- Added a network sandbox proxy with policy enforcement to better control outbound network access. (#8442)
- Introduced the first phase of connectors support via the app server and MCP integration, including new config/docs updates. (#9667)
- Shipped collaboration mode as beta in the TUI, with a clearer plan → execute handoff and simplified mode selection (Coding vs Plan). (#9690, #9712, #9802, #9834)
- Added ephemeral threads and improved collaboration tool provenance metadata for spawned threads. (#9765, #9769)
- WebSocket connections now support proxy configuration. (#9719)
## Bug Fixes
- Fixed exec policy parsing for multiline quoted arguments. (#9565)
- `--yolo` now skips the git repository check instead of failing outside a repo. (#9590)
- Improved resume reliability by handling out-of-order events and prompting for the working directory when it differs. (#9512, #9731)
- Backspace no longer deletes a text element when the cursor is at the element’s left edge. (#9630)
- Config loading errors are clearer and more actionable across surfaces. (#9746)
- Default model selection now respects filtered presets to avoid invalid defaults. (#9782)
## Documentation
- Corrected a typo in the experimental collaboration prompt template. (#9716)
- Added documentation for the new connectors configuration surface. (#9667)
## Chores
- Refreshed the bundled model catalog/presets. (#9726)
- Updated GitHub Actions for Node 24 compatibility. (#9722)
## Changelog
Full Changelog: https://github.com/openai/codex/compare/rust-v0.89.0...rust-v0.90.0
- #9715 feat: fix formatting of `codex features list` @bolinfest
- #9716 Fix typo in experimental_prompt.md @iudizm
- #9719 feat: support proxy for ws connection @apanasenko-oai
- #9712 TUI: prompt to implement plan and switch to Execute @charley-oai
- #9713 use machine scope instead of user scope for dpapi. @iceweasel-oai
- #9726 Update models.json @github-actions
- #9667 [connectors] Support connectors part 1 - App server & MCP @mzeng-openai
- #9674 feat(app-server) Expose `personality` @dylan-hurd-oai
- #9733 Change the prompt for planning and reasoning effort @aibrahim-oai
- #9730 Hide mode cycle hint while a task is running @charley-oai
- #9720 feat: add session source as otel metadata tag @alexsong-oai
- #9565 Fix execpolicy parsing for multiline quoted args @jdsalasca
- #9745 chore: use some raw strings to reduce quoting @bolinfest
- #9753 nit: exclude PWD for rc sourcing @jif-oai
- #9690 feat: tui beta for collab @jif-oai
- #9116 Persist text element ranges and attached images across history/resume @charley-oai
- #9777 plan mode prompt change @aibrahim-oai
- #9590 fix(exec): skip git repo check when --yolo flag is used @zerone0x
- #9722 Upgrade GitHub Actions for Node 24 compatibility @salmanmkc
- #9611 Print warning if we skip config loading @gt-oai
- #9782 Select default model from filtered presets @aibrahim-oai
- #9512 Fix resume picker when user event appears after head @jdsalasca
- #9787 Remove stale TODO comment from defs.bzl @jcoens-openai
- #9700 still load skills @gt-oai
- #9791 Load untrusted rules @gt-oai
- #9707 bundle sandbox helper binaries in main zip, for winget. @iceweasel-oai
- #9792 Chore: remove mode from header @aibrahim-oai
- #9793 change collaboration mode to struct @aibrahim-oai
- #8442 feat: introducing a network sandbox proxy @viyatb-oai
- #9802 Have a coding mode and only show coding and plan @aibrahim-oai
- #9746 Another round of improvements for config error messages @etraut-openai
- #9812 Remove batman reference from experimental prompt @charley-oai
- #9769 feat: add thread spawn source for collab tools @jif-oai
- #9765 feat: ephemeral threads @jif-oai
- #9819 fix: libcc link @jif-oai
- #9820 fix: musl build @jif-oai
- #9316 fix(windows-sandbox): remove request files after read @MaxMiksa
- #9630 Prevent backspace from removing a text element when the cursor is at the element’s left edge @charley-oai
- #9840 Revert "fix: musl build" @aibrahim-oai
- #9778 Raise welcome animation breakpoint to 37 rows @mzeng-openai
- #9731 Ask for cwd choice when resuming session from different cwd @charley-oai
- #9841 Revert "fix: libcc link" @aibrahim-oai
- #9806 Use collaboration mode masks without mutating base settings @aibrahim-oai
- #9834 Mark collab as beta @pakrym-oai
- #9847 Revert "Revert "fix: musl build"" @jif-oai
- #9855 feat: cap number of agents @jif-oai
Keep an unmasked base collaboration mode and apply the active mask on
demand. Simplify the TUI mask helpers and update tests/docs to match the
mask contract.
# Summary
- Fix resume/fork config rebuild so cwd changes inside the TUI produce a
fully rebuilt Config (trust/approval/sandbox) instead of mutating only
the cwd.
- Preserve `--add-dir` behavior across resume/fork by normalizing
relative roots to absolute paths once (based on the original cwd).
- Prefer latest `TurnContext.cwd` for resume/fork prompts but fall back
to `SessionMeta.cwd` if the latest cwd no longer exists.
- Align resume/fork selection handling and ensure UI config matches the
resumed thread config.
- Fix Windows test TOML path escaping in trust-level test.
# Details
- Rebuild Config via `ConfigBuilder` when resuming into a different cwd;
carry forward runtime approval/sandbox overrides.
- Add `normalize_harness_overrides_for_cwd` to resolve relative
`additional_writable_roots` against the initial cwd before reuse.
- Guard `read_session_cwd` with filesystem existence check for the
latest `TurnContext.cwd`.
- Update naming/flow around cwd comparison and prompt selection.
<img width="603" height="150" alt="Screenshot 2026-01-23 at 5 42 13 PM"
src="https://github.com/user-attachments/assets/d1897386-bb28-4e8a-98cf-187fdebbecb0"
/>
And proof the model understands the new cwd:
<img width="828" height="353" alt="Screenshot 2026-01-22 at 5 36 45 PM"
src="https://github.com/user-attachments/assets/12aed8ca-dec3-4b64-8dae-c6b8cff78387"
/>
### Motivation
- The large ASCII welcome animation can push onboarding content below
the fold on default-height terminals, making the CLI appear
unresponsive; raising the breakpoint prevents that.
- The existing test measured an arbitrary row count rather than
asserting the welcome line position relative to the animation frame,
which made the intent unclear.
### Description
- Increase `MIN_ANIMATION_HEIGHT` from `20` to `37` in
`codex-rs/tui/src/onboarding/welcome.rs` so the animation is skipped
unless there is enough vertical space.
- Replace the brittle measurement logic in the welcome render test with
a `row_containing` helper and assert the welcome row equals the frame
height plus the spacer line (`frame_lines + 1`).
- Add a regression test
`welcome_skips_animation_below_height_breakpoint` that verifies the
animation is not rendered when the viewport height is one row below the
breakpoint.
### Testing
- Ran formatting with `~/.cargo/bin/just fmt` which completed
successfully.
- Ran unit tests for the crate with `cargo test -p codex-tui --lib` and
they passed (unit test suite succeeded).
- Ran `cargo test -p codex-tui` which reported a failing integration
test in this environment because the test cannot locate the `codex`
binary, so full crate tests are blocked here (environment limitation).
------
[Codex
Task](https://chatgpt.com/codex/tasks/task_i_6973b0a710d4832c9ff36fac26eb1519)
**Summary**
- Prevent backspace from removing a text element when the cursor is at
the element’s left edge.
- Instead just delete the char before the placeholder (moving it to the
left).
## Summary
- Remove elevated runner request files after read (best-effort cleanup
on errors)
- Add a unit test to cover request file lifecycle
## Testing
- `cargo test -p codex-windows-sandbox` (Windows)
Fixes#9315
In a [recent PR](https://github.com/openai/codex/pull/9182), I made some
improvements to config error messages so errors didn't leave app server
clients in a dead state. This is a follow-on PR to make these error
messages more readable and actionable for both TUI and GUI users. For
example, see #9668 where the user was understandably confused about the
source of the problem and how to fix it.
The improved error message:
1. Clearly identifies the config file where the error was found (which
is more important now that we support layered configs)
2. Provides a line and column number of the error
3. Displays the line where the error occurred and underlines it
For example, if my `config.toml` includes the following:
```toml
[features]
collaboration_modes = "true"
```
Here's the current CLI error message:
```
Error loading config.toml: invalid type: string "true", expected a boolean in `features`
```
And here's the improved message:
```
Error loading config.toml:
/Users/etraut/.codex/config.toml:43:23: invalid type: string "true", expected a boolean
|
43 | collaboration_modes = "true"
| ^^^^^^
```
The bulk of the new logic is contained within a new module
`config_loader/diagnostics.rs` that is responsible for calculating the
text range for a given toml path (which is more involved than I would
have expected).
In addition, this PR adds the file name and text range to the
`ConfigWarningNotification` app server struct. This allows GUI clients
to present the user with a better error message and an optional link to
open the errant config file. This was a suggestion from @.bolinfest when
he reviewed my previous PR.
This add a new crate, `codex-network-proxy`, a local network proxy
service used by Codex to enforce fine-grained network policy (domain
allow/deny) and to surface blocked network events for interactive
approvals.
- New crate: `codex-rs/network-proxy/` (`codex-network-proxy` binary +
library)
- Core capabilities:
- HTTP proxy support (including CONNECT tunneling)
- SOCKS5 proxy support (in the later PR)
- policy evaluation (allowed/denied domain lists; denylist wins;
wildcard support)
- small admin API for polling/reload/mode changes
- optional MITM support for HTTPS CONNECT to enforce “limited mode”
method restrictions (later PR)
Will follow up integration with codex in subsequent PRs.
## Testing
- `cd codex-rs && cargo build -p codex-network-proxy`
- `cd codex-rs && cargo run -p codex-network-proxy -- proxy`
### Motivation
- Remove an outdated comment in `defs.bzl` referencing
`cargo_build_script` that is no longer relevant.
### Description
- Delete the stale `# TODO(zbarsky): cargo_build_script support?` line
so the logic flows directly from `binaries` to `lib_srcs` in `defs.bzl`.
### Testing
- Ran `git diff --check` which produced no errors.
------
[Codex
Task](https://chatgpt.com/codex/tasks/task_i_6973d9ac757c8331be475a8fb0f90a88)
Fixes#9501
Contributing guide:
https://github.com/openai/codex/blob/main/docs/contributing.md
## Summary
The resume picker requires a session_meta line and at least one
user_message event within the initial head scan. Some rollout files
contain multiple session_meta entries before the first user_message, so
the user event can fall outside the default head window and the session
is omitted from the picker even though it is resumable by ID.
This PR keeps the head summary bounded but extends scanning for a
user_message once a session_meta has been observed. The summary still
caps stored head entries, but we allow a small, bounded extra scan to
find the first user event so valid sessions are not filtered out.
## Changes
- Continue scanning past the head limit (bounded) when session_meta is
present but no user_message has been seen yet.
- Mark session_meta as seen even if the head summary buffer is already
full.
- Add a regression test with multiple session_meta lines before the
first user_message.
## Why This Is Safe
- The head summary remains bounded to avoid unbounded memory usage.
- The extra scan is capped (USER_EVENT_SCAN_LIMIT) and only triggers
after a session_meta is seen.
- Behavior is unchanged for typical files where the user_message appears
early.
## Testing
- cargo test -p codex-core --lib
test_list_threads_scans_past_head_for_user_event
## Summary
Upgrade GitHub Actions to their latest versions to ensure compatibility
with Node 24, as Node 20 will reach end-of-life in April 2026.
## Changes
| Action | Old Version(s) | New Version | Release | Files |
|--------|---------------|-------------|---------|-------|
| `actions/cache` |
[`v4`](https://github.com/actions/cache/releases/tag/v4) |
[`v5`](https://github.com/actions/cache/releases/tag/v5) |
[Release](https://github.com/actions/cache/releases/tag/v5) | bazel.yml
|
## Context
Per [GitHub's
announcement](https://github.blog/changelog/2025-09-19-deprecation-of-node-20-on-github-actions-runners/),
Node 20 is being deprecated and runners will begin using Node 24 by
default starting March 4th, 2026.
### Why this matters
- **Node 20 EOL**: April 2026
- **Node 24 default**: March 4th, 2026
- **Action**: Update to latest action versions that support Node 24
### Security Note
Actions that were previously pinned to commit SHAs remain pinned to SHAs
(updated to the latest release SHA) to maintain the security benefits of
immutable references.
### Testing
These changes only affect CI/CD workflow configurations and should not
impact application functionality. The workflows should be tested by
running them on a branch before merging.
Signed-off-by: Salman Muin Kayser Chishti <13schishti@gmail.com>
## Summary
Fixes#7522
The `--yolo` (`--dangerously-bypass-approvals-and-sandbox`) flag is
documented to skip all confirmation prompts and execute commands without
sandboxing, intended solely for running in environments that are
externally sandboxed. However, it was not bypassing the trusted
directory (git repo) check, requiring users to also specify
`--skip-git-repo-check`.
This change makes `--yolo` also skip the git repo check, matching the
documented behavior and user expectations.
## Changes
- Modified `codex-rs/exec/src/lib.rs` to check for
`dangerously_bypass_approvals_and_sandbox` flag in addition to
`skip_git_repo_check` when determining whether to skip the git repo
check
## Testing
- Verified the code compiles with `cargo check -p codex-exec`
- Ran existing tests with `cargo test -p codex-exec` (34 passed, 8
integration tests failed due to unrelated API connectivity issues)
---
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-authored-by: Claude <noreply@anthropic.com>
# 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**
- Backtrack selection now rehydrates `text_elements` and
`local_image_paths` from the chosen user history cell so Esc‑Esc history
edits preserve image placeholders and attachments.
- Composer prefill uses the preserved elements/attachments in both `tui`
and `tui2`.
- Extended backtrack selection tests to cover image placeholder elements
and local image paths.
**Changes**
- `tui/src/app_backtrack.rs`: Backtrack selection now carries text
elements + local image paths; composer prefill uses them (removes TODO).
- `tui2/src/app_backtrack.rs`: Same as above.
- `tui/src/app.rs`: Updated backtrack test to assert restored
elements/paths.
- `tui2/src/app.rs`: Same test updates.
### The original scope of this PR (threading text elements and image
attachments through the codex harness thoroughly/persistently) was
broken into the following PRs other than this one:
The diff of this PR was reduced by changing types in a starter PR:
https://github.com/openai/codex/pull/9235
Then text element metadata was added to protocol, app server, and core
in this PR: https://github.com/openai/codex/pull/9331
Then the end-to-end flow was completed by wiring TUI/TUI2 input,
history, and restore behavior in
https://github.com/openai/codex/pull/9393
Prompt expansion was supported in this PR:
https://github.com/openai/codex/pull/9518
TextElement optional placeholder field was protected in
https://github.com/openai/codex/pull/9545
## What
Fix bash command parsing to accept double-quoted strings that contain
literal newlines so execpolicy can match allow rules.
## Why
Allow rules like [git, commit] should still match when commit messages
include a newline in a quoted argument; the parser currently rejects
these strings and falls back to the outer shell invocation.
## How
- Validate double-quoted strings by ensuring all named children are
string_content and then stripping the outer quotes from the raw node
text so embedded newlines are preserved.
- Reuse the helper for concatenated arguments.
- Ensure large SI suffix formatting uses the caller-provided locale
formatter for grouping.
- Add coverage for newline-containing quoted arguments.
Fixes#9541.
## Tests
- cargo test -p codex-core
- just fix -p codex-core
- cargo test -p codex-protocol
- just fix -p codex-protocol
- cargo test --all-features
## Summary
- hide the “(shift+tab to cycle)” suffix on the collaboration mode label
while a task is running
- keep the cycle hint visible when idle
- add a snapshot to cover the running-task label state
### Motivation
Exposes a per-thread / per-turn `personality` override in the v2
app-server API so clients can influence model communication style at
thread/turn start. Ensures the override is passed into the session
configuration resolution so it becomes effective for subsequent turns
and headless runners.
### Testing
- [x] Add an integration-style test
`turn_start_accepts_personality_override_v2` in
`codex-rs/app-server/tests/suite/v2/turn_start.rs` that verifies a
`/personality` override results in a developer update message containing
`<personality_spec>` in the outbound model request.
------
[Codex
Task](https://chatgpt.com/codex/tasks/task_i_6971d646b1c08322a689a54d2649f3fe)
In order to make Codex work with connectors, we add a built-in gateway
MCP that acts as a transparent proxy between the client and the
connectors. The gateway MCP collects actions that are accessible to the
user and sends them down to the user, when a connector action is chosen
to be called, the client invokes the action through the gateway MCP as
well.
- [x] Add the system built-in gateway MCP to list and run connectors.
- [x] Add the app server methods and protocol
This fixes a bug where the elevated sandbox setup encrypts sandbox user
passwords as an admin user, but normal command execution attempts to
decrypt them as a different user.
Machine scope allows all users to encyrpt/decrypt
this PR also moves the encrypted file to a different location
.codex/.sandbox-secrets which the sandbox users cannot read.
## Summary
- Replace the plan‑implementation prompt with a standard selection
popup.
- “Yes” submits a user turn in Execute via a dedicated app event to
preserve normal transcript behavior.
- “No” simply dismisses the popup.
<img width="977" height="433" alt="Screenshot 2026-01-22 at 2 00 54 PM"
src="https://github.com/user-attachments/assets/91fad06f-7b7a-4cd8-9051-f28a19b750b2"
/>
## Changes
- Add a plan‑implementation popup using `SelectionViewParams`.
- Add `SubmitUserMessageWithMode` so “Yes” routes through
`submit_user_message` (ensures user history + separator state).
- Track `saw_plan_update_this_turn` so the prompt appears even when only
`update_plan` is emitted.
- Suppress the plan popup on replayed turns, when messages are queued,
or when a rate‑limit prompt is pending.
- Add `execute_mode` helper for collaboration modes.
- Add tests for replay/queued/rate‑limit guards and plan update without
final message.
- Add snapshots for both the default and “No”‑selected popup states.
The formatting of `codex features list` made it hard to follow. This PR
introduces column width math to make things nice.
Maybe slightly hard to machine-parse (since not a simple `\t`), but we
should introduce a `--json` option if that's really important.
You can see the before/after in the screenshot:
<img width="1119" height="932" alt="image"
src="https://github.com/user-attachments/assets/c99dce85-899a-4a2d-b4af-003938f5e1df"
/>
## Summary
Support updating Personality mid-Thread via UserTurn/OverwriteTurn. This
is explicitly unused by the clients so far, to simplify PRs - app-server
and tui implementations will be follow-ups.
## Testing
- [x] added integration tests
