com/BurntSushi/ripgrep/compare/globset-0.4.16...globset-0.4.18">compare
view</a></li>
</ul>
</details>
<br />
[](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)
Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.
[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)
---
<details>
<summary>Dependabot commands and options</summary>
<br />
You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot merge` will merge this PR after your CI passes on it
- `@dependabot squash and merge` will squash and merge this PR after
your CI passes on it
- `@dependabot cancel merge` will cancel a previously requested merge
and block automerging
- `@dependabot reopen` will reopen this PR if it is closed
- `@dependabot close` will close this PR and stop Dependabot recreating
it. You can achieve the same result by closing it manually
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)
</details>
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Bumps [tokio-test](https://github.com/tokio-rs/tokio) from 0.4.4 to
0.4.5.
<details>
<summary>Commits</summary>
<ul>
<li><a
href="41d1877689"><code>41d1877</code></a>
chore: prepare tokio-test 0.4.5 (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7831">#7831</a>)</li>
<li><a
href="60b083b630"><code>60b083b</code></a>
chore: prepare tokio-stream 0.1.18 (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7830">#7830</a>)</li>
<li><a
href="9cc02cc88d"><code>9cc02cc</code></a>
chore: prepare tokio-util 0.7.18 (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7829">#7829</a>)</li>
<li><a
href="d2799d791b"><code>d2799d7</code></a>
task: improve the docs of <code>Builder::spawn_local</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7828">#7828</a>)</li>
<li><a
href="4d4870f291"><code>4d4870f</code></a>
task: doc that task drops before JoinHandle completion (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7825">#7825</a>)</li>
<li><a
href="fdb150901a"><code>fdb1509</code></a>
fs: check for io-uring opcode support (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7815">#7815</a>)</li>
<li><a
href="426a562780"><code>426a562</code></a>
rt: remove <code>allow(dead_code)</code> after <code>JoinSet</code>
stabilization (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7826">#7826</a>)</li>
<li><a
href="e3b89bbefa"><code>e3b89bb</code></a>
chore: prepare Tokio v1.49.0 (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7824">#7824</a>)</li>
<li><a
href="4f577b84e9"><code>4f577b8</code></a>
Merge 'tokio-1.47.3' into 'master'</li>
<li><a
href="f320197693"><code>f320197</code></a>
chore: prepare Tokio v1.47.3 (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7823">#7823</a>)</li>
<li>Additional commits viewable in <a
href="https://github.com/tokio-rs/tokio/compare/tokio-test-0.4.4...tokio-test-0.4.5">compare
view</a></li>
</ul>
</details>
<br />
[](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)
Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.
[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)
---
<details>
<summary>Dependabot commands and options</summary>
<br />
You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot merge` will merge this PR after your CI passes on it
- `@dependabot squash and merge` will squash and merge this PR after
your CI passes on it
- `@dependabot cancel merge` will cancel a previously requested merge
and block automerging
- `@dependabot reopen` will reopen this PR if it is closed
- `@dependabot close` will close this PR and stop Dependabot recreating
it. You can achieve the same result by closing it manually
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)
</details>
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Currently `apply_patch` will fail on Windows if the file contents happen
to have a multi-byte character at the point where the `preview` function
truncates.
I've used the existing `take_bytes_at_char_boundary` helper and added a
regression test (that fails without the fix).
This is related to #4013 but doesn't fix it.
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
### Motivation
- Allow MCP OAuth flows to request scopes defined in `config.toml`
instead of requiring users to always pass `--scopes` on the CLI.
CLI/remote parameters should still override config values.
### Description
- Add optional `scopes: Option<Vec<String>>` to `McpServerConfig` and
`RawMcpServerConfig`, and propagate it through deserialization and the
built config types.
- Serialize `scopes` into the MCP server TOML via
`serialize_mcp_server_table` in `core/src/config/edit.rs` and include
`scopes` in the generated config schema (`core/config.schema.json`).
- CLI: update `codex-rs/cli/src/mcp_cmd.rs` `run_login` to fall back to
`server.scopes` when the `--scopes` flag is empty, with explicit CLI
scopes still taking precedence.
- App server: update
`codex-rs/app-server/src/codex_message_processor.rs`
`mcp_server_oauth_login` to use `params.scopes.or_else(||
server.scopes.clone())` so the RPC path also respects configured scopes.
- Update many test fixtures to initialize the new `scopes` field (set to
`None`) so test code builds with the new struct field.
### Testing
- Ran config tooling and formatters: `just write-config-schema`
(succeeded), `just fmt` (succeeded), and `just fix -p codex-core`, `just
fix -p codex-cli`, `just fix -p codex-app-server` (succeeded where
applicable).
- Ran unit tests for the CLI: `cargo test -p codex-cli` (passed).
- Ran unit tests for core: `cargo test -p codex-core` (ran; many tests
passed but several failed, including model refresh/403-related tests,
shell snapshot/timeouts, and several `unified_exec` expectations).
- Ran app-server tests: `cargo test -p codex-app-server` (ran; many
integration-suite tests failed due to mocked/remote HTTP 401/403
responses and wiremock expectations).
If you want, I can split the tests into smaller focused runs or help
debug the failing integration tests (they appear to be unrelated to the
config change and stem from external HTTP/mocking behaviors encountered
during the test runs).
------
[Codex
Task](https://chatgpt.com/codex/tasks/task_i_69718f505914832ea1f334b3ba064553)
We've recently standardized a [feature maturity
model](https://developers.openai.com/codex/feature-maturity) that we're
using in our docs and support forums to communicate expectations to
users. This PR updates the internal stage names and descriptions to
match.
This change involves a simple internal rename and updates to a few
user-visible strings. No functional change.
## Summary
- Adds a new `thread/unarchive` RPC to move archived thread rollouts
back into the active `sessions/` tree.
## What changed
- **Protocol**
- Adds `thread/unarchive` request/response types and wiring.
- **Server**
- Implements `thread_unarchive` in the app server.
- Validates the archived rollout path and thread ID.
- Restores the rollout to `sessions/YYYY/MM/DD/...` based on the rollout
filename timestamp.
- **Core**
- Adds `find_archived_thread_path_by_id_str` helper for archived
rollouts.
- **Docs**
- Documents the new RPC and usage example.
- **Tests**
- Adds an end-to-end server test that:
1) starts a thread,
2) archives it,
3) unarchives it,
4) asserts the file is restored to `sessions/`.
## How to use
```json
{ "method": "thread/unarchive", "id": 24, "params": { "threadId": "<thread-id>" } }
```
## Author Codex Session
`codex resume 019bf158-54b6-7960-a696-9d85df7e1bc1` (soon I'll make this
kind of session UUID forkable by anyone with the right
`session_object_storage_url` line in their config, but for now just
pasting it here for my reference)
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
Reproduce with a prompt like this with collab enabled:
```
Examine the code at <some subdirectory with a deeply nested project>. Find the most urgent issue to resolve and describe it to me.
```
Existing behavior causes the top-level agent to busy wait on subagents.
### Summary
Add `isOther` to question object from request_user_input tool input and
remove `other` option from the tool prompt to better handle tool input.
I've seen this test fail with:
```
- Mock #1.
Expected range of matching incoming requests: == 2
Number of matched incoming requests: 1
```
This is because we pop the wrong task_complete events and then the test
exits. I think this is because the MCP events are now buffered after
https://github.com/openai/codex/pull/8874.
So:
1. clear the buffer before we do any user message sending
2. additionally listen for task start before task complete
3. use the ID from task start to find the correct task complete event.
Sessions' `updated_at` times are truncated to seconds, with the UUID
session ID used to break ties. If the two test sessions are created in
the same second, AND the session B UUID < session A UUID, the test
fails.
Fix this by mutating the session mtimes, from which we derive the
updated_at time, to ensure session B is updated_at later than session A.
## Summary
Add dynamic tool injection to thread startup in API v2, wire dynamic
tool calls through the app server to clients, and plumb responses back
into the model tool pipeline.
### Flow (high level)
- Thread start injects `dynamic_tools` into the model tool list for that
thread (validation is done here).
- When the model emits a tool call for one of those names, core raises a
`DynamicToolCallRequest` event.
- The app server forwards it to the client as `item/tool/call`, waits
for the client’s response, then submits a `DynamicToolResponse` back to
core.
- Core turns that into a `function_call_output` in the next model
request so the model can continue.
### What changed
- Added dynamic tool specs to v2 thread start params and protocol types;
introduced `item/tool/call` (request/response) for dynamic tool
execution.
- Core now registers dynamic tool specs at request time and routes those
calls via a new dynamic tool handler.
- App server validates tool names/schemas, forwards dynamic tool call
requests to clients, and publishes tool outputs back into the session.
- Integration tests
## Summary
Move `model_instructions_template` config to the experimental slug while
we iterate on this feature
## Testing
- [x] Tested locally, unit tests still pass
## Summary
Adds /personality selector in the TUI, which leverages the new core
interface in #9644
Notes:
- We are doing some of our own state management for model_info loading
here, but not sure if that's ideal. open to opinions on simpler
approach, but would like to avoid blocking on a larger refactor
- Right now, the `/personality` selector just hides when the model
doesn't support it. we can update this behavior down the line
## Testing
- [x] Tested locally
- [x] Added snapshot tests
