Compare commits

...

154 Commits

Author SHA1 Message Date
Dylan Hurd
3f75c61e62 chore(ci) add .bazelversion 2026-01-20 22:55:14 -08:00
Ahmed Ibrahim
ebc88f29f8 don't ask for approval for just fix (#9586)
It blocks all my skills from executing because it asks to run just fmt.
It's quick command that doesn't need approval.


<img width="967" height="120" alt="image"
src="https://github.com/user-attachments/assets/f8e6ca76-a650-49e9-beb2-ce98ba48d310"
/>
2026-01-21 04:56:11 +00:00
Ahmed Ibrahim
465da00d02 fix CI by running pnpm (#9587) 2026-01-20 20:54:15 -08:00
pakrym-oai
527b7b4c02 Feature to auto-enable websockets transport (#9578) 2026-01-20 20:32:06 -08:00
alexsong-oai
fabc2bcc32 feat: add skill injected counter metric (#9575) 2026-01-20 19:05:37 -08:00
charley-oai
0523a259c8 Reject ask user question tool in Execute and Custom (#9560)
## Summary
- Keep `request_user_input` in the tool list but reject it at runtime in
Execute/Custom modes with a clear model-facing error.
- Add a session accessor for current collaboration mode and enforce the
gate in the request_user_input handler.
- Update core/app-server tests to use Plan mode for success and add
Execute/Custom rejection coverage.
2026-01-20 18:32:17 -08:00
charley-oai
531748a080 Prompt Expansion: Preserve Text Elements (#9518)
Summary
- Preserve `text_elements` through custom prompt argument parsing and
expansion (named and numeric placeholders).
- Translate text element ranges through Shlex parsing using sentinel
substitution, and rehydrate text + element ranges per arg.
- Drop image attachments when their placeholder does not survive prompt
expansion, keeping attachments consistent with rendered elements.
- Mirror changes in TUI2 and expand tests for prompt parsing/expansion
edge cases.

Tests
- placeholders with spaces as single tokens (positional + key=value,
quoted + unquoted),
  - prompt expansion with image placeholders,
  - large paste + image arg combinations,
  - unused image arg dropped after expansion.
2026-01-20 18:30:20 -08:00
Michael Bolin
f4d55319d1 feat: rename experimental_instructions_file to model_instructions_file (#9555)
A user who has `experimental_instructions_file` set will now see this:

<img width="888" height="660" alt="image"
src="https://github.com/user-attachments/assets/51c98312-eb9b-4881-81f1-bea6677e158d"
/>

And a `codex exec` would include this warning:

<img width="888" height="660" alt="image"
src="https://github.com/user-attachments/assets/a89f62be-1edf-4593-a75e-e0b4a762ed7d"
/>
2026-01-21 02:25:08 +00:00
Ahmed Ibrahim
3a0eeb8edf Show session header before configuration (#9568)
We were skipping if we know the model. We shouldn't
2026-01-21 02:13:54 +00:00
Michael Bolin
ac2090caf2 fix: bminor/bash is no longer on GitHub so use bolinfest/bash instead (#9563)
This should fix CI.
2026-01-21 00:35:42 +00:00
Josh McKinney
0a26675155 feat(tui2): add /experimental menu (#9562)
Adds an /experimental slash command and bottom-pane view to toggle beta
features.

Persists feature-flag updates to config.toml, matching tui behavior.
2026-01-21 00:20:57 +00:00
Jeff Mickey
c14e6813fb [codex-tui] exit when terminal is dumb (#9293)
Using terminal with TERM=dumb specifically mean that TUIs and the like
don't work. Ensure that codex doesn't run in these environments and exit
with odd errors like crossterm's "Error: The cursor position could not
be read within a normal duration"

---------

Co-authored-by: Josh McKinney <joshka@openai.com>
2026-01-20 16:17:38 -08:00
HDCode
80f80181c2 fix(core): require approval for force delete on Windows (#8590)
### What
Implemented detection for dangerous "force delete" commands on Windows
to trigger the user approval prompt when `--ask-for-approval on-request`
is set. This aligns Windows behavior with the existing safety checks for
`rm -rf` on Linux.

### Why
Fixes #8567 - a critical safety gap where destructive Windows commands
could bypass the approval prompt. This prevents accidental data loss by
ensuring the user explicitly confirms operations that would otherwise
suppress the OS's native confirmation prompts.

### How
Updated the Windows command safety module to identify and flag the
following patterns as dangerous:
*   **PowerShell**:
* Detects `Remove-Item` (and aliases `rm`, `ri`, `del`, `erase`, `rd`,
`rmdir`) when used with the `-Force` flag.
* Uses token-based analysis to robustly detect these patterns even
inside script blocks (`{...}`), sub-expression `(...)`, or
semicolon-chained sequences.
*   **CMD**:
    *   Detects `del /f` (force delete files).
    *   Detects `rd /s /q` (recursive delete quiet).
* **Command Chaining**: Added support for analyzing chained commands
(using `&`, `&&`, `|`, `||`) to separate and check individual commands
(e.g., catching `del /f` hidden in `echo log & del /f data`).

### Testing
Added comprehensive unit tests covering:
* **PowerShell**: `Remove-Item -Path 'test' -Recurse -Force` (Exact
reproduction case).
* **Complex Syntax**: Verified detection inside blocks (e.g., `if
($true) { rm -Force }`) and with trailing punctuation.
*   **CMD**:
    *   `del /f` (Flagged).
    *   `rd /s /q` (Flagged).
    *   Chained commands: `echo hi & del /f file` (Flagged).
*   **False Positives**:
    *   `rd /s` (Not flagged - relies on native prompt).
    *   Standard deletions without force flags.

Verified with `cargo test` and `cargo clippy`.

---------

Co-authored-by: Eric Traut <etraut@openai.com>
2026-01-20 15:25:27 -08:00
Ahmed Ibrahim
fbd8afad81 queue only when task is working (#9558) 2026-01-20 15:24:45 -08:00
Ahmed Ibrahim
de4980d2ac Enable remote models (#9554) 2026-01-20 23:17:22 +00:00
charley-oai
64678f895a Improve UI spacing for queued messages (#9162)
Despite good spacing between queued messages and assistant message text:
<img width="462" height="322" alt="Screenshot 2026-01-12 at 4 54 50 PM"
src="https://github.com/user-attachments/assets/e8b46252-0b33-40d2-b431-cb73b9a3bd2e"
/>

Codex has confusing spacing between queued messages and shimmering
status text (making the queued message seem like a sub-item of the
shimmering status text)
<img width="615" height="217" alt="Screenshot 2026-01-12 at 4 54 18 PM"
src="https://github.com/user-attachments/assets/ee5e6095-8fe9-4863-88d2-10472cab8bd6"
/>

This PR changes the spacing between the queued message(s) and shimmering
status text to make it less confusing:
<img width="440" height="240" alt="Screenshot 2026-01-13 at 11 20 36 AM"
src="https://github.com/user-attachments/assets/02dcc690-cbe9-4943-87de-c7300ef51120"
/>

While working on the status/queued spacing change, we noticed two
paste‑burst tests were timing‑sensitive and could fail
on slower CI. We added a small test‑only helper to keep the paste‑burst
state active and refreshed during these tests. This
removes dependence on tight timing and makes the tests deterministic
without affecting runtime behavior.
2026-01-20 14:54:49 -08:00
zerone0x
ca23b0da5b fix(cli): add execute permission to bin/codex.js (#9532)
## Summary
Fixes #9520

The `bin/codex.js` file was missing execute permissions (`644` instead
of `755`), causing the `codex` command to fail after npm global
installation.

## Changes
- Added execute permission (`+x`) to `codex-cli/bin/codex.js`

## Verification
After this fix, npm tarballs will include the correct file permissions:
```bash
# Before: -rw-r--r-- (644)
# After:  -rwxr-xr-x (755)
```

---
🤖 Generated with Claude Code

Co-authored-by: Claude <noreply@anthropic.com>
2026-01-20 14:53:14 -08:00
charley-oai
be9e55c5fc Add total (non-partial) TextElement placeholder accessors (#9545)
## Summary
- Make `TextElement` placeholders private and add a text-backed accessor
to avoid assuming `Some`.
- Since they are optional in the protocol, we want to make sure any
accessors properly handle the None case (getting the placeholder using
the byte range in the text)
- Preserve placeholders during protocol/app-server conversions using the
accessor fallback.
- Update TUI composer/remap logic and tests to use the new
constructor/accessor.
2026-01-20 14:04:11 -08:00
Ahmed Ibrahim
56fe5e7bea merge remote models (#9547)
We have `models.json` and `/models` response
Behavior:
1. New models from models endpoint gets added
2. Shared models get replaced by remote ones
3. Existing models in `models.json` but not `/models` are kept
4. Mark highest priority as default
2026-01-20 14:02:07 -08:00
Max Kong
c73a11d55e fix(windows-sandbox): parse PATH list entries for audit roots (#9319)
## Summary
- Use `std::env::split_paths` to parse PATH entries in audit candidate
collection
- Add a unit test covering multiple PATH entries (including spaces)

## Testing
- `cargo test -p codex-windows-sandbox` (Windows)

Fixes #9317
2026-01-20 14:00:27 -08:00
Max Kong
f2de920185 fix(windows-sandbox): deny .git file entries under writable roots (#9314)
## Summary
- Deny `.git` entries under writable roots even when `.git` is a file
(worktrees/submodules)
- Add a unit test for `.git` file handling

## Testing
- `cargo test -p codex-windows-sandbox` (Windows)

Fixes #9313
2026-01-20 13:59:59 -08:00
iceweasel-oai
9ea8e3115e lookup system SIDs instead of hardcoding English strings. (#9552)
The elevated setup does not work on non-English windows installs where
Users/Administrators/etc are in different languages. This PR uses the
well-known SIDs instead, which do not vary based on locale
2026-01-20 13:55:37 -08:00
Owen Lin
b0049ab644 fix(core): don't update the file's mtime on resume (#9553)
Remove `FileTimes::new().set_modified(SystemTime::now())` when resuming
a thread.

Context: It's awkward in UI built on top of app-server that resuming a
thread bumps the `updated_at` timestamp, even if no message is sent. So
if you open a thread (perhaps to just view its contents), it
automatically reorders it to the top which is almost certainly not what
you want.
2026-01-20 21:39:31 +00:00
Skylar Graika
b236f1c95d fix: prevent repeating interrupted turns (#9043)
## What
Record a model-visible `<turn_aborted>` marker in history when a turn is
interrupted, and treat it as a session prefix.

## Why
When a turn is interrupted, Codex emits `TurnAborted` but previously did
not persist anything model-visible in the conversation history. On the
next user turn, the model can’t tell the previous work was aborted and
may resume/repeat earlier actions (including duplicated side effects
like re-opening PRs).

Fixes: https://github.com/openai/codex/issues/9042

## How
On `TurnAbortReason::Interrupted`, append a hidden user message
containing a `<turn_aborted>…</turn_aborted>` marker and flush.
Treat `<turn_aborted>` like `<environment_context>` for session-prefix
filtering.
Add a regression test to ensure follow-up turns don’t repeat side
effects from an aborted turn.

## Testing
`just fmt`
`just fix -p codex-core`
`cargo test -p codex-core -- --test-threads=1`
`cargo test --all-features -- --test-threads=1`

---------

Co-authored-by: Skylar Graika <sgraika127@gmail.com>
Co-authored-by: jif-oai <jif@openai.com>
Co-authored-by: Eric Traut <etraut@openai.com>
2026-01-20 13:07:28 -08:00
Eric Traut
79c5bf9835 Fixed config merging issue with profiles (#9509)
This PR fixes a small issue with chained (layered) config.toml file
merging. The old logic didn't properly handle profiles.

In particular, if a lower-layer config overrides a profile defined in a
higher-layer config, the override did not take effect. This prevents
users from having project-specific profile overrides and contradicts the
(soon-to-be) documented behavior of config merging.

The change adds a unit test for this case. It also exposes a function
from the config crate that is needed by the app server code paths to
implement support for layered configs.
2026-01-20 12:18:00 -08:00
jif-oai
0b3c802a54 fix: memory leak issue (#9543)
Co-authored-by: Josh McKinney <joshka@openai.com>
2026-01-20 20:14:14 +00:00
Dylan Hurd
714151eb4e feat(personality) introduce model_personality config (#9459)
## Summary
Introduces the concept of a config model_personality. I would consider
this an MVP for testing out the feature. There are a number of
follow-ups to this PR:

- More sophisticated templating with validation
- In-product experience to manage this

## Testing
- [x] Testing locally
2026-01-20 11:06:14 -08:00
Simon Willison
46a4a03083 Fix typo in feature name from 'Mult-agents' to 'Multi-agents' (#9542)
Fixes a typo in a feature description.
2026-01-20 10:55:36 -08:00
Tiffany Citra
2c3843728c fix: writable_roots doesn't recognize home directory symbol in non-windows OS (#9193)
Fixes:
```
[sandbox_workspace_write]
writable_roots = ["~/code/"]
```

translates to
```
/Users/ccunningham/.codex/~/code
```
(i.e. the home dir symbol isn't recognized)
2026-01-20 10:55:01 -08:00
Ahmed Ibrahim
5ae6e70801 Tui: use collaboration mode instead of model and effort (#9507)
- Only use collaboration modes in the tui state to track model and
effort.
- No behavior change without the collaboration modes flag.
- Change model and effort on /model, /collab (behind a flag), and
shift+tab (behind flag)
2026-01-20 10:26:12 -08:00
Anton Panasenko
7b27aa7707 feat: support proxy for ws connection (#9409)
unfortunately tokio-tungstenite doesn't support proxy configuration
outbox, while https://github.com/snapview/tokio-tungstenite/pull/370 is
in review, we can depend on source code for now.
2026-01-20 09:36:30 -08:00
gt-oai
7351c12999 Only load config from trusted folders (#9533)
Config includes multiple code execution entrypoints. 

Now, we load the config from predetermined locations first
(~/.codex/config.toml etc), use those to learn which folders are
'trusted', and only load additional config from the CWD if it is
trusted.
2026-01-20 15:44:21 +00:00
jif-oai
3a9f436ce0 feat: metrics on shell snapshot (#9527) 2026-01-20 13:18:24 +00:00
jif-oai
6bbf506120 feat: metrics on remote models (#9528) 2026-01-20 13:02:55 +00:00
jif-oai
a3a97f3ea9 feat: record timer with additional tags (#9529) 2026-01-20 13:01:55 +00:00
jif-oai
9ec20ba065 nit: do not render terminal interactions if no task running (#9374)
To prevent race where the terminal interaction message is processed
after the last message
2026-01-20 10:20:17 +00:00
jif-oai
483239d861 chore: collab in experimental (#9525) 2026-01-20 10:19:06 +00:00
Dylan Hurd
3078eedb24 fix(tui) fix user message light mode background (#9407)
## Summary
Fixes the user message styles for light mode.

## Testing
Attaching 2 screenshots from ghostty, but I also tried various styles in
Terminal.app and iTerm2.
**Before**
<img width="888" height="560" alt="Screenshot 2026-01-16 at 5 22 36 PM"
src="https://github.com/user-attachments/assets/73d9decb-a01a-4ece-b88e-ea49a33cc0c6"
/>

**After**
<img width="890" height="281" alt="Screenshot 2026-01-16 at 5 22 59 PM"
src="https://github.com/user-attachments/assets/6689e286-d699-4ceb-b0cb-579a31b047bf"
/>
2026-01-19 23:58:44 -08:00
charley-oai
eb90e20c0b Persist text elements through TUI input and history (#9393)
Continuation of breaking up this PR
https://github.com/openai/codex/pull/9116

## Summary
- Thread user text element ranges through TUI/TUI2 input, submission,
queueing, and history so placeholders survive resume/edit flows.
- Preserve local image attachments alongside text elements and rehydrate
placeholders when restoring drafts.
- Keep model-facing content shapes clean by attaching UI metadata only
to user input/events (no API content changes).

## Key Changes
- TUI/TUI2 composer now captures text element ranges, trims them with
text edits, and restores them when submission is suppressed.
- User history cells render styled spans for text elements and keep
local image paths for future rehydration.
- Initial chat widget bootstraps accept empty `initial_text_elements` to
keep initialization uniform.
- Protocol/core helpers updated to tolerate the new InputText field
shape without changing payloads sent to the API.
2026-01-19 23:49:34 -08:00
Dylan Hurd
675f165c56 fix(core) Preserve base_instructions in SessionMeta (#9427)
## Summary
This PR consolidates base_instructions onto SessionMeta /
SessionConfiguration, so we ensure `base_instructions` is set once per
session and should be (mostly) immutable, unless:
- overridden by config on resume / fork
- sub-agent tasks, like review or collab


In a future PR, we should convert all references to `base_instructions`
to consistently used the typed struct, so it's less likely that we put
other strings there. See #9423. However, this PR is already quite
complex, so I'm deferring that to a follow-up.

## Testing
- [x] Added a resume test to assert that instructions are preserved. In
particular, `resume_switches_models_preserves_base_instructions` fails
against main.

Existing test coverage thats assert base instructions are preserved
across multiple requests in a session:
- Manual compact keeps baseline instructions:
core/tests/suite/compact.rs:199
- Auto-compact keeps baseline instructions:
core/tests/suite/compact.rs:1142
- Prompt caching reuses the same instructions across two requests:
core/tests/suite/prompt_caching.rs:150 and
core/tests/suite/prompt_caching.rs:157
- Prompt caching with explicit expected string across two requests:
core/tests/suite/prompt_caching.rs:213 and
core/tests/suite/prompt_caching.rs:222
- Resume with model switch keeps original instructions:
core/tests/suite/resume.rs:136
- Compact/resume/fork uses request 0 instructions for later expected
payloads: core/tests/suite/compact_resume_fork.rs:215
2026-01-19 21:59:36 -08:00
Ahmed Ibrahim
65d3b9e145 Migrate tui to use UserTurn (#9497)
- `tui/` and `tui2/` submit `Op::UserTurn` and own full turn context
(cwd/approval/sandbox/model/etc.).
- `Op::UserInput` is documented as legacy in `codex-protocol` (doc-only;
no `#[deprecated]` to avoid `-D warnings` fallout).
- Remove obsolete `#[allow(deprecated)]` and the unused `ConversationId`
alias/re-export.
2026-01-19 13:40:39 -08:00
prateek-oai
0c0c5aeddc tui: avoid Esc interrupt when skill popup active (#9451)
Fixes #9450

## What
- When a task is running and the skills autocomplete popup is open,
`Esc` now dismisses the popup instead of sending `Op::Interrupt`.
- `Esc` still interrupts a running task when no popup is active.

## Tests
- `cargo test -p codex-tui`

---------

Co-authored-by: prateek <199982+prateek@users.noreply.github.com>
2026-01-19 12:52:04 -08:00
Shijie Rao
d544adf71a Feat: plan mode prompt update (#9495)
### Summary
* Added instruction on using `request_user_input`
* Added the output to be json with `plan` key and the actual plan as the
value.
* Remove `PLAN.md` write because that gets into sandbox issue. We can
add it back later.
2026-01-19 12:17:29 -08:00
jif-oai
070935d5e8 chore: fix beta VS experimental (#9496) 2026-01-19 19:39:23 +00:00
Ahmed Ibrahim
b11e96fb04 Act on reasoning-included per turn (#9402)
- Reset reasoning-included flag each turn and update compaction test
2026-01-19 11:23:25 -08:00
Shijie Rao
57ec3a8277 Feat: request user input tool (#9472)
### Summary
* Add `requestUserInput` tool that the model can use for gather
feedback/asking question mid turn.


### Tool input schema
```
{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "title": "requestUserInput input",
  "type": "object",
  "additionalProperties": false,
  "required": ["questions"],
  "properties": {
    "questions": {
      "type": "array",
      "description": "Questions to show the user (1-3). Prefer 1 unless multiple independent decisions block progress.",
      "minItems": 1,
      "maxItems": 3,
      "items": {
        "type": "object",
        "additionalProperties": false,
        "required": ["id", "header", "question"],
        "properties": {
          "id": {
            "type": "string",
            "description": "Stable identifier for mapping answers (snake_case)."
          },
          "header": {
            "type": "string",
            "description": "Short header label shown in the UI (12 or fewer chars)."
          },
          "question": {
            "type": "string",
            "description": "Single-sentence prompt shown to the user."
          },
          "options": {
            "type": "array",
            "description": "Optional 2-3 mutually exclusive choices. Put the recommended option first and suffix its label with \"(Recommended)\". Only include \"Other\" option if we want to include a free form option. If the question is free form in nature, do not include any option.",
            "minItems": 2,
            "maxItems": 3,
            "items": {
              "type": "object",
              "additionalProperties": false,
              "required": ["value", "label", "description"],
              "properties": {
                "value": {
                  "type": "string",
                  "description": "Machine-readable value (snake_case)."
                },
                "label": {
                  "type": "string",
                  "description": "User-facing label (1-5 words)."
                },
                "description": {
                  "type": "string",
                  "description": "One short sentence explaining impact/tradeoff if selected."
                }
              }
            }
          }
        }
      }
    }
  }
}
```

### Tool output schema
```
{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "title": "requestUserInput output",
  "type": "object",
  "additionalProperties": false,
  "required": ["answers"],
  "properties": {
    "answers": {
      "type": "object",
      "description": "Map of question id to user answer.",
      "additionalProperties": {
        "type": "object",
        "additionalProperties": false,
        "required": ["selected"],
        "properties": {
          "selected": {
            "type": "array",
            "items": { "type": "string" }
          },
          "other": {
            "type": ["string", "null"]
          }
        }
      }
    }
  }
}
```
2026-01-19 10:17:30 -08:00
Ahmed Ibrahim
bf430ad9fe TUI: collaboration mode UX + always submit UserTurn when enabled (#9461)
- Adds experimental collaboration modes UX in TUI: Plan / Pair
Programming / Execute.
- Gated behind `Feature::CollaborationModes`; existing behavior remains
unchanged when disabled.
- Selection UX:
- `Shift+Tab` cycles modes while idle (no task running, no modal/popup).
- `/collab` cycles; `/collab <plan|pair|pp|execute|exec>` sets
explicitly.
- Footer flash after changes + shortcut overlay shows `Shift+Tab` “to
change mode”.
  - `/status` shows “Collaboration mode”.
- Submission semantics:
- When enabled: every submit uses `Op::UserTurn` and always includes
`collaboration_mode: Some(...)` (default Pair Programming).
  - Removes the one-shot “pending collaboration mode” behavior.
- Implementation:
- New `tui/src/collaboration_modes.rs` (selection enum/cycle, `/collab`
parsing, resolve to `CollaborationMode`, footer flash line).
- Fallback: `resolve_mode_or_fallback` synthesizes a `CollaborationMode`
when presets are missing (uses current model + reasoning effort; no
`developer_instructions`) to avoid core falling back to `Custom`.
  - TODO: migrate TUI to use `Op::UserTurn`.
2026-01-19 09:32:04 -08:00
Eric Traut
3788e2cc0f Fixed stale link to MCP documentation (#9490)
This was noted in #9482
2026-01-19 09:10:54 -08:00
jif-oai
92cf2a1c3a chore: warning metric (#9487) 2026-01-19 17:07:04 +00:00
Ahmed Ibrahim
31415ebfcf Remove unused protocol collaboration mode prompts (#9463)
Delete duplicate collaboration mode markdown under protocol prompts;
core templates remain the single source of truth.
2026-01-19 08:54:19 -08:00
Eric Traut
264d40efdc Fix invalid input error on Azure endpoint (#9387)
Users of Azure endpoints are reporting that when they use `/review`,
they sometimes see an error "Invalid 'input[3].id". I suspect this is
specific to the Azure implementation of the `responses` API. The Azure
team generally copies the OpenAI code for this endpoint, but they do
have minor differences and sometimes lag in rolling out bug fixes or
updates.

The error appears to be triggered because the `/review` implementation
is using a user ID with a colon in it.

Addresses #9360
2026-01-19 08:21:27 -08:00
jif-oai
3c28c85063 prompt 3 (#9479) 2026-01-19 11:56:38 +00:00
jif-oai
dc1b62acbd feat: detach non-tty childs (#9477)
Thanks to the investigations made by
* @frantic-openai https://github.com/openai/codex/pull/9403
* @kfiramar https://github.com/openai/codex/pull/9388
2026-01-19 11:35:34 +00:00
jif-oai
186794dbb3 feat: close all threads in /new (#9478) 2026-01-19 11:35:03 +00:00
jif-oai
7ebe13f692 feat: timer total turn metrics (#9382) 2026-01-19 10:44:31 +00:00
Eric Traut
a803467f52 Fixed TUI regression related to image paste in WSL (#9473)
This affects quoted paths in WSL. The fix is to normalize quoted Windows
paths before WSL conversion.

This addresses #9456
2026-01-18 23:02:02 -08:00
dependabot[bot]
a5e5d7a384 chore(deps): bump chrono from 0.4.42 to 0.4.43 in /codex-rs (#9465)
Bumps [chrono](https://github.com/chronotope/chrono) from 0.4.42 to
0.4.43.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/chronotope/chrono/releases">chrono's
releases</a>.</em></p>
<blockquote>
<h2>0.4.43</h2>
<h2>What's Changed</h2>
<ul>
<li>Install extra components for lint workflow by <a
href="https://github.com/djc"><code>@​djc</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1741">chronotope/chrono#1741</a></li>
<li>Upgrade windows-bindgen to 0.64 by <a
href="https://github.com/djc"><code>@​djc</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1742">chronotope/chrono#1742</a></li>
<li>Improve windows-bindgen setup by <a
href="https://github.com/djc"><code>@​djc</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1744">chronotope/chrono#1744</a></li>
<li>Drop stabilized feature doc_auto_cfg by <a
href="https://github.com/djc"><code>@​djc</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1745">chronotope/chrono#1745</a></li>
<li>Faster RFC 3339 parsing by <a
href="https://github.com/djc"><code>@​djc</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1748">chronotope/chrono#1748</a></li>
<li>Update windows-bindgen requirement from 0.64 to 0.65 by <a
href="https://github.com/dependabot"><code>@​dependabot</code></a>[bot]
in <a
href="https://redirect.github.com/chronotope/chrono/pull/1751">chronotope/chrono#1751</a></li>
<li>add <code>NaiveDate::abs_diff</code> by <a
href="https://github.com/Kinrany"><code>@​Kinrany</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1752">chronotope/chrono#1752</a></li>
<li>Add feature gated defmt support. by <a
href="https://github.com/pebender"><code>@​pebender</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1747">chronotope/chrono#1747</a></li>
<li>Drop deny lints, eager Debug impls are a mixed blessing by <a
href="https://github.com/djc"><code>@​djc</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1753">chronotope/chrono#1753</a></li>
<li>chore: minor improvement for docs by <a
href="https://github.com/spuradage"><code>@​spuradage</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1756">chronotope/chrono#1756</a></li>
<li>Added doctest for the NaiveDate years_since function by <a
href="https://github.com/LucasBou"><code>@​LucasBou</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1755">chronotope/chrono#1755</a></li>
<li>Prepare 0.4.43 by <a
href="https://github.com/djc"><code>@​djc</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1765">chronotope/chrono#1765</a></li>
<li>Update copyright year to 2026 in LICENSE.txt by <a
href="https://github.com/taozui472"><code>@​taozui472</code></a> in <a
href="https://redirect.github.com/chronotope/chrono/pull/1767">chronotope/chrono#1767</a></li>
</ul>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="45caaa970c"><code>45caaa9</code></a>
Update copyright year to 2026 in LICENSE.txt</li>
<li><a
href="1c0b8f011a"><code>1c0b8f0</code></a>
Bump version to 0.4.43</li>
<li><a
href="a03e43b1c3"><code>a03e43b</code></a>
Upgrade windows-bindgen to 0.66</li>
<li><a
href="4fedaba2a2"><code>4fedaba</code></a>
Ignore bincode advisory</li>
<li><a
href="f4b7bbda67"><code>f4b7bbd</code></a>
Bump actions/checkout from 5 to 6</li>
<li><a
href="db129730e8"><code>db12973</code></a>
Added doctest for the NaiveDate years_since function (<a
href="https://redirect.github.com/chronotope/chrono/issues/1755">#1755</a>)</li>
<li><a
href="34b5f49e9d"><code>34b5f49</code></a>
chore: minor improvement for docs</li>
<li><a
href="8c827116b9"><code>8c82711</code></a>
Bump actions/setup-node from 5 to 6</li>
<li><a
href="ea1f11b356"><code>ea1f11b</code></a>
Drop deny lints, eager Debug impls are a mixed blessing</li>
<li><a
href="35f9f2daef"><code>35f9f2d</code></a>
Add feature gated defmt support.</li>
<li>Additional commits viewable in <a
href="https://github.com/chronotope/chrono/compare/v0.4.42...v0.4.43">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=chrono&package-manager=cargo&previous-version=0.4.42&new-version=0.4.43)](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>
2026-01-19 06:43:34 +00:00
dependabot[bot]
66b74efbc6 chore(deps): bump ctor from 0.5.0 to 0.6.3 in /codex-rs (#9469)
Bumps [ctor](https://github.com/mmastrac/rust-ctor) from 0.5.0 to 0.6.3.
<details>
<summary>Commits</summary>
<ul>
<li>See full diff in <a
href="https://github.com/mmastrac/rust-ctor/commits">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=ctor&package-manager=cargo&previous-version=0.5.0&new-version=0.6.3)](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>
2026-01-18 22:37:11 -08:00
dependabot[bot]
78a359f7fa chore(deps): bump arc-swap from 1.7.1 to 1.8.0 in /codex-rs (#9468)
Bumps [arc-swap](https://github.com/vorner/arc-swap) from 1.7.1 to
1.8.0.
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/vorner/arc-swap/blob/master/CHANGELOG.md">arc-swap's
changelog</a>.</em></p>
<blockquote>
<h1>1.8.0</h1>
<ul>
<li>Support for Pin (<a
href="https://redirect.github.com/vorner/arc-swap/issues/185">#185</a>,
<a
href="https://redirect.github.com/vorner/arc-swap/issues/183">#183</a>).</li>
<li>Fix (hopefully) crash on ARM (<a
href="https://redirect.github.com/vorner/arc-swap/issues/164">#164</a>).</li>
<li>Fix Miri check (<a
href="https://redirect.github.com/vorner/arc-swap/issues/186">#186</a>,
<a
href="https://redirect.github.com/vorner/arc-swap/issues/156">#156</a>).</li>
<li>Fix support for Rust 1.31.0.</li>
<li>Some minor clippy lints.</li>
</ul>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="2540d266a8"><code>2540d26</code></a>
Version bump to 1.8.0</li>
<li><a
href="9981e3af23"><code>9981e3a</code></a>
Keep &quot;old&quot; Cargo.lock around</li>
<li><a
href="57a8abbfc4"><code>57a8abb</code></a>
Fix documentation links</li>
<li><a
href="346c5b642b"><code>346c5b6</code></a>
Fix some clippy warnings</li>
<li><a
href="0bd349a56b"><code>0bd349a</code></a>
Fix support for Rust 1.31.0</li>
<li><a
href="57aa5224c1"><code>57aa522</code></a>
Merge pull request <a
href="https://redirect.github.com/vorner/arc-swap/issues/185">#185</a>
from SpriteOvO/pin</li>
<li><a
href="4c0c4ab321"><code>4c0c4ab</code></a>
Implement <code>RefCnt</code> for <code>Pin\&lt;Arc&gt;</code> and
<code>Pin\&lt;Rc&gt;</code></li>
<li><a
href="e596275acf"><code>e596275</code></a>
Avoid warnings about hidden lifetimes</li>
<li><a
href="d849a2d17e"><code>d849a2d</code></a>
Use SeqCst in debt-lists</li>
<li><a
href="1f9b221da9"><code>1f9b221</code></a>
Merge pull request <a
href="https://redirect.github.com/vorner/arc-swap/issues/186">#186</a>
from nbdd0121/prov</li>
<li>Additional commits viewable in <a
href="https://github.com/vorner/arc-swap/compare/v1.7.1...v1.8.0">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=arc-swap&package-manager=cargo&previous-version=1.7.1&new-version=1.8.0)](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>
2026-01-18 22:25:23 -08:00
dependabot[bot]
274af30525 chore(deps): bump tokio from 1.48.0 to 1.49.0 in /codex-rs (#9467)
Bumps [tokio](https://github.com/tokio-rs/tokio) from 1.48.0 to 1.49.0.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/tokio-rs/tokio/releases">tokio's
releases</a>.</em></p>
<blockquote>
<h2>Tokio v1.49.0</h2>
<h1>1.49.0 (January 3rd, 2026)</h1>
<h3>Added</h3>
<ul>
<li>net: add support for <code>TCLASS</code> option on IPv6 (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7781">#7781</a>)</li>
<li>runtime: stabilize <code>runtime::id::Id</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7125">#7125</a>)</li>
<li>task: implement <code>Extend</code> for <code>JoinSet</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7195">#7195</a>)</li>
<li>task: stabilize the <code>LocalSet::id()</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7776">#7776</a>)</li>
</ul>
<h3>Changed</h3>
<ul>
<li>net: deprecate <code>{TcpStream,TcpSocket}::set_linger</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7752">#7752</a>)</li>
</ul>
<h3>Fixed</h3>
<ul>
<li>macros: fix the hygiene issue of <code>join!</code> and
<code>try_join!</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7766">#7766</a>)</li>
<li>runtime: revert &quot;replace manual vtable definitions with
Wake&quot; (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7699">#7699</a>)</li>
<li>sync: return <code>TryRecvError::Disconnected</code> from
<code>Receiver::try_recv</code> after <code>Receiver::close</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7686">#7686</a>)</li>
<li>task: remove unnecessary trait bounds on the <code>Debug</code>
implementation (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7720">#7720</a>)</li>
</ul>
<h3>Unstable</h3>
<ul>
<li>fs: handle <code>EINTR</code> in <code>fs::write</code> for io-uring
(<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7786">#7786</a>)</li>
<li>fs: support io-uring with <code>tokio::fs::read</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7696">#7696</a>)</li>
<li>runtime: disable io-uring on <code>EPERM</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7724">#7724</a>)</li>
<li>time: add alternative timer for better multicore scalability (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7467">#7467</a>)</li>
</ul>
<h3>Documented</h3>
<ul>
<li>docs: fix a typos in <code>bounded.rs</code> and
<code>park.rs</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7817">#7817</a>)</li>
<li>io: add <code>SyncIoBridge</code> cross-references to
<code>copy</code> and <code>copy_buf</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7798">#7798</a>)</li>
<li>io: doc that <code>AsyncWrite</code> does not inherit from
<code>std::io::Write</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7705">#7705</a>)</li>
<li>metrics: clarify that <code>num_alive_tasks</code> is not strongly
consistent (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7614">#7614</a>)</li>
<li>net: clarify the cancellation safety of the
<code>TcpStream::peek</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7305">#7305</a>)</li>
<li>net: clarify the drop behavior of <code>unix::OwnedWriteHalf</code>
(<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7742">#7742</a>)</li>
<li>net: clarify the platform-dependent backlog in
<code>TcpSocket</code> docs (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7738">#7738</a>)</li>
<li>runtime: mention <code>LocalRuntime</code> in
<code>new_current_thread</code> docs (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7820">#7820</a>)</li>
<li>sync: add missing period to <code>mpsc::Sender::try_send</code> docs
(<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7721">#7721</a>)</li>
<li>sync: clarify the cancellation safety of
<code>oneshot::Receiver</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7780">#7780</a>)</li>
<li>sync: improve the docs for the <code>errors</code> of mpsc (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7722">#7722</a>)</li>
<li>task: add example for <code>spawn_local</code> usage on local
runtime (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7689">#7689</a>)</li>
</ul>
<p><a
href="https://redirect.github.com/tokio-rs/tokio/issues/7125">#7125</a>:
<a
href="https://redirect.github.com/tokio-rs/tokio/pull/7125">tokio-rs/tokio#7125</a>
<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7195">#7195</a>:
<a
href="https://redirect.github.com/tokio-rs/tokio/pull/7195">tokio-rs/tokio#7195</a>
<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7305">#7305</a>:
<a
href="https://redirect.github.com/tokio-rs/tokio/pull/7305">tokio-rs/tokio#7305</a>
<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7467">#7467</a>:
<a
href="https://redirect.github.com/tokio-rs/tokio/pull/7467">tokio-rs/tokio#7467</a>
<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7614">#7614</a>:
<a
href="https://redirect.github.com/tokio-rs/tokio/pull/7614">tokio-rs/tokio#7614</a>
<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7686">#7686</a>:
<a
href="https://redirect.github.com/tokio-rs/tokio/pull/7686">tokio-rs/tokio#7686</a>
<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7689">#7689</a>:
<a
href="https://redirect.github.com/tokio-rs/tokio/pull/7689">tokio-rs/tokio#7689</a></p>
<!-- raw HTML omitted -->
</blockquote>
<p>... (truncated)</p>
</details>
<details>
<summary>Commits</summary>
<ul>
<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><a
href="ea6b144cd1"><code>ea6b144</code></a>
ci: freeze rustc on nightly-2025-01-25 in <code>netlify.toml</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7652">#7652</a>)</li>
<li><a
href="264e703296"><code>264e703</code></a>
Merge <code>tokio-1.43.4</code> into <code>tokio-1.47.x</code> (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7822">#7822</a>)</li>
<li><a
href="dfb0f00838"><code>dfb0f00</code></a>
chore: prepare Tokio v1.43.4 (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7821">#7821</a>)</li>
<li><a
href="4a91f197b0"><code>4a91f19</code></a>
ci: fix wasm32-wasip1 tests (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7788">#7788</a>)</li>
<li><a
href="601c383ab6"><code>601c383</code></a>
ci: upgrade FreeBSD from 14.2 to 14.3 (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7758">#7758</a>)</li>
<li><a
href="484cb52d8d"><code>484cb52</code></a>
sync: return <code>TryRecvError::Disconnected</code> from
<code>Receiver::try_recv</code> after `Re...</li>
<li><a
href="16f20c34ed"><code>16f20c3</code></a>
rt: mention <code>LocalRuntime</code> in <code>new_current_thread</code>
docs (<a
href="https://redirect.github.com/tokio-rs/tokio/issues/7820">#7820</a>)</li>
<li>Additional commits viewable in <a
href="https://github.com/tokio-rs/tokio/compare/tokio-1.48.0...tokio-1.49.0">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=tokio&package-manager=cargo&previous-version=1.48.0&new-version=1.49.0)](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>
2026-01-18 22:24:55 -08:00
dependabot[bot]
efa9326f08 chore(deps): bump log from 0.4.28 to 0.4.29 in /codex-rs (#9466)
Bumps [log](https://github.com/rust-lang/log) from 0.4.28 to 0.4.29.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/rust-lang/log/releases">log's
releases</a>.</em></p>
<blockquote>
<h2>0.4.29</h2>
<h2>MSRV</h2>
<p>This release increases <code>log</code>'s MSRV from
<code>1.61.0</code> to <code>1.68.0</code>.</p>
<h2>What's Changed</h2>
<ul>
<li>docs: Add missing impls from README.md by <a
href="https://github.com/AldaronLau"><code>@​AldaronLau</code></a> in <a
href="https://redirect.github.com/rust-lang/log/pull/703">rust-lang/log#703</a></li>
<li>Point to new URLs for favicon and logo by <a
href="https://github.com/AldaronLau"><code>@​AldaronLau</code></a> in <a
href="https://redirect.github.com/rust-lang/log/pull/704">rust-lang/log#704</a></li>
<li>perf: reduce llvm-lines of FromStr for <code>Level</code> and
<code>LevelFilter</code> by <a
href="https://github.com/dishmaker"><code>@​dishmaker</code></a> in <a
href="https://redirect.github.com/rust-lang/log/pull/709">rust-lang/log#709</a></li>
<li>Replace serde with serde_core by <a
href="https://github.com/Thomasdezeeuw"><code>@​Thomasdezeeuw</code></a>
in <a
href="https://redirect.github.com/rust-lang/log/pull/712">rust-lang/log#712</a></li>
<li>Fix clippy lints by <a
href="https://github.com/Thomasdezeeuw"><code>@​Thomasdezeeuw</code></a>
in <a
href="https://redirect.github.com/rust-lang/log/pull/713">rust-lang/log#713</a></li>
<li>Use GitHub Actions to install Rust and cargo-hack by <a
href="https://github.com/Thomasdezeeuw"><code>@​Thomasdezeeuw</code></a>
in <a
href="https://redirect.github.com/rust-lang/log/pull/715">rust-lang/log#715</a></li>
<li>Exclude old unstable_kv features from testing matrix by <a
href="https://github.com/Thomasdezeeuw"><code>@​Thomasdezeeuw</code></a>
in <a
href="https://redirect.github.com/rust-lang/log/pull/716">rust-lang/log#716</a></li>
<li>Fix up CI by <a
href="https://github.com/KodrAus"><code>@​KodrAus</code></a> in <a
href="https://redirect.github.com/rust-lang/log/pull/718">rust-lang/log#718</a></li>
<li>Prepare for 0.4.29 release by <a
href="https://github.com/KodrAus"><code>@​KodrAus</code></a> in <a
href="https://redirect.github.com/rust-lang/log/pull/719">rust-lang/log#719</a></li>
</ul>
<h2>New Contributors</h2>
<ul>
<li><a
href="https://github.com/AldaronLau"><code>@​AldaronLau</code></a> made
their first contribution in <a
href="https://redirect.github.com/rust-lang/log/pull/703">rust-lang/log#703</a></li>
<li><a href="https://github.com/dishmaker"><code>@​dishmaker</code></a>
made their first contribution in <a
href="https://redirect.github.com/rust-lang/log/pull/709">rust-lang/log#709</a></li>
</ul>
<p><strong>Full Changelog</strong>: <a
href="https://github.com/rust-lang/log/compare/0.4.28...0.4.29">https://github.com/rust-lang/log/compare/0.4.28...0.4.29</a></p>
</blockquote>
</details>
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/rust-lang/log/blob/master/CHANGELOG.md">log's
changelog</a>.</em></p>
<blockquote>
<h2>[0.4.29] - 2025-12-02</h2>
<h2>What's Changed</h2>
<ul>
<li>perf: reduce llvm-lines of FromStr for <code>Level</code> and
<code>LevelFilter</code> by <a
href="https://github.com/dishmaker"><code>@​dishmaker</code></a> in <a
href="https://redirect.github.com/rust-lang/log/pull/709">rust-lang/log#709</a></li>
<li>Replace serde with serde_core by <a
href="https://github.com/Thomasdezeeuw"><code>@​Thomasdezeeuw</code></a>
in <a
href="https://redirect.github.com/rust-lang/log/pull/712">rust-lang/log#712</a></li>
</ul>
<h2>New Contributors</h2>
<ul>
<li><a
href="https://github.com/AldaronLau"><code>@​AldaronLau</code></a> made
their first contribution in <a
href="https://redirect.github.com/rust-lang/log/pull/703">rust-lang/log#703</a></li>
<li><a href="https://github.com/dishmaker"><code>@​dishmaker</code></a>
made their first contribution in <a
href="https://redirect.github.com/rust-lang/log/pull/709">rust-lang/log#709</a></li>
</ul>
<p><strong>Full Changelog</strong>: <a
href="https://github.com/rust-lang/log/compare/0.4.28...0.4.29">https://github.com/rust-lang/log/compare/0.4.28...0.4.29</a></p>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="b1e2df7bce"><code>b1e2df7</code></a>
Merge pull request <a
href="https://redirect.github.com/rust-lang/log/issues/719">#719</a>
from rust-lang/cargo/0.4.29</li>
<li><a
href="3fe1a546dc"><code>3fe1a54</code></a>
prepare for 0.4.29 release</li>
<li><a
href="7a432d9ab5"><code>7a432d9</code></a>
Merge pull request <a
href="https://redirect.github.com/rust-lang/log/issues/718">#718</a>
from rust-lang/ci/msrv</li>
<li><a
href="0689d56847"><code>0689d56</code></a>
rebump msrv to 1.68.0</li>
<li><a
href="46b448e2a7"><code>46b448e</code></a>
try drop msrv back to 1.61.0</li>
<li><a
href="929ab3812e"><code>929ab38</code></a>
fix up doc test feature gate</li>
<li><a
href="957cece478"><code>957cece</code></a>
bump serde-dependent crates</li>
<li><a
href="bea40c847c"><code>bea40c8</code></a>
bump msrv to 1.68.0</li>
<li><a
href="c540184ee9"><code>c540184</code></a>
Merge pull request <a
href="https://redirect.github.com/rust-lang/log/issues/716">#716</a>
from rust-lang/ci-smaller-matrix2</li>
<li><a
href="c971e636c4"><code>c971e63</code></a>
Merge branch 'master' into ci-smaller-matrix2</li>
<li>Additional commits viewable in <a
href="https://github.com/rust-lang/log/compare/0.4.28...0.4.29">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=log&package-manager=cargo&previous-version=0.4.28&new-version=0.4.29)](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>
2026-01-18 22:24:34 -08:00
Eric Traut
1271d450b1 Fixed symlink support for config.toml (#9445)
We already support reading from `config.toml` through a symlink, but the
code was not properly handling updates to a symlinked config file. This
PR generalizes safe symlink-chain resolution and atomic writes into
path_utils, updating all config write paths to use the shared logic
(including set_default_oss_provider, which previously didn't use the
common path), and adds tests for symlink chains and cycles.

This resolves #6646.

Notes:
* Symlink cycles or resolution failures replace the top-level symlink
with a real file.
* Shared config write path now handles symlinks consistently across
edits, defaults, and empty-user-layer creation.

This PR was inspired by https://github.com/openai/codex/pull/9437, which
was contributed by @ryoppippi
2026-01-18 19:22:28 -08:00
Michael Bolin
c87a7d9043 fix(tui2): running /mcp was not printing any output until another event triggered a flush (#9457)
This seems to fix things, but I'm not sure if this is the
correct/minimal fix.
2026-01-18 19:19:37 -08:00
Ahmed Ibrahim
f72f87fbee Add collaboration modes test prompts (#9443)
# 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.
2026-01-18 11:39:08 -08:00
SlKzᵍᵐ
0a568a47fd tui: allow forward navigation in backtrack preview (#9059)
Fixes #9058

## Summary
When the transcript backtrack preview is armed (press `Esc`), allow
navigating to newer user messages with the `→` arrow, in addition to
navigating backwards with `Esc`/`←`, before confirming with `Enter`.

## Changes
- Backtrack preview navigation: `Esc`/`←` steps to older user messages,
`→` steps to newer ones, `Enter` edits the selected message (clamped at
bounds, no wrap-around).
- Transcript overlay footer hints updated to advertise `esc/←`, `→`, and
`enter` when a message is highlighted.

## Related
- WSL shortcut-overlay snapshot determinism: #9359

## Testing
- `just fmt`
- `just fix -p codex-tui`
- `just fix -p codex-tui2`
- `cargo test -p codex-tui app_backtrack::`
- `cargo test -p codex-tui pager_overlay::`
- `cargo test -p codex-tui2 app_backtrack::`
- `cargo test -p codex-tui2 pager_overlay::`

---------

Co-authored-by: Josh McKinney <joshka@openai.com>
2026-01-18 04:10:24 +00:00
Ahmed Ibrahim
aeaff26451 Preserve slash command order in search (#9425)
Keep slash popup search results in presentation order for built-ins and
prompts.
2026-01-17 19:46:07 -08:00
Ahmed Ibrahim
1478a88eb0 Add collaboration developer instructions (#9424)
- Add additional instructions when they are available
- Make sure to update them on change either UserInput or UserTurn
2026-01-18 01:31:14 +00:00
Dylan Hurd
80d7a5d7fe chore(instructions) Remove unread SessionMeta.instructions field (#9423)
### Description
- Remove the now-unused `instructions` field from the session metadata
to simplify SessionMeta and stop propagating transient instruction text
through the rollout recorder API. This was only saving
user_instructions, and was never being read.
- Stop passing user instructions into the rollout writer at session
creation so the rollout header only contains canonical session metadata.

### Testing

- Ran `just fmt` which completed successfully.
- Ran `just fix -p codex-protocol`, `just fix -p codex-core`, `just fix
-p codex-app-server`, `just fix -p codex-tui`, and `just fix -p
codex-tui2` which completed (Clippy fixes applied) as part of
verification.
- Ran `cargo test -p codex-protocol` which passed (28 tests).
- Ran `cargo test -p codex-core` which showed failures in a small set of
tests (not caused by the protocol type change directly):
`default_client::tests::test_create_client_sets_default_headers`,
several `models_manager::manager::tests::refresh_available_models_*`,
and `shell_snapshot::tests::linux_sh_snapshot_includes_sections` (these
tests failed in this CI run).
- Ran `cargo test -p codex-app-server` which reported several failing
integration tests (including
`suite::codex_message_processor_flow::test_codex_jsonrpc_conversation_flow`,
`suite::output_schema::send_user_turn_*`, and
`suite::user_agent::get_user_agent_returns_current_codex_user_agent`).
- `cargo test -p codex-tui` and `cargo test -p codex-tui2` were
attempted but aborted due to disk space exhaustion (`No space left on
device`).

------
[Codex
Task](https://chatgpt.com/codex/tasks/task_i_696bd8ce632483228d298cf07c7eb41c)
2026-01-17 16:02:28 -08:00
Dylan Hurd
bffe9b33e9 chore(core) Create instructions module (#9422)
## Summary
We have a variety of things we refer to as instructions in the code
base: our current canonical terms are:
- base instructions (raw string)
- developer instructions (has a type in protocol)
- user instructions

We also have `instructions` floating around in various places. We should
standardize on the above, and start using types to prevent them from
ending up in the wrong place. There will be additional PRs, but I'm
going to keep these small so we can easily follow them!

## Testing
- [x] Tests pass, this is purely a file move
2026-01-17 16:01:26 -08:00
Ahmed Ibrahim
8f0e0300d2 Expose collaboration presets (#9421)
Expose collaboration presets for clients

---------

Co-authored-by: Josh McKinney <joshka@openai.com>
2026-01-17 12:32:50 -08:00
Alex Hornby
b877a2041e fix unified_exec::tests::unified_exec_timeouts to use a more unique match value (#9414)
Fix unified_exec_timeouts to use a unique variable value rather than
"codex" which was causing false positives when running tests locally
(presumably from my bash prompts). Discovered while running tests to
validate another change.

Fixes https://github.com/openai/codex/issues/9413

Test Plan:

Ran test locally on my fedora 43 x86_64 machine with:
```
cd codex/cargo-rs
cargo nextest run --all-features --no-fail-fast unified_exec::tests::unified_exec_timeouts
```

Before, unified_exec_timeouts fails:
```
    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.38s
────────────
 Nextest run ID fa2b4949-a66c-408c-8002-32c52c70ec4f with nextest profile: default
    Starting 1 test across 107 binaries (3211 tests skipped)
        FAIL [   5.667s] codex-core unified_exec::tests::unified_exec_timeouts
  stdout ───

    running 1 test
    test unified_exec::tests::unified_exec_timeouts ... FAILED

    failures:

    failures:
        unified_exec::tests::unified_exec_timeouts

    test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 774 filtered out; finished in 5.66s

  stderr ───

    thread 'unified_exec::tests::unified_exec_timeouts' (459601) panicked at core/src/unified_exec/mod.rs:381:9:
    timeout too short should yield incomplete output
    note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

────────────
     Summary [   5.677s] 1 test run: 0 passed, 1 failed, 3211 skipped
        FAIL [   5.667s] codex-core unified_exec::tests::unified_exec_timeouts
error: test run failed

```

After, works:
```
    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.34s
────────────
 Nextest run ID f49e9004-e30b-4049-b0ff-283b543a1cd7 with nextest profile: default
    Starting 1 test across 107 binaries (3211 tests skipped)
        SLOW [> 15.000s] codex-core unified_exec::tests::unified_exec_timeouts
        PASS [  17.666s] codex-core unified_exec::tests::unified_exec_timeouts
────────────
     Summary [  17.676s] 1 test run: 1 passed (1 slow), 3211 skipped
```
2026-01-17 09:05:53 -08:00
Ahmed Ibrahim
764f3c7d03 fix(tui) Defer backtrack trim until rollback confirms (#9401)
Document the backtrack/rollback state machine and invariants between the
transcript overlay, in-flight “live tail”, and core thread state (tui + tui2).

Also adjust behavior for correctness:
- Track a single pending rollback and block additional rollbacks until core responds.
- Defer trimming transcript cells until ThreadRolledBack for the active session.
- Clear the guard on ThreadRollbackFailed so the user can retry.
- After a confirmed trim, schedule a one-shot scrollback refresh on the next draw.
- Clear stale pending rollback state when switching sessions.

---------

Co-authored-by: Josh McKinney <joshka@openai.com>
2026-01-17 06:29:41 +00:00
Fouad Matin
93a5e0fe1c fix(codex-api): treat invalid_prompt as non-retryable (#9400)
**Goal**: Prevent response.failed events with `invalid_prompt` from
being treated as retryable errors so the UI shows the actual error
message instead of continually retrying.

**Before**: Codex would continue to retry despite the prompt being
marked as disallowed
**After**: Codex will stop retrying once prompt is marked disallowed
2026-01-16 22:22:08 -08:00
Ahmed Ibrahim
146d54cede Add collaboration_mode override to turns (#9408) 2026-01-16 21:51:25 -08:00
xl-openai
ad8bf59cbf Support enable/disable skill via config/api. (#9328)
In config.toml:
```
[[skills.config]]
path = "/Users/xl/.codex/skills/my_skill/SKILL.md"
enabled = false
```

API:
skills/list, skills/config/write
2026-01-16 20:22:05 -08:00
Ahmed Ibrahim
246f506551 Introduce collaboration modes (#9340)
- Merge `model` and `reasoning_effort` under collaboration modes.
- Add additional instructions for custom collaboration mode
- Default to Custom to not change behavior
2026-01-17 00:28:22 +00:00
Anton Panasenko
c26fe64539 feat: show forked from session id in /status (#9330)
Summary:
- Add forked_from to SessionMeta/SessionConfiguredEvent and persist it
for forked sessions.
- Surface forked_from in /status for tui + tui2 and add snapshots.
2026-01-16 13:41:46 -08:00
Owen Lin
f1653dd4d3 feat(app-server, core): return threads by created_at or updated_at (#9247)
Add support for returning threads by either `created_at` OR `updated_at`
descending. Previously core always returned threads ordered by
`created_at`.

This PR:
- updates core to be able to list threads by `updated_at` OR
`created_at` descending based on what the caller wants
- also update `thread/list` in app-server to expose this (default to
`created_at` if not specified)

All existing codepaths (app-server, TUI) still default to `created_at`,
so no behavior change is expected with this PR.

**Implementation**
To sort by `updated_at` is a bit nontrivial (whereas `created_at` is
easy due to the way we structure the folders and filenames on disk,
which are all based on `created_at`).

The most naive way to do this without introducing a cache file or sqlite
DB (which we have to implement/maintain) is to scan files in reverse
`created_at` order on disk, and look at the file's mtime (last modified
timestamp according to the filesystem) until we reach `MAX_SCAN_FILES`
(currently set to 10,000). Then, we can return the most recent N
threads.

Based on some quick and dirty benchmarking on my machine with ~1000
rollout files, calling `thread/list` with limit 50, the `updated_at`
path is slower as expected due to all the I/O:
- updated-at: average 103.10 ms
- created-at: average 41.10 ms

Those absolute numbers aren't a big deal IMO, but we can certainly
optimize this in a followup if needed by introducing more state stored
on disk.

**Caveat**
There's also a limitation in that any files older than `MAX_SCAN_FILES`
will be excluded, which means if a user continues a REALLY old thread,
it's possible to not be included. In practice that should not be too big
of an issue.

If a user makes...
- 1000 rollouts/day → threads older than 10 days won't show up
- 100 rollouts/day → ~100 days

If this becomes a problem for some reason, even more motivation to
implement an updated_at cache.
2026-01-16 20:58:55 +00:00
Anton Panasenko
e893e83eb9 feat: /fork the current session instead of opening session picker (#9385)
Implemented /fork to fork the current session directly (no picker),
handling it via a new ForkCurrentSession app event in both tui and tui2.
Updated slash command descriptions/tooltips and adjusted the fork tests
accordingly. Removed the unused in-session fork picker event.
2026-01-16 11:28:52 -08:00
viyatb-oai
f89a40a849 chore: upgrade to Rust 1.92.0 (#8860)
**Summary**
- Upgrade Rust toolchain used by CI to 1.92.0.
- Address new clippy `derivable_impls` warnings by deriving `Default`
for enums across protocol, core, backend openapi models, and
windows-sandbox setup.
- Tidy up related test/config behavior (originator header handling, env
override cleanup) and remove a now-unused assignment in TUI/TUI2 render
layout.

**Testing**
- `just fmt`
- `just fix -p codex-tui`
- `just fix -p codex-tui2`
- `just fix -p codex-windows-sandbox`
- `cargo test -p codex-tui`
- `cargo test -p codex-tui2`
- `cargo test -p codex-windows-sandbox`
- `cargo test -p codex-core --test all`
- `cargo test -p codex-app-server --test all`
- `cargo test -p codex-mcp-server --test all`
- `cargo test --all-features`
2026-01-16 11:12:52 -08:00
jif-oai
e650d4b02c feat: tool call duration metric (#9364) 2026-01-16 18:33:14 +01:00
Ahmed Ibrahim
ebdd8795e9 Turn-state sticky routing per turn (#9332)
- capture the header from SSE/WS handshakes, store it per
ModelClientSession using `Oncelock`, echo it on turn-scoped requests,
and add SSE+WS integration tests for within-turn persistence +
cross-turn reset.

- keep `x-codex-turn-state` sticky within a user turn to maintain
routing continuity for retries/tool follow-ups.
2026-01-16 09:30:11 -08:00
Jeremy Rose
4125c825f9 add codex cloud list (#9324)
for listing cloud tasks.
2026-01-16 08:56:38 -08:00
Eric Traut
9147df0e60 Made codex exec resume --last consistent with codex resume --last (#9352)
PR #9245 made `codex resume --last` honor cwd, but I forgot to make the
same change for `codex exec resume --last`. This PR fixes the
inconsistency.

This addresses #8700
2026-01-16 08:53:47 -08:00
Matthew Zeng
131590066e [device-auth] Add device code auth as a standalone option when headless environment is detected. (#9333) 2026-01-16 08:35:03 -08:00
jif-oai
2691e1ce21 fix: flaky tests (#9373) 2026-01-16 17:24:41 +01:00
jif-oai
1668ca726f chore: close pipe on non-pty processes (#9369)
Closing the STDIN of piped process when starting them to avoid commands
like `rg` to wait for content on STDIN and hangs for ever
2026-01-16 15:54:32 +01:00
jif-oai
7905e99d03 prompt collab (#9367) 2026-01-16 15:12:41 +01:00
jif-oai
7fc49697dd feat: CODEX_CI (#9366) 2026-01-16 13:52:16 +00:00
jif-oai
c576756c81 feat: collab wait multiple IDs (#9294) 2026-01-16 12:05:04 +01:00
jif-oai
c1ac5223e1 feat: run user commands under user snapshot (#9357)
The initial goal is for user snapshots to have access to aliases etc
2026-01-16 11:49:28 +01:00
jif-oai
f5b3e738fb feat: propagate approval request of unsubscribed threads (#9232)
A thread can now be spawned by another thread. In order to process the
approval requests of such sub-threads, we need to detect those event and
show them in the TUI.

This is a temporary solution while the UX is being figured out. This PR
should be reverted once done
2026-01-16 11:23:01 +01:00
Ahmed Ibrahim
0cce6ebd83 rename model turn to sampling request (#9336)
We have two type of turns now: model and user turns. It's always
confusing to refer to either. Model turn is basically a sampling
request.
2026-01-16 10:06:24 +01:00
Eric Traut
1fc72c647f Fix token estimate during compaction (#9337)
This addresses #9287
2026-01-15 19:48:11 -08:00
Michael Bolin
99f47d6e9a fix(mcp): include threadId in both content and structuredContent in CallToolResult (#9338) 2026-01-15 18:33:11 -08:00
Thanh Nguyen
a6324ab34b fix(tui): only show 'Worked for' separator when actual work was performed (#8958)
Fixes #7919.

This PR addresses a TUI display bug where the "Worked for" separator
would appear prematurely during the planning stage.

**Changes:**
- Added `had_work_activity` flag to `ChatWidget` to track if actual work
(exec commands, MCP tool calls, patches) was performed in the current
turn.
- Updated `handle_streaming_delta` to only display the
`FinalMessageSeparator` if both `needs_final_message_separator` AND
`had_work_activity` are true.
- Updated `handle_exec_end_now`, `handle_patch_apply_end_now`, and
`handle_mcp_end_now` to set `had_work_activity = true`.

**Verification:**
- Ran `cargo test -p codex-tui` to ensure no regressions.
- Manual verification confirms the separator now only appears after
actual work is completed.

---------

Co-authored-by: Josh McKinney <joshka@openai.com>
2026-01-16 01:41:43 +00:00
Dylan Hurd
3cabb24210 chore(windows) Enable Powershell UTF8 feature (#9195)
## Summary
We've received a lot of positive feedback about this feature, so we're
going to enable it by default.
2026-01-16 01:29:12 +00:00
charley-oai
1fa8350ae7 Add text element metadata to protocol, app server, and core (#9331)
The second part of breaking up PR
https://github.com/openai/codex/pull/9116

Summary:

- Add `TextElement` / `ByteRange` to protocol user inputs and user
message events with defaults.
- Thread `text_elements` through app-server v1/v2 request handling and
history rebuild.
- Preserve UI metadata only in user input/events (not `ContentItem`)
while keeping local image attachments in user events for rehydration.

Details:

- Protocol: `UserInput::Text` carries `text_elements`;
`UserMessageEvent` carries `text_elements` + `local_images`.
Serialization includes empty vectors for backward compatibility.
- app-server-protocol: v1 defines `V1TextElement` / `V1ByteRange` in
camelCase with conversions; v2 uses its own camelCase wrapper.
- app-server: v1/v2 input mapping includes `text_elements`; thread
history rebuilds include them.
- Core: user event emission preserves UI metadata while model history
stays clean; history replay round-trips the metadata.
2026-01-15 17:26:41 -08:00
Yuvraj Angad Singh
004a74940a fix: send non-null content on elicitation Accept (#9196)
## Summary

- When a user accepts an MCP elicitation request, send `content:
Some(json!({}))` instead of `None`
- MCP servers that use elicitation expect content to be present when
action is Accept
- This matches the expected behavior shown in tests at
`exec-server/tests/common/lib.rs:171`

## Root Cause

In `codex-rs/core/src/codex.rs`, the `resolve_elicitation` function
always sent `content: None`:

```rust
let response = ElicitationResponse {
    action,
    content: None,  // Always None, even for Accept
};
```

## Fix

Send an empty object when accepting:

```rust
let content = match action {
    ElicitationAction::Accept => Some(serde_json::json!({})),
    ElicitationAction::Decline | ElicitationAction::Cancel => None,
};
```

## Test plan

- [x] Code compiles with `cargo check -p codex-core`
- [x] Formatted with `just fmt`
- [ ] Integration test `accept_elicitation_for_prompt_rule` (requires
MCP server binary)

Fixes #9053
2026-01-15 14:20:57 -08:00
Ahmed Ibrahim
749b58366c Revert empty paste image handling (#9318)
Revert #9049 behavior so empty paste events no longer trigger a
clipboard image read.
2026-01-15 14:16:09 -08:00
pap-openai
d886a8646c remove needs_follow_up error log (#9272) 2026-01-15 21:20:54 +00:00
sayan-oai
169201b1b5 [search] allow explicitly disabling web search (#9249)
moving `web_search` rollout serverside, so need a way to explicitly
disable search + signal eligibility from the client.

- Add `x‑oai‑web‑search‑eligible` header that signifies whether the
request can have web search.
- Only attach the `web_search` tool when the resolved `WebSearchMode` is
`Live` or `Cached`.
2026-01-15 11:28:57 -08:00
xl-openai
42fa4c237f Support SKILL.toml file. (#9125)
We’re introducing a new SKILL.toml to hold skill metadata so Codex can
deliver a richer Skills experience.

Initial focus is the interface block:
```
[interface]
display_name = "Optional user-facing name"
short_description = "Optional user-facing description"
icon_small = "./assets/small-400px.png"
icon_large = "./assets/large-logo.svg"
brand_color = "#3B82F6"
default_prompt = "Optional surrounding prompt to use the skill with"
```

All fields are exposed via the app server API.
display_name and short_description are consumed by the TUI.
2026-01-15 11:20:04 -08:00
Eric Traut
5f10548772 Revert recent styling change for input prompt placeholder text (#9307)
A recent change in commit ccba737d26 modified the styling of the
placeholder text (e.g. "Implement {feature}") in the input box of the
CLI, changing it from non-italic to italic. I think this was likely
unintentional. It results in a bad display appearance on some terminal
emulators, and several users have complained about it.

This change switches back to non-italic styling, restoring the older
behavior.

It addresses #9262
2026-01-15 10:58:12 -08:00
jif-oai
da44569fef nit: clean unified exec background processes (#9304)
To fix the occurences where the End event is received after the listener
stopped listenning
2026-01-15 18:34:33 +00:00
jif-oai
393a5a0311 chore: better orchestrator prompt (#9301) 2026-01-15 18:11:43 +00:00
viyatb-oai
55bda1a0f2 revert: remove pre-Landlock bind mounts apply (#9300)
**Description**

This removes the pre‑Landlock read‑only bind‑mount step from the Linux
sandbox so filesystem restrictions rely solely on Landlock again.
`mounts.rs` is kept in place but left unused. The linux‑sandbox README
is updated to match the new behavior and manual test expectations.
2026-01-15 09:47:57 -08:00
李琼羽
b4d240c3ae fix(exec): improve stdin prompt decoding (#9151)
Fixes #8733.

- Read prompt from stdin as raw bytes and decode more helpfully.
- Strip UTF-8 BOM; decode UTF-16LE/UTF-16BE when a BOM is present.
- For other non-UTF8 input, fail with an actionable message (offset +
iconv hint).

Tests: `cargo test -p codex-exec`.
2026-01-15 09:29:05 -08:00
gt-oai
f6df1596eb Propagate MCP disabled reason (#9207)
Indicate why MCP servers are disabled when they are disabled by
requirements:

```
➜  codex git:(main) ✗ just codex mcp list
cargo run --bin codex -- "$@"
    Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.27s
     Running `target/debug/codex mcp list`
Name         Command          Args  Env  Cwd  Status                                                                  Auth
docs         docs-mcp         -     -    -    disabled: requirements (MDM com.openai.codex:requirements_toml_base64)  Unsupported
hello_world  hello-world-mcp  -     -    -    disabled: requirements (MDM com.openai.codex:requirements_toml_base64)  Unsupported

➜  codex git:(main) ✗ just c
cargo run --bin codex -- "$@"
    Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.90s
     Running `target/debug/codex`
╭─────────────────────────────────────────────╮
│ >_ OpenAI Codex (v0.0.0)                    │
│                                             │
│ model:     gpt-5.2 xhigh   /model to change │
│ directory: ~/code/codex/codex-rs            │
╰─────────────────────────────────────────────╯

/mcp

🔌  MCP Tools

  • No MCP tools available.

  • docs (disabled)
    • Reason: requirements (MDM com.openai.codex:requirements_toml_base64)

  • hello_world (disabled)
    • Reason: requirements (MDM com.openai.codex:requirements_toml_base64)
```
2026-01-15 17:24:00 +00:00
Eric Traut
ae96a15312 Changed codex resume --last to honor the current cwd (#9245)
This PR changes `codex resume --last` to work consistently with `codex
resume`. Namely, it filters based on the cwd when selecting the last
session. It also supports the `--all` modifier as an override.

This addresses #8700
2026-01-15 17:05:08 +00:00
jif-oai
3fc487e0e0 feat: basic tui for event emission (#9209) 2026-01-15 15:53:02 +00:00
jif-oai
faeb08c1e1 feat: add interrupt capabilities to send_input (#9276) 2026-01-15 14:59:07 +00:00
jif-oai
05b960671d feat: add agent roles to collab tools (#9275)
Add `agent_type` parameter to the collab tool `spawn_agent` that
contains a preset to apply on the config when spawning this agent
2026-01-15 13:33:52 +00:00
jif-oai
bad4c12b9d feat: collab tools app-server event mapping (#9213) 2026-01-15 09:03:26 +00:00
viyatb-oai
2259031d64 fix: fallback to Landlock-only when user namespaces unavailable and set PR_SET_NO_NEW_PRIVS early (#9250)
fixes https://github.com/openai/codex/issues/9236

### Motivation
- Prevent sandbox setup from failing when unprivileged user namespaces
are denied so Landlock-only protections can still be applied.
- Ensure `PR_SET_NO_NEW_PRIVS` is set before installing seccomp and
Landlock restrictions to avoid kernel `EPERM`/`LandlockRestrict`
ordering issues.

### Description
- Add `is_permission_denied` helper that detects `EPERM` /
`PermissionDenied` from `CodexErr` to drive fallback logic.
- In `apply_read_only_mounts` skip read-only bind-mount setup and return
`Ok(())` when `unshare_user_and_mount_namespaces()` fails with
permission-denied so Landlock rules can still be installed.
- Add `set_no_new_privs()` and call it from
`apply_sandbox_policy_to_current_thread` before installing seccomp
filters and Landlock rules when disk or network access is restricted.
2026-01-14 22:24:34 -08:00
Ahmed Ibrahim
a09711332a Add migration_markdown in model_info (#9219)
Next step would be to clean Model Upgrade in model presets

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: aibrahim-oai <219906144+aibrahim-oai@users.noreply.github.com>
2026-01-15 01:55:22 +00:00
charley-oai
4a9c2bcc5a Add text element metadata to types (#9235)
Initial type tweaking PR to make the diff of
https://github.com/openai/codex/pull/9116 smaller

This should not change any behavior, just adds some fields to types
2026-01-14 16:41:50 -08:00
Michael Bolin
2a68b74b9b fix: increase timeout for release builds from 30 to 60 minutes (#9242)
Windows builds have been tripping the 30 minute timeout. For sure, we
need to improve this, but as a quick fix, let's just increase the
timeout.

Perhaps we should switch to `lto = "thin"` for release builds, at least
for Windows:


3728db11b8/codex-rs/Cargo.toml (L288)

See https://doc.rust-lang.org/cargo/reference/profiles.html#lto for
details.
2026-01-15 00:38:25 +00:00
Michael Bolin
3728db11b8 fix: eliminate unnecessary clone() for each SSE event (#9238)
Given how many SSE events we get, seems worth fixing.
2026-01-15 00:06:09 +00:00
viyatb-oai
e59e7d163d fix: correct linux sandbox uid/gid mapping after unshare (#9234)
fixes https://github.com/openai/codex/issues/9233
## Summary
- capture effective uid/gid before unshare for user namespace maps
- pass captured ids into uid/gid map writer

## Testing
- just fmt
- just fix -p codex-linux-sandbox
- cargo test -p codex-linux-sandbox
2026-01-14 15:35:53 -08:00
willwang-openai
71a2973fd9 upgrade runners in rust-ci.yml to use the larger runners (#9106)
Upgrades runners in rust-ci.yaml to larger runners

ubuntu-24.04 (x64 and arm64) -> custom 16 core ubuntu 24.04 runners
macos-14 -> mac0s-15-xlarge
[TODO] windows (x64 and arm64) -> custom 16 core windows runners
2026-01-14 15:22:59 -08:00
Eric Traut
24b88890cb Updated issue template to ask for terminal emulator (#9231)
Co-authored-by: Josh McKinney <joshka@openai.com>
2026-01-14 23:00:15 +00:00
sayan-oai
5e426ac270 add WebSearchMode enum (#9216)
### What
Add `WebSearchMode` enum (disabled, cached live, defaults to cached) to
config + V2 protocol. This enum takes precedence over legacy flags:
`web_search_cached`, `web_search_request`, and `tools.web_search`.

Keep `--search` as live.

### Tests
Added tests
2026-01-14 12:51:42 -08:00
Josh McKinney
27da8a68d3 fix(tui): disable double-press quit shortcut (#9220)
Disables the default Ctrl+C/Ctrl+D double-press quit UX (keeps the code
path behind a const) while we rethink the quit/interrupt flow.

Tests:
- just fmt
- cargo clippy --fix --all-features --tests --allow-dirty --allow-no-vcs
-p codex-tui
- cargo test -p codex-tui --lib
2026-01-14 20:28:18 +00:00
Josh McKinney
0471ddbe74 fix(tui2): align Steer submit keys (#9218)
- Remove legacy Ctrl+K queuing in tui2; Tab is the queue key.
- Make Enter queue when Steer is disabled and submit immediately when
Steer is enabled.
- Add Steer keybinding docs on both tui and tui2 chat composers.
2026-01-14 19:35:55 +00:00
pakrym-oai
e6d2ef432d Rename hierarchical_agents to child_agents_md (#9215)
Clearer name
2026-01-14 19:14:24 +00:00
jif-oai
577e1fd1b2 feat: adding piped process to replace PTY when needed (#8797) 2026-01-14 18:44:04 +00:00
gt-oai
fe1e0da102 s/mcp_server_requirements/mcp_servers (#9212)
A simple `s/mcp_server_requirements/mcp_servers/g` for an unreleased
feature. @bolinfest correctly pointed out, it's already in
`requirements.toml` so the `_requirements` is redundant.
2026-01-14 18:41:52 +00:00
pakrym-oai
e958d0337e Log headers in trace mode (#9214)
To enable:

```
export RUST_LOG="warn,codex_=trace"
```

Sample: 
```
Request completed method=POST url=https://chatgpt.com/backend-api/codex/responses status=200 OK headers={"date": "Wed, 14 Jan 2026 18:21:21 GMT", "transfer-encoding": "chunked", "connection": "keep-alive", "x-codex-plan-type": "business", "x-codex-primary-used-percent": "3", "x-codex-secondary-used-percent": "6", "x-codex-primary-window-minutes": "300", "x-codex-primary-over-secondary-limit-percent": "0", "x-codex-secondary-window-minutes": "10080", "x-codex-primary-reset-after-seconds": "9944", "x-codex-secondary-reset-after-seconds": "171121", "x-codex-primary-reset-at": "1768424824", "x-codex-secondary-reset-at": "1768586001", "x-codex-credits-has-credits": "False", "x-codex-credits-balance": "", "x-codex-credits-unlimited": "False", "x-models-etag": "W/\"7a7ffbc83c159dbd7a2a73aaa9c91b7a\"", "x-oai-request-id": "ffedcd30-6d8a-4c4d-be10-8ebb23c142c8", "x-envoy-upstream-service-time": "417", "x-openai-proxy-wasm": "v0.1", "cf-cache-status": "DYNAMIC", "set-cookie": "__cf_bm=xFKeaMbWNbKO5ZX.K5cJBhj34OA1QvnF_3nkdMThjlA-1768414881-1.0.1.1-uLpsE_BDkUfcmOMaeKVQmv_6_2ytnh_R3lO_il5N5K3YPQEkBo0cOMTdma6bK0Gz.hQYcIesFwKIJht1kZ9JKqAYYnjgB96hF4.sii2U3cE; path=/; expires=Wed, 14-Jan-26 18:51:21 GMT; domain=.chatgpt.com; HttpOnly; Secure; SameSite=None", "report-to": "{\"endpoints\":[{\"url\":\"https:\/\/a.nel.cloudflare.com\/report\/v4?s=4Kc7g4zUhKkIm3xHuB6ba4jyIUqqZ07ETwIPAYQASikRjA8JesbtUKDP9tSrZ5PnzWldaiSz5dZVQFI579LEsCMlMUSelTvmyQ8j4FbFDawi%2FprWZ5iRePiaSalr\"}],\"group\":\"cf-nel\",\"max_age\":604800}", "nel": "{\"success_fraction\":0.01,\"report_to\":\"cf-nel\",\"max_age\":604800}", "strict-transport-security": "max-age=31536000; includeSubDomains; preload", "x-content-type-options": "nosniff", "cross-origin-opener-policy": "same-origin-allow-popups", "referrer-policy": "strict-origin-when-cross-origin", "server": "cloudflare", "cf-ray": "9bdf270adc7aba3a-SEA"} version=HTTP/1.1
```
2026-01-14 18:38:12 +00:00
Ahmed Ibrahim
8e937fbba9 Get model on session configured (#9191)
- Don't try to precompute model unless you know it from `config`
- Block `/model` on session configured
- Queue messages until session configured
- show "loading" in status until session configured
2026-01-14 10:20:41 -08:00
Celia Chen
02f67bace8 fix: Emit response.completed immediately for Responses SSE (#9170)
we see windows test failures like this:
https://github.com/openai/codex/actions/runs/20930055601/job/60138344260.

The issue is that SSE connections sometimes remain open after the
completion event esp. for windows. We should emit the completion event
and return immediately. this is consistent with the protocol:

> The Model streams responses back in an SSE, which are collected until
"completed" message and the SSE terminates

from
https://github.com/openai/codex/blob/dev/cc/fix-windows-test/codex-rs/docs/protocol_v1.md#L37.

this helps us achieve parity with responses websocket logic here:
https://github.com/openai/codex/blob/dev/cc/fix-windows-test/codex-rs/codex-api/src/endpoint/responses_websocket.rs#L220-L227.
2026-01-14 10:05:00 -08:00
jif-oai
3d322fa9d8 feat: add collab prompt (#9208)
Adding a prompt for collab tools. This is only for internal use and the
prompt won't be gated for now as it is not stable yet.

The goal of this PR is to provide the tool required to iterate on the
prompt
2026-01-14 09:58:15 -08:00
jif-oai
6a939ed7a4 feat: emit events around collab tools (#9095)
Emit the following events around the collab tools. On the `app-server`
this will be under `item/started` and `item/completed`
```
#[derive(Debug, Clone, Deserialize, Serialize, PartialEq, JsonSchema, TS)]
pub struct CollabAgentSpawnBeginEvent {
    /// Identifier for the collab tool call.
    pub call_id: String,
    /// Thread ID of the sender.
    pub sender_thread_id: ThreadId,
    /// Initial prompt sent to the agent. Can be empty to prevent CoT leaking at the
    /// beginning.
    pub prompt: String,
}

#[derive(Debug, Clone, Deserialize, Serialize, PartialEq, JsonSchema, TS)]
pub struct CollabAgentSpawnEndEvent {
    /// Identifier for the collab tool call.
    pub call_id: String,
    /// Thread ID of the sender.
    pub sender_thread_id: ThreadId,
    /// Thread ID of the newly spawned agent, if it was created.
    pub new_thread_id: Option<ThreadId>,
    /// Initial prompt sent to the agent. Can be empty to prevent CoT leaking at the
    /// beginning.
    pub prompt: String,
    /// Last known status of the new agent reported to the sender agent.
    pub status: AgentStatus,
}

#[derive(Debug, Clone, Deserialize, Serialize, PartialEq, JsonSchema, TS)]
pub struct CollabAgentInteractionBeginEvent {
    /// Identifier for the collab tool call.
    pub call_id: String,
    /// Thread ID of the sender.
    pub sender_thread_id: ThreadId,
    /// Thread ID of the receiver.
    pub receiver_thread_id: ThreadId,
    /// Prompt sent from the sender to the receiver. Can be empty to prevent CoT
    /// leaking at the beginning.
    pub prompt: String,
}

#[derive(Debug, Clone, Deserialize, Serialize, PartialEq, JsonSchema, TS)]
pub struct CollabAgentInteractionEndEvent {
    /// Identifier for the collab tool call.
    pub call_id: String,
    /// Thread ID of the sender.
    pub sender_thread_id: ThreadId,
    /// Thread ID of the receiver.
    pub receiver_thread_id: ThreadId,
    /// Prompt sent from the sender to the receiver. Can be empty to prevent CoT
    /// leaking at the beginning.
    pub prompt: String,
    /// Last known status of the receiver agent reported to the sender agent.
    pub status: AgentStatus,
}

#[derive(Debug, Clone, Deserialize, Serialize, PartialEq, JsonSchema, TS)]
pub struct CollabWaitingBeginEvent {
    /// Thread ID of the sender.
    pub sender_thread_id: ThreadId,
    /// Thread ID of the receiver.
    pub receiver_thread_id: ThreadId,
    /// ID of the waiting call.
    pub call_id: String,
}

#[derive(Debug, Clone, Deserialize, Serialize, PartialEq, JsonSchema, TS)]
pub struct CollabWaitingEndEvent {
    /// Thread ID of the sender.
    pub sender_thread_id: ThreadId,
    /// Thread ID of the receiver.
    pub receiver_thread_id: ThreadId,
    /// ID of the waiting call.
    pub call_id: String,
    /// Last known status of the receiver agent reported to the sender agent.
    pub status: AgentStatus,
}

#[derive(Debug, Clone, Deserialize, Serialize, PartialEq, JsonSchema, TS)]
pub struct CollabCloseBeginEvent {
    /// Identifier for the collab tool call.
    pub call_id: String,
    /// Thread ID of the sender.
    pub sender_thread_id: ThreadId,
    /// Thread ID of the receiver.
    pub receiver_thread_id: ThreadId,
}

#[derive(Debug, Clone, Deserialize, Serialize, PartialEq, JsonSchema, TS)]
pub struct CollabCloseEndEvent {
    /// Identifier for the collab tool call.
    pub call_id: String,
    /// Thread ID of the sender.
    pub sender_thread_id: ThreadId,
    /// Thread ID of the receiver.
    pub receiver_thread_id: ThreadId,
    /// Last known status of the receiver agent reported to the sender agent before
    /// the close.
    pub status: AgentStatus,
}
```
2026-01-14 17:55:57 +00:00
Josh McKinney
4283a7432b tui: double-press Ctrl+C/Ctrl+D to quit (#8936)
## Problem

Codex’s TUI quit behavior has historically been easy to trigger
accidentally and hard to reason
about.

- `Ctrl+C`/`Ctrl+D` could terminate the UI immediately, which is a
common key to press while trying
  to dismiss a modal, cancel a command, or recover from a stuck state.
- “Quit” and “shutdown” were not consistently separated, so some exit
paths could bypass the
  shutdown/cleanup work that should run before the process terminates.

This PR makes quitting both safer (harder to do by accident) and more
uniform across quit
gestures, while keeping the shutdown-first semantics explicit.

## Mental model

After this change, the system treats quitting as a UI request that is
coordinated by the app
layer.

- The UI requests exit via `AppEvent::Exit(ExitMode)`.
- `ExitMode::ShutdownFirst` is the normal user path: the app triggers
`Op::Shutdown`, continues
rendering while shutdown runs, and only ends the UI loop once shutdown
has completed.
- `ExitMode::Immediate` exists as an escape hatch (and as the
post-shutdown “now actually exit”
signal); it bypasses cleanup and should not be the default for
user-triggered quits.

User-facing quit gestures are intentionally “two-step” for safety:

- `Ctrl+C` and `Ctrl+D` no longer exit immediately.
- The first press arms a 1-second window and shows a footer hint (“ctrl
+ <key> again to quit”).
- Pressing the same key again within the window requests a
shutdown-first quit; otherwise the
  hint expires and the next press starts a fresh window.

Key routing remains modal-first:

- A modal/popup gets first chance to consume `Ctrl+C`.
- If a modal handles `Ctrl+C`, any armed quit shortcut is cleared so
dismissing a modal cannot
  prime a subsequent `Ctrl+C` to quit.
- `Ctrl+D` only participates in quitting when the composer is empty and
no modal/popup is active.

The design doc `docs/exit-confirmation-prompt-design.md` captures the
intended routing and the
invariants the UI should maintain.

## Non-goals

- This does not attempt to redesign modal UX or make modals uniformly
dismissible via `Ctrl+C`.
It only ensures modals get priority and that quit arming does not leak
across modal handling.
- This does not introduce a persistent confirmation prompt/menu for
quitting; the goal is to keep
  the exit gesture lightweight and consistent.
- This does not change the semantics of core shutdown itself; it changes
how the UI requests and
  sequences it.

## Tradeoffs

- Quitting via `Ctrl+C`/`Ctrl+D` now requires a deliberate second
keypress, which adds friction for
  users who relied on the old “instant quit” behavior.
- The UI now maintains a small time-bounded state machine for the armed
shortcut, which increases
  complexity and introduces timing-dependent behavior.

This design was chosen over alternatives (a modal confirmation prompt or
a long-lived “are you
sure” state) because it provides an explicit safety barrier while
keeping the flow fast and
keyboard-native.

## Architecture

- `ChatWidget` owns the quit-shortcut state machine and decides when a
quit gesture is allowed
  (idle vs cancellable work, composer state, etc.).
- `BottomPane` owns rendering and local input routing for modals/popups.
It is responsible for
consuming cancellation keys when a view is active and for
showing/expiring the footer hint.
- `App` owns shutdown sequencing: translating
`AppEvent::Exit(ShutdownFirst)` into `Op::Shutdown`
  and only terminating the UI loop when exit is safe.

This keeps “what should happen” decisions (quit vs interrupt vs ignore)
in the chat/widget layer,
while keeping “how it looks and which view gets the key” in the
bottom-pane layer.

## Observability

You can tell this is working by running the TUIs and exercising the quit
gestures:

- While idle: pressing `Ctrl+C` (or `Ctrl+D` with an empty composer and
no modal) shows a footer
hint for ~1 second; pressing again within that window exits via
shutdown-first.
- While streaming/tools/review are active: `Ctrl+C` interrupts work
rather than quitting.
- With a modal/popup open: `Ctrl+C` dismisses/handles the modal (if it
chooses to) and does not
arm a quit shortcut; a subsequent quick `Ctrl+C` should not quit unless
the user re-arms it.

Failure modes are visible as:

- Quits that happen immediately (no hint window) from `Ctrl+C`/`Ctrl+D`.
- Quits that occur while a modal is open and consuming `Ctrl+C`.
- UI termination before shutdown completes (cleanup skipped).

## Tests

- Updated/added unit and snapshot coverage in `codex-tui` and
`codex-tui2` to validate:
  - The quit hint appears and expires on the expected key.
- Double-press within the window triggers a shutdown-first quit request.
- Modal-first routing prevents quit bypass and clears any armed shortcut
when a modal consumes
    `Ctrl+C`.

These tests focus on the UI-level invariants and rendered output; they
do not attempt to validate
real terminal key-repeat timing or end-to-end process shutdown behavior.

---
Screenshot:
<img width="912" height="740" alt="Screenshot 2026-01-13 at 1 05 28 PM"
src="https://github.com/user-attachments/assets/18f3d22e-2557-47f2-a369-ae7a9531f29f"
/>
2026-01-14 17:42:52 +00:00
pakrym-oai
92472e7baa Use current model for review (#9179)
Instead of having a hard-coded default review model, use the current
model for running `/review` unless one is specified in the config.

Also inherit current reasoning effort
2026-01-14 08:59:41 -08:00
viyatb-oai
e1447c3009 feat: add support for read-only bind mounts in the linux sandbox (#9112)
### Motivation

- Landlock alone cannot prevent writes to sensitive in-repo files like
`.git/` when the repo root is writable, so explicit mount restrictions
are required for those paths.
- The sandbox must set up any mounts before calling Landlock so Landlock
can still be applied afterwards and the two mechanisms compose
correctly.

### Description

- Add a new `linux-sandbox` helper `apply_read_only_mounts` in
`linux-sandbox/src/mounts.rs` that: unshares namespaces, maps uids/gids
when required, makes mounts private, bind-mounts targets, and remounts
them read-only.
- Wire the mount step into the sandbox flow by calling
`apply_read_only_mounts(...)` before network/seccomp and before applying
Landlock rules in `linux-sandbox/src/landlock.rs`.
2026-01-14 08:30:46 -08:00
jif-oai
bcd7858ced feat: add auto refresh on thread listeners (#9105)
This PR is in the scope of multi-agent work. 

An agent (=thread) can now spawn other agents. Those other agents are
not attached to any clients. We need a way to make sure that the clients
are aware of the new threads to look at (for approval for example). This
PR adds a channel to the `ThreadManager` that pushes the ID of those
newly created agents such that the client (here the app-server) can also
subscribe to those ones.
2026-01-14 16:26:01 +00:00
jif-oai
32b1795ff4 chore: clamp min yield time for empty write_stdin (#9156)
After evals, 0 impact on performance
2026-01-14 16:25:40 +00:00
Ahmed Ibrahim
bdae0035ec Render exec output deltas inline (#9194) 2026-01-14 08:11:12 -08:00
jif-oai
7532f34699 fix: drop double waiting header in TUI (#9145) 2026-01-14 09:52:34 +00:00
jif-oai
bc6d9ef6fc feat: only source shell snapshot if the file exists (#9197) 2026-01-14 01:28:25 -08:00
jif-oai
dc3deaa3e7 feat: return an error if the image sent by the user is a bad image (#9146)
## Before
When we detect an `InvalidImageRequest`, we replace the image by a
placeholder and keep going

## Now
In such `InvalidImageRequest`, we check if the image is due to a user
message or a tool call output. For tool call output we still replace it
with a placeholder to avoid breaking the agentic loop bu tif this is
because of a user message, we send an error to the user
2026-01-14 09:07:45 +00:00
jif-oai
6fbb89e858 fix: shell snapshot clean-up (#9155)
Clean all shell snapshot files corresponding to sessions that have not
been updated in 7 days
Those files should never leak. The only known cases were it can leak are
during non graceful interrupt of the process (`kill -9, `panic`, OS
crash, ...)
2026-01-14 09:05:46 +00:00
jif-oai
258fc4b401 feat: add sourcing of rc files to shell snapshot (#9150) 2026-01-14 08:58:10 +00:00
Ahmed Ibrahim
b9ff4ec830 change api default model (#9188) 2026-01-13 22:33:34 -08:00
Michael Bolin
0c09dc3c03 feat: add threadId to MCP server messages (#9192)
This favors `threadId` instead of `conversationId` so we use the same
terms as https://developers.openai.com/codex/sdk/.

To test the local build:

```
cd codex-rs
cargo build --bin codex
npx -y @modelcontextprotocol/inspector ./target/debug/codex mcp-server
```

I sent:

```json
{
  "method": "tools/call",
  "params": {
    "name": "codex",
    "arguments": {
      "prompt": "favorite ls option?"
    },
    "_meta": {
      "progressToken": 0
    }
  }
}
```

and got:

```json
{
  "content": [
    {
      "type": "text",
      "text": "`ls -lah` (or `ls -alh`) — long listing, includes dotfiles, human-readable sizes."
    }
  ],
  "structuredContent": {
    "threadId": "019bbb20-bff6-7130-83aa-bf45ab33250e"
  }
}
```

and successfully used the `threadId` in the follow-up with the
`codex-reply` tool call:

```json
{
  "method": "tools/call",
  "params": {
    "name": "codex-reply",
    "arguments": {
      "prompt": "what is the long versoin",
      "threadId": "019bbb20-bff6-7130-83aa-bf45ab33250e"
    },
    "_meta": {
      "progressToken": 1
    }
  }
}
```

whose response also has the `threadId`:

```json
{
  "content": [
    {
      "type": "text",
      "text": "Long listing is `ls -l` (adds permissions, owner/group, size, timestamp)."
    }
  ],
  "structuredContent": {
    "threadId": "019bbb20-bff6-7130-83aa-bf45ab33250e"
  }
}
```

Fixes https://github.com/openai/codex/issues/3712.
2026-01-13 22:14:41 -08:00
sayan-oai
5675af5190 chore: clarify default shell for unified_exec (#8997)
The description of the `shell` arg for `exec_command` states the default
is `/bin/bash`, but AFAICT it's the user's default shell.

Default logic
[here](2a06d64bc9/codex-rs/core/src/tools/handlers/unified_exec.rs (L123)).

EDIT: #9004 has an alternative where we inform the model of the default
shell itself.
2026-01-13 21:37:41 -08:00
Eric Traut
31d9b6f4d2 Improve handling of config and rules errors for app server clients (#9182)
When an invalid config.toml key or value is detected, the CLI currently
just quits. This leaves the VSCE in a dead state.

This PR changes the behavior to not quit and bubble up the config error
to users to make it actionable. It also surfaces errors related to
"rules" parsing.

This allows us to surface these errors to users in the VSCE, like this:

<img width="342" height="129" alt="Screenshot 2026-01-13 at 4 29 22 PM"
src="https://github.com/user-attachments/assets/a79ffbe7-7604-400c-a304-c5165b6eebc4"
/>

<img width="346" height="244" alt="Screenshot 2026-01-13 at 4 45 06 PM"
src="https://github.com/user-attachments/assets/de874f7c-16a2-4a95-8c6d-15f10482e67b"
/>
2026-01-13 17:57:09 -08:00
Ahmed Ibrahim
5a82a72d93 Use offline cache for tui migrations (#9186) 2026-01-13 17:44:12 -08:00
Josh McKinney
ce49e92848 fix(tui): harden paste-burst state transitions (#9124)
User-facing symptom: On terminals that deliver pastes as rapid
KeyCode::Char/Enter streams (notably Windows), paste-burst transient
state
can leak into the next input. Users can see Enter insert a newline when
they meant to submit, or see characters appear late / handled through
the
wrong path.

System problem: PasteBurst is time-based. Clearing only the
classification window (e.g. via clear_window_after_non_char()) can erase
last_plain_char_time without emitting buffered text. If a buffer is
still
non-empty after that, flush_if_due() no longer has a timeout clock to
flush against, so the buffer can get "stuck" until another plain char
arrives.

This was surfaced while adding deterministic regression tests for
paste-burst behavior.

Fix: when disabling burst detection, defuse any in-flight burst state:
flush held/buffered text through handle_paste() (so it follows normal
paste integration), then clear timing and Enter suppression.

Document the rationale inline and update docs/tui-chat-composer.md so
"disable_paste_burst" matches the actual behavior.
2026-01-14 01:42:21 +00:00
Ahmed Ibrahim
4d787a2cc2 Renew cache ttl on etag match (#9174)
so we don't do unnecessary fetches
2026-01-14 01:21:41 +00:00
Ahmed Ibrahim
c96c26cf5b [CODEX-4427] improve parsed commands (#8933)
**make command summaries more accurate by distinguishing
list/search/read operations across common CLI tools.**
- Added parsing helpers to centralize operand/flag handling (cd_target,
sed_read_path, first_non_flag_operand, single_non_flag_operand,
parse_grep_like, awk_data_file_operand, python_walks_files,
is_python_command) and reused them in summarize_main_tokens/shell
parsing.
- Newly parsed list-files commands: git ls-files, rg --files (incl.
rga/ripgrep-all), eza/exa, tree, du, python -c file-walks, plus fd/find
map to ListFiles when no query.
- Newly parsed search commands: git grep, grep/egrep/fgrep, ag/ack/pt,
rg/rga files-with-matches flags (-l/-L, --files-with-matches,
--files-without-match), with improved flag skipping to avoid
misclassifying args as paths.
- Newly parsed read commands: bat/batcat, less, more, awk <file>, and
more flexible sed -n range + file detection.
- refine “small formatting command” detection for awk/sed, handle cd
with -- or multiple operands, keep pipeline summaries focused on primary
command.
2026-01-13 16:59:33 -08:00
Ahmed Ibrahim
7e33ac7eb6 clean models manager (#9168)
Have only the following Methods:
- `list_models`: getting current available models
- `try_list_models`: sync version no refresh for tui use
- `get_default_model`: get the default model (should be tightened to
core and received on session configuration)
- `get_model_info`: get `ModelInfo` for a specific model (should be
tightened to core but used in tests)
- `refresh_if_new_etag`: trigger refresh on different etags

Also move the cache to its own struct
2026-01-13 16:55:33 -08:00
github-actions[bot]
ebbbee70c6 Update models.json (#9136)
Automated update of models.json.

Co-authored-by: aibrahim-oai <219906144+aibrahim-oai@users.noreply.github.com>
2026-01-13 16:46:24 -08:00
364 changed files with 28405 additions and 4629 deletions

1
.bazelversion Normal file
View File

@@ -0,0 +1 @@
8.5.1

View File

@@ -40,11 +40,18 @@ body:
description: |
For MacOS and Linux: copy the output of `uname -mprs`
For Windows: copy the output of `"$([Environment]::OSVersion | ForEach-Object VersionString) $(if ([Environment]::Is64BitOperatingSystem) { "x64" } else { "x86" })"` in the PowerShell console
- type: input
id: terminal
attributes:
label: What terminal emulator and version are you using (if applicable)?
description: Also note any multiplexer in use (screen / tmux / zellij)
description: |
E.g, VSCode, Terminal.app, iTerm2, Ghostty, Windows Terminal (WSL / PowerShell)
- type: textarea
id: actual
attributes:
label: What issue are you seeing?
description: Please include the full error messages and prompts with PII redacted. If possible, please provide text instead of a screenshot.
description: Please include the full error messages and prompts with PII redacted. If possible, please provide text instead of a screenshot.
validations:
required: true
- type: textarea

View File

@@ -59,7 +59,7 @@ jobs:
working-directory: codex-rs
steps:
- uses: actions/checkout@v6
- uses: dtolnay/rust-toolchain@1.90
- uses: dtolnay/rust-toolchain@1.92
with:
components: rustfmt
- name: cargo fmt
@@ -77,7 +77,7 @@ jobs:
working-directory: codex-rs
steps:
- uses: actions/checkout@v6
- uses: dtolnay/rust-toolchain@1.90
- uses: dtolnay/rust-toolchain@1.92
- uses: taiki-e/install-action@44c6d64aa62cd779e873306675c7a58e86d6d532 # v2
with:
tool: cargo-shear
@@ -88,7 +88,7 @@ jobs:
# --- CI to validate on different os/targets --------------------------------
lint_build:
name: Lint/Build — ${{ matrix.runner }} - ${{ matrix.target }}${{ matrix.profile == 'release' && ' (release)' || '' }}
runs-on: ${{ matrix.runner }}
runs-on: ${{ matrix.runs_on || matrix.runner }}
timeout-minutes: 30
needs: changed
# Keep job-level if to avoid spinning up runners when not needed
@@ -106,51 +106,78 @@ jobs:
fail-fast: false
matrix:
include:
- runner: macos-14
- runner: macos-15-xlarge
target: aarch64-apple-darwin
profile: dev
- runner: macos-14
- runner: macos-15-xlarge
target: x86_64-apple-darwin
profile: dev
- runner: ubuntu-24.04
target: x86_64-unknown-linux-musl
profile: dev
runs_on:
group: codex-runners
labels: codex-linux-x64
- runner: ubuntu-24.04
target: x86_64-unknown-linux-gnu
profile: dev
runs_on:
group: codex-runners
labels: codex-linux-x64
- runner: ubuntu-24.04-arm
target: aarch64-unknown-linux-musl
profile: dev
runs_on:
group: codex-runners
labels: codex-linux-arm64
- runner: ubuntu-24.04-arm
target: aarch64-unknown-linux-gnu
profile: dev
- runner: windows-latest
runs_on:
group: codex-runners
labels: codex-linux-arm64
- runner: windows-x64
target: x86_64-pc-windows-msvc
profile: dev
- runner: windows-11-arm
runs_on:
group: codex-runners
labels: codex-windows-x64
- runner: windows-arm64
target: aarch64-pc-windows-msvc
profile: dev
runs_on:
group: codex-runners
labels: codex-windows-arm64
# Also run representative release builds on Mac and Linux because
# there could be release-only build errors we want to catch.
# Hopefully this also pre-populates the build cache to speed up
# releases.
- runner: macos-14
- runner: macos-15-xlarge
target: aarch64-apple-darwin
profile: release
- runner: ubuntu-24.04
target: x86_64-unknown-linux-musl
profile: release
- runner: windows-latest
runs_on:
group: codex-runners
labels: codex-linux-x64
- runner: windows-x64
target: x86_64-pc-windows-msvc
profile: release
- runner: windows-11-arm
runs_on:
group: codex-runners
labels: codex-windows-x64
- runner: windows-arm64
target: aarch64-pc-windows-msvc
profile: release
runs_on:
group: codex-runners
labels: codex-windows-arm64
steps:
- uses: actions/checkout@v6
- uses: dtolnay/rust-toolchain@1.90
- uses: dtolnay/rust-toolchain@1.92
with:
targets: ${{ matrix.target }}
components: clippy
@@ -336,7 +363,7 @@ jobs:
tests:
name: Tests — ${{ matrix.runner }} - ${{ matrix.target }}
runs-on: ${{ matrix.runner }}
runs-on: ${{ matrix.runs_on || matrix.runner }}
timeout-minutes: 30
needs: changed
if: ${{ needs.changed.outputs.codex == 'true' || needs.changed.outputs.workflows == 'true' || github.event_name == 'push' }}
@@ -353,46 +380,43 @@ jobs:
fail-fast: false
matrix:
include:
- runner: macos-14
- runner: macos-15-xlarge
target: aarch64-apple-darwin
profile: dev
- runner: ubuntu-24.04
target: x86_64-unknown-linux-gnu
profile: dev
runs_on:
group: codex-runners
labels: codex-linux-x64
- runner: ubuntu-24.04-arm
target: aarch64-unknown-linux-gnu
profile: dev
- runner: windows-latest
runs_on:
group: codex-runners
labels: codex-linux-arm64
- runner: windows-x64
target: x86_64-pc-windows-msvc
profile: dev
- runner: windows-11-arm
runs_on:
group: codex-runners
labels: codex-windows-x64
- runner: windows-arm64
target: aarch64-pc-windows-msvc
profile: dev
runs_on:
group: codex-runners
labels: codex-windows-arm64
steps:
- uses: actions/checkout@v6
# We have been running out of space when running this job on Linux for
# x86_64-unknown-linux-gnu, so remove some unnecessary dependencies.
- name: Remove unnecessary dependencies to save space
if: ${{ startsWith(matrix.runner, 'ubuntu') }}
shell: bash
run: |
set -euo pipefail
sudo rm -rf \
/usr/local/lib/android \
/usr/share/dotnet \
/usr/local/share/boost \
/usr/local/lib/node_modules \
/opt/ghc
sudo apt-get remove -y docker.io docker-compose podman buildah
# Some integration tests rely on DotSlash being installed.
# See https://github.com/openai/codex/pull/7617.
- name: Install DotSlash
uses: facebook/install-dotslash@v2
- uses: dtolnay/rust-toolchain@1.90
- uses: dtolnay/rust-toolchain@1.92
with:
targets: ${{ matrix.target }}

View File

@@ -49,7 +49,7 @@ jobs:
needs: tag-check
name: Build - ${{ matrix.runner }} - ${{ matrix.target }}
runs-on: ${{ matrix.runner }}
timeout-minutes: 30
timeout-minutes: 60
permissions:
contents: read
id-token: write
@@ -80,7 +80,7 @@ jobs:
steps:
- uses: actions/checkout@v6
- uses: dtolnay/rust-toolchain@1.90
- uses: dtolnay/rust-toolchain@1.92
with:
targets: ${{ matrix.target }}

View File

@@ -24,7 +24,7 @@ jobs:
node-version: 22
cache: pnpm
- uses: dtolnay/rust-toolchain@1.90
- uses: dtolnay/rust-toolchain@1.92
- name: build codex
run: cargo build --bin codex

View File

@@ -93,7 +93,7 @@ jobs:
- name: Checkout repository
uses: actions/checkout@v6
- uses: dtolnay/rust-toolchain@1.90
- uses: dtolnay/rust-toolchain@1.92
with:
targets: ${{ matrix.target }}
@@ -198,7 +198,7 @@ jobs:
shell: bash
run: |
set -euo pipefail
git clone --depth 1 https://github.com/bminor/bash /tmp/bash
git clone --depth 1 https://github.com/bolinfest/bash /tmp/bash
cd /tmp/bash
git fetch --depth 1 origin a8a1c2fac029404d3f42cd39f5a20f24b6e4fe4b
git checkout a8a1c2fac029404d3f42cd39f5a20f24b6e4fe4b
@@ -240,7 +240,7 @@ jobs:
shell: bash
run: |
set -euo pipefail
git clone --depth 1 https://github.com/bminor/bash /tmp/bash
git clone --depth 1 https://github.com/bolinfest/bash /tmp/bash
cd /tmp/bash
git fetch --depth 1 origin a8a1c2fac029404d3f42cd39f5a20f24b6e4fe4b
git checkout a8a1c2fac029404d3f42cd39f5a20f24b6e4fe4b

View File

@@ -18,8 +18,7 @@ In the codex-rs folder where the rust code lives:
Run `just fmt` (in `codex-rs` directory) automatically after making Rust code changes; do not ask for approval to run it. Before finalizing a change to `codex-rs`, run `just fix -p <project>` (in `codex-rs` directory) to fix any linter issues in the code. Prefer scoping with `-p` to avoid slow workspacewide Clippy builds; only run `just fix` without `-p` if you changed shared crates. Additionally, run the tests:
1. Run the test for the specific project that was changed. For example, if changes were made in `codex-rs/tui`, run `cargo test -p codex-tui`.
2. Once those pass, if any changes were made in common, core, or protocol, run the complete test suite with `cargo test --all-features`.
When running interactively, ask the user before running `just fix` to finalize. `just fmt` does not require approval. project-specific or individual tests can be run without asking the user, but do ask the user before running the complete test suite.
2. Once those pass, if any changes were made in common, core, or protocol, run the complete test suite with `cargo test --all-features`. project-specific or individual tests can be run without asking the user, but do ask the user before running the complete test suite.
## TUI style conventions

0
codex-cli/bin/codex.js Normal file → Executable file
View File

94
codex-rs/Cargo.lock generated
View File

@@ -360,16 +360,19 @@ dependencies = [
"objc2-foundation",
"parking_lot",
"percent-encoding",
"windows-sys 0.52.0",
"windows-sys 0.60.2",
"wl-clipboard-rs",
"x11rb",
]
[[package]]
name = "arc-swap"
version = "1.7.1"
version = "1.8.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "69f7f8c3906b62b754cd5326047894316021dcfe5a194c8ea52bdd94934a3457"
checksum = "51d03449bb8ca2cc2ef70869af31463d1ae5ccc8fa3e334b307203fbf815207e"
dependencies = [
"rustversion",
]
[[package]]
name = "arrayvec"
@@ -861,9 +864,9 @@ dependencies = [
[[package]]
name = "chrono"
version = "0.4.42"
version = "0.4.43"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "145052bdd345b87320e369255277e3fb5152762ad123a901ef5c262dd38fe8d2"
checksum = "fac4744fb15ae8337dc853fee7fb3f4e48c0fbaa23d0afe49c447b4fab126118"
dependencies = [
"iana-time-zone",
"js-sys",
@@ -1298,7 +1301,7 @@ dependencies = [
"codex-windows-sandbox",
"core-foundation 0.9.4",
"core_test_support",
"ctor 0.5.0",
"ctor 0.6.3",
"dirs",
"dunce",
"encoding_rs",
@@ -1524,6 +1527,7 @@ dependencies = [
"codex-utils-absolute-path",
"landlock",
"libc",
"pretty_assertions",
"seccompiler",
"tempfile",
"tokio",
@@ -1680,7 +1684,7 @@ dependencies = [
"anyhow",
"clap",
"codex-process-hardening",
"ctor 0.5.0",
"ctor 0.6.3",
"libc",
"reqwest",
"serde",
@@ -1884,7 +1888,9 @@ dependencies = [
name = "codex-utils-absolute-path"
version = "0.0.0"
dependencies = [
"dirs",
"path-absolutize",
"pretty_assertions",
"schemars 0.8.22",
"serde",
"serde_json",
@@ -1938,8 +1944,10 @@ dependencies = [
"anyhow",
"filedescriptor",
"lazy_static",
"libc",
"log",
"portable-pty",
"pretty_assertions",
"shared_library",
"tokio",
"winapi",
@@ -2259,9 +2267,9 @@ dependencies = [
[[package]]
name = "ctor"
version = "0.5.0"
version = "0.6.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "67773048316103656a637612c4a62477603b777d91d9c62ff2290f9cde178fdb"
checksum = "424e0138278faeb2b401f174ad17e715c829512d74f3d1e81eb43365c2e0590e"
dependencies = [
"ctor-proc-macro",
"dtor",
@@ -2269,9 +2277,9 @@ dependencies = [
[[package]]
name = "ctor-proc-macro"
version = "0.0.6"
version = "0.0.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e2931af7e13dc045d8e9d26afccc6fa115d64e115c9c84b1166288b46f6782c2"
checksum = "52560adf09603e58c9a7ee1fe1dcb95a16927b17c127f0ac02d6e768a0e25bc1"
[[package]]
name = "darling"
@@ -2826,7 +2834,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "778e2ac28f6c47af28e4907f13ffd1e1ddbd400980a9abd7c8df189bf578a5ad"
dependencies = [
"libc",
"windows-sys 0.52.0",
"windows-sys 0.60.2",
]
[[package]]
@@ -2923,7 +2931,7 @@ checksum = "0ce92ff622d6dadf7349484f42c93271a0d49b7cc4d466a936405bacbe10aa78"
dependencies = [
"cfg-if",
"rustix 1.0.8",
"windows-sys 0.52.0",
"windows-sys 0.59.0",
]
[[package]]
@@ -3449,7 +3457,7 @@ dependencies = [
"tokio",
"tokio-rustls",
"tower-service",
"webpki-roots",
"webpki-roots 1.0.2",
]
[[package]]
@@ -3499,7 +3507,7 @@ dependencies = [
"libc",
"percent-encoding",
"pin-project-lite",
"socket2 0.6.1",
"socket2 0.5.10",
"system-configuration",
"tokio",
"tower-service",
@@ -3864,7 +3872,7 @@ checksum = "e04d7f318608d35d4b61ddd75cbdaee86b023ebe2bd5a66ee0915f0bf93095a9"
dependencies = [
"hermit-abi",
"libc",
"windows-sys 0.52.0",
"windows-sys 0.59.0",
]
[[package]]
@@ -4157,9 +4165,9 @@ dependencies = [
[[package]]
name = "log"
version = "0.4.28"
version = "0.4.29"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "34080505efa8e45a4b816c349525ebe327ceaa8559756f0356cba97ef3bf7432"
checksum = "5e5032e24019045c762d3c0f28f5b6b8bbf38563a65908389bf7978758920897"
[[package]]
name = "logos"
@@ -5336,7 +5344,7 @@ dependencies = [
"quinn-udp",
"rustc-hash",
"rustls",
"socket2 0.6.1",
"socket2 0.5.10",
"thiserror 2.0.17",
"tokio",
"tracing",
@@ -5373,9 +5381,9 @@ dependencies = [
"cfg_aliases 0.2.1",
"libc",
"once_cell",
"socket2 0.6.1",
"socket2 0.5.10",
"tracing",
"windows-sys 0.52.0",
"windows-sys 0.60.2",
]
[[package]]
@@ -5658,7 +5666,7 @@ dependencies = [
"wasm-bindgen-futures",
"wasm-streams",
"web-sys",
"webpki-roots",
"webpki-roots 1.0.2",
]
[[package]]
@@ -5754,7 +5762,7 @@ dependencies = [
"errno",
"libc",
"linux-raw-sys 0.4.15",
"windows-sys 0.52.0",
"windows-sys 0.59.0",
]
[[package]]
@@ -5767,7 +5775,7 @@ dependencies = [
"errno",
"libc",
"linux-raw-sys 0.9.4",
"windows-sys 0.52.0",
"windows-sys 0.60.2",
]
[[package]]
@@ -7068,9 +7076,9 @@ checksum = "1f3ccbac311fea05f86f61904b462b55fb3df8837a366dfc601a0161d0532f20"
[[package]]
name = "tokio"
version = "1.48.0"
version = "1.49.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ff360e02eab121e0bc37a2d3b4d4dc622e6eda3a8e5253d5435ecf5bd4c68408"
checksum = "72a2903cd7736441aac9df9d7688bd0ce48edccaadf181c3b90be801e81d3d86"
dependencies = [
"bytes",
"libc",
@@ -7141,14 +7149,18 @@ dependencies = [
[[package]]
name = "tokio-tungstenite"
version = "0.21.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "c83b561d025642014097b66e6c1bb422783339e0909e4429cde4749d1990bc38"
version = "0.28.0"
source = "git+https://github.com/JakkuSakura/tokio-tungstenite?rev=2ae536b0de793f3ddf31fc2f22d445bf1ef2023d#2ae536b0de793f3ddf31fc2f22d445bf1ef2023d"
dependencies = [
"futures-util",
"log",
"rustls",
"rustls-native-certs",
"rustls-pki-types",
"tokio",
"tokio-rustls",
"tungstenite",
"webpki-roots 0.26.11",
]
[[package]]
@@ -7547,20 +7559,19 @@ dependencies = [
[[package]]
name = "tungstenite"
version = "0.21.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9ef1a641ea34f399a848dea702823bbecfb4c486f911735368f1f137cb8257e1"
version = "0.28.0"
source = "git+https://github.com/JakkuSakura/tungstenite-rs?rev=f514de8644821113e5d18a027d6d28a5c8cc0a6e#f514de8644821113e5d18a027d6d28a5c8cc0a6e"
dependencies = [
"byteorder",
"bytes",
"data-encoding",
"http 1.3.1",
"httparse",
"log",
"rand 0.8.5",
"rand 0.9.2",
"rustls",
"rustls-pki-types",
"sha1",
"thiserror 1.0.69",
"url",
"thiserror 2.0.17",
"utf-8",
]
@@ -8031,6 +8042,15 @@ dependencies = [
"rustls-pki-types",
]
[[package]]
name = "webpki-roots"
version = "0.26.11"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "521bc38abb08001b01866da9f51eb7c5d647a19260e00054a8c7fd5f9e57f7a9"
dependencies = [
"webpki-roots 1.0.2",
]
[[package]]
name = "webpki-roots"
version = "1.0.2"
@@ -8085,7 +8105,7 @@ version = "0.1.9"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "cf221c93e13a30d793f7645a0e7762c55d169dbb0a49671918a2319d289b10bb"
dependencies = [
"windows-sys 0.52.0",
"windows-sys 0.59.0",
]
[[package]]

View File

@@ -122,12 +122,12 @@ axum = { version = "0.8", default-features = false }
base64 = "0.22.1"
bytes = "1.10.1"
chardetng = "0.1.17"
chrono = "0.4.42"
chrono = "0.4.43"
clap = "4"
clap_complete = "4"
color-eyre = "0.6.3"
crossterm = "0.28.1"
ctor = "0.5.0"
ctor = "0.6.3"
derive_more = "2"
diffy = "0.4.2"
dirs = "6"
@@ -182,7 +182,7 @@ ratatui-core = "0.1.0"
ratatui-macros = "0.6.0"
regex = "1.12.2"
regex-lite = "0.1.8"
reqwest = "0.12"
reqwest = { version = "0.12", features = ["rustls-tls-webpki-roots", "rustls-tls-native-roots"] }
rmcp = { version = "0.12.0", default-features = false }
schemars = "0.8.22"
seccompiler = "0.5.0"
@@ -212,7 +212,7 @@ tiny_http = "0.12"
tokio = "1"
tokio-stream = "0.1.18"
tokio-test = "0.4"
tokio-tungstenite = "0.21.0"
tokio-tungstenite = { version = "0.28.0", features = ["rustls-tls-webpki-roots", "rustls-tls-native-roots", "proxy"] }
tokio-util = "0.7.18"
toml = "0.9.5"
toml_edit = "0.24.0"
@@ -303,6 +303,10 @@ opt-level = 0
# ratatui = { path = "../../ratatui" }
crossterm = { git = "https://github.com/nornagon/crossterm", branch = "nornagon/color-query" }
ratatui = { git = "https://github.com/nornagon/ratatui", branch = "nornagon-v0.29.0-patch" }
tokio-tungstenite = { git = "https://github.com/JakkuSakura/tokio-tungstenite", rev = "2ae536b0de793f3ddf31fc2f22d445bf1ef2023d" }
# Uncomment to debug local changes.
# rmcp = { path = "../../rust-sdk/crates/rmcp" }
[patch."ssh://git@github.com/JakkuSakura/tungstenite-rs.git"]
tungstenite = { git = "https://github.com/JakkuSakura/tungstenite-rs", rev = "f514de8644821113e5d18a027d6d28a5c8cc0a6e" }

View File

@@ -133,6 +133,10 @@ client_request_definitions! {
params: v2::SkillsListParams,
response: v2::SkillsListResponse,
},
SkillsConfigWrite => "skills/config/write" {
params: v2::SkillsConfigWriteParams,
response: v2::SkillsConfigWriteResponse,
},
TurnStart => "turn/start" {
params: v2::TurnStartParams,
response: v2::TurnStartResponse,
@@ -150,6 +154,11 @@ client_request_definitions! {
params: v2::ModelListParams,
response: v2::ModelListResponse,
},
/// EXPERIMENTAL - list collaboration mode presets.
CollaborationModeList => "collaborationMode/list" {
params: v2::CollaborationModeListParams,
response: v2::CollaborationModeListResponse,
},
McpServerOauthLogin => "mcpServer/oauth/login" {
params: v2::McpServerOauthLoginParams,
@@ -501,6 +510,12 @@ server_request_definitions! {
response: v2::FileChangeRequestApprovalResponse,
},
/// EXPERIMENTAL - Request input from the user for a tool call.
ToolRequestUserInput => "item/tool/requestUserInput" {
params: v2::ToolRequestUserInputParams,
response: v2::ToolRequestUserInputResponse,
},
/// DEPRECATED APIs below
/// Request to approve a patch.
/// This request is used for Turns started via the legacy APIs (i.e. SendUserTurn, SendUserMessage).
@@ -567,6 +582,7 @@ server_notification_definitions! {
ReasoningTextDelta => "item/reasoning/textDelta" (v2::ReasoningTextDeltaNotification),
ContextCompacted => "thread/compacted" (v2::ContextCompactedNotification),
DeprecationNotice => "deprecationNotice" (v2::DeprecationNoticeNotification),
ConfigWarning => "configWarning" (v2::ConfigWarningNotification),
/// Notifies the user of world-writable directories on Windows, which cannot be protected by the sandbox.
WindowsWorldWritableWarning => "windows/worldWritableWarning" (v2::WindowsWorldWritableWarningNotification),
@@ -873,4 +889,21 @@ mod tests {
);
Ok(())
}
#[test]
fn serialize_list_collaboration_modes() -> Result<()> {
let request = ClientRequest::CollaborationModeList {
request_id: RequestId::Integer(7),
params: v2::CollaborationModeListParams::default(),
};
assert_eq!(
json!({
"method": "collaborationMode/list",
"id": 7,
"params": {}
}),
serde_json::to_value(&request)?,
);
Ok(())
}
}

View File

@@ -197,6 +197,12 @@ impl ThreadHistoryBuilder {
if !payload.message.trim().is_empty() {
content.push(UserInput::Text {
text: payload.message.clone(),
text_elements: payload
.text_elements
.iter()
.cloned()
.map(Into::into)
.collect(),
});
}
if let Some(images) = &payload.images {
@@ -204,6 +210,9 @@ impl ThreadHistoryBuilder {
content.push(UserInput::Image { url: image.clone() });
}
}
for path in &payload.local_images {
content.push(UserInput::LocalImage { path: path.clone() });
}
content
}
}
@@ -244,6 +253,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "First turn".into(),
images: Some(vec!["https://example.com/one.png".into()]),
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentMessage(AgentMessageEvent {
message: "Hi there".into(),
@@ -257,6 +268,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "Second turn".into(),
images: None,
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentMessage(AgentMessageEvent {
message: "Reply two".into(),
@@ -277,6 +290,7 @@ mod tests {
content: vec![
UserInput::Text {
text: "First turn".into(),
text_elements: Vec::new(),
},
UserInput::Image {
url: "https://example.com/one.png".into(),
@@ -308,7 +322,8 @@ mod tests {
ThreadItem::UserMessage {
id: "item-4".into(),
content: vec![UserInput::Text {
text: "Second turn".into()
text: "Second turn".into(),
text_elements: Vec::new(),
}],
}
);
@@ -327,6 +342,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "Turn start".into(),
images: None,
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentReasoning(AgentReasoningEvent {
text: "first summary".into(),
@@ -371,6 +388,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "Please do the thing".into(),
images: None,
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentMessage(AgentMessageEvent {
message: "Working...".into(),
@@ -381,6 +400,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "Let's try again".into(),
images: None,
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentMessage(AgentMessageEvent {
message: "Second attempt complete.".into(),
@@ -398,7 +419,8 @@ mod tests {
ThreadItem::UserMessage {
id: "item-1".into(),
content: vec![UserInput::Text {
text: "Please do the thing".into()
text: "Please do the thing".into(),
text_elements: Vec::new(),
}],
}
);
@@ -418,7 +440,8 @@ mod tests {
ThreadItem::UserMessage {
id: "item-3".into(),
content: vec![UserInput::Text {
text: "Let's try again".into()
text: "Let's try again".into(),
text_elements: Vec::new(),
}],
}
);
@@ -437,6 +460,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "First".into(),
images: None,
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentMessage(AgentMessageEvent {
message: "A1".into(),
@@ -444,6 +469,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "Second".into(),
images: None,
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentMessage(AgentMessageEvent {
message: "A2".into(),
@@ -452,6 +479,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "Third".into(),
images: None,
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentMessage(AgentMessageEvent {
message: "A3".into(),
@@ -469,6 +498,7 @@ mod tests {
id: "item-1".into(),
content: vec![UserInput::Text {
text: "First".into(),
text_elements: Vec::new(),
}],
},
ThreadItem::AgentMessage {
@@ -486,6 +516,7 @@ mod tests {
id: "item-3".into(),
content: vec![UserInput::Text {
text: "Third".into(),
text_elements: Vec::new(),
}],
},
ThreadItem::AgentMessage {
@@ -504,6 +535,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "One".into(),
images: None,
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentMessage(AgentMessageEvent {
message: "A1".into(),
@@ -511,6 +544,8 @@ mod tests {
EventMsg::UserMessage(UserMessageEvent {
message: "Two".into(),
images: None,
text_elements: Vec::new(),
local_images: Vec::new(),
}),
EventMsg::AgentMessage(AgentMessageEvent {
message: "A2".into(),

View File

@@ -16,6 +16,8 @@ use codex_protocol::protocol::ReviewDecision;
use codex_protocol::protocol::SandboxPolicy;
use codex_protocol::protocol::SessionSource;
use codex_protocol::protocol::TurnAbortReason;
use codex_protocol::user_input::ByteRange as CoreByteRange;
use codex_protocol::user_input::TextElement as CoreTextElement;
use codex_utils_absolute_path::AbsolutePathBuf;
use schemars::JsonSchema;
use serde::Deserialize;
@@ -126,6 +128,7 @@ pub struct ConversationSummary {
pub path: PathBuf,
pub preview: String,
pub timestamp: Option<String>,
pub updated_at: Option<String>,
pub model_provider: String,
pub cwd: PathBuf,
pub cli_version: String,
@@ -444,9 +447,71 @@ pub struct RemoveConversationListenerParams {
#[serde(rename_all = "camelCase")]
#[serde(tag = "type", content = "data")]
pub enum InputItem {
Text { text: String },
Image { image_url: String },
LocalImage { path: PathBuf },
Text {
text: String,
/// UI-defined spans within `text` used to render or persist special elements.
#[serde(default)]
text_elements: Vec<V1TextElement>,
},
Image {
image_url: String,
},
LocalImage {
path: PathBuf,
},
}
#[derive(Serialize, Deserialize, Debug, Clone, Copy, PartialEq, Eq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(rename = "ByteRange")]
pub struct V1ByteRange {
/// Start byte offset (inclusive) within the UTF-8 text buffer.
pub start: usize,
/// End byte offset (exclusive) within the UTF-8 text buffer.
pub end: usize,
}
impl From<CoreByteRange> for V1ByteRange {
fn from(value: CoreByteRange) -> Self {
Self {
start: value.start,
end: value.end,
}
}
}
impl From<V1ByteRange> for CoreByteRange {
fn from(value: V1ByteRange) -> Self {
Self {
start: value.start,
end: value.end,
}
}
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(rename = "TextElement")]
pub struct V1TextElement {
/// Byte range in the parent `text` buffer that this element occupies.
pub byte_range: V1ByteRange,
/// Optional human-readable placeholder for the element, displayed in the UI.
pub placeholder: Option<String>,
}
impl From<CoreTextElement> for V1TextElement {
fn from(value: CoreTextElement) -> Self {
Self {
byte_range: value.byte_range.into(),
placeholder: value._placeholder_for_conversion_only().map(str::to_string),
}
}
}
impl From<V1TextElement> for CoreTextElement {
fn from(value: V1TextElement) -> Self {
Self::new(value.byte_range.into(), value.placeholder)
}
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]

View File

@@ -4,10 +4,12 @@ use std::path::PathBuf;
use crate::protocol::common::AuthMode;
use codex_protocol::account::PlanType;
use codex_protocol::approvals::ExecPolicyAmendment as CoreExecPolicyAmendment;
use codex_protocol::config_types::CollaborationMode;
use codex_protocol::config_types::ForcedLoginMethod;
use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::config_types::SandboxMode as CoreSandboxMode;
use codex_protocol::config_types::Verbosity;
use codex_protocol::config_types::WebSearchMode;
use codex_protocol::items::AgentMessageContent as CoreAgentMessageContent;
use codex_protocol::items::TurnItem as CoreTurnItem;
use codex_protocol::models::ResponseItem;
@@ -15,6 +17,7 @@ use codex_protocol::openai_models::ReasoningEffort;
use codex_protocol::parse_command::ParsedCommand as CoreParsedCommand;
use codex_protocol::plan_tool::PlanItemArg as CorePlanItemArg;
use codex_protocol::plan_tool::StepStatus as CorePlanStepStatus;
use codex_protocol::protocol::AgentStatus as CoreAgentStatus;
use codex_protocol::protocol::AskForApproval as CoreAskForApproval;
use codex_protocol::protocol::CodexErrorInfo as CoreCodexErrorInfo;
use codex_protocol::protocol::CreditsSnapshot as CoreCreditsSnapshot;
@@ -23,10 +26,13 @@ use codex_protocol::protocol::RateLimitSnapshot as CoreRateLimitSnapshot;
use codex_protocol::protocol::RateLimitWindow as CoreRateLimitWindow;
use codex_protocol::protocol::SessionSource as CoreSessionSource;
use codex_protocol::protocol::SkillErrorInfo as CoreSkillErrorInfo;
use codex_protocol::protocol::SkillInterface as CoreSkillInterface;
use codex_protocol::protocol::SkillMetadata as CoreSkillMetadata;
use codex_protocol::protocol::SkillScope as CoreSkillScope;
use codex_protocol::protocol::TokenUsage as CoreTokenUsage;
use codex_protocol::protocol::TokenUsageInfo as CoreTokenUsageInfo;
use codex_protocol::user_input::ByteRange as CoreByteRange;
use codex_protocol::user_input::TextElement as CoreTextElement;
use codex_protocol::user_input::UserInput as CoreUserInput;
use codex_utils_absolute_path::AbsolutePathBuf;
use mcp_types::ContentBlock as McpContentBlock;
@@ -327,6 +333,7 @@ pub struct ProfileV2 {
pub model_reasoning_effort: Option<ReasoningEffort>,
pub model_reasoning_summary: Option<ReasoningSummary>,
pub model_verbosity: Option<Verbosity>,
pub web_search: Option<WebSearchMode>,
pub chatgpt_base_url: Option<String>,
#[serde(default, flatten)]
pub additional: HashMap<String, JsonValue>,
@@ -355,6 +362,7 @@ pub struct Config {
pub sandbox_workspace_write: Option<SandboxWorkspaceWrite>,
pub forced_chatgpt_workspace_id: Option<String>,
pub forced_login_method: Option<ForcedLoginMethod>,
pub web_search: Option<WebSearchMode>,
pub tools: Option<ToolsV2>,
pub profile: Option<String>,
#[serde(default)]
@@ -909,6 +917,20 @@ pub struct ModelListResponse {
pub next_cursor: Option<String>,
}
/// EXPERIMENTAL - list collaboration mode presets.
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Default, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct CollaborationModeListParams {}
/// EXPERIMENTAL - collaboration mode presets response.
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct CollaborationModeListResponse {
pub data: Vec<CollaborationMode>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
@@ -1175,11 +1197,21 @@ pub struct ThreadListParams {
pub cursor: Option<String>,
/// Optional page size; defaults to a reasonable server-side value.
pub limit: Option<u32>,
/// Optional sort key; defaults to created_at.
pub sort_key: Option<ThreadSortKey>,
/// Optional provider filter; when set, only sessions recorded under these
/// providers are returned. When present but empty, includes all providers.
pub model_providers: Option<Vec<String>>,
}
#[derive(Serialize, Deserialize, Debug, Clone, Copy, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "snake_case")]
#[ts(export_to = "v2/")]
pub enum ThreadSortKey {
CreatedAt,
UpdatedAt,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
@@ -1248,11 +1280,34 @@ pub enum SkillScope {
pub struct SkillMetadata {
pub name: String,
pub description: String,
#[ts(optional)]
#[serde(default, skip_serializing_if = "Option::is_none")]
#[ts(optional)]
/// Legacy short_description from SKILL.md. Prefer SKILL.toml interface.short_description.
pub short_description: Option<String>,
#[serde(default, skip_serializing_if = "Option::is_none")]
#[ts(optional)]
pub interface: Option<SkillInterface>,
pub path: PathBuf,
pub scope: SkillScope,
pub enabled: bool,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct SkillInterface {
#[ts(optional)]
pub display_name: Option<String>,
#[ts(optional)]
pub short_description: Option<String>,
#[ts(optional)]
pub icon_small: Option<PathBuf>,
#[ts(optional)]
pub icon_large: Option<PathBuf>,
#[ts(optional)]
pub brand_color: Option<String>,
#[ts(optional)]
pub default_prompt: Option<String>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
@@ -1272,14 +1327,44 @@ pub struct SkillsListEntry {
pub errors: Vec<SkillErrorInfo>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct SkillsConfigWriteParams {
pub path: PathBuf,
pub enabled: bool,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct SkillsConfigWriteResponse {
pub effective_enabled: bool,
}
impl From<CoreSkillMetadata> for SkillMetadata {
fn from(value: CoreSkillMetadata) -> Self {
Self {
name: value.name,
description: value.description,
short_description: value.short_description,
interface: value.interface.map(SkillInterface::from),
path: value.path,
scope: value.scope.into(),
enabled: true,
}
}
}
impl From<CoreSkillInterface> for SkillInterface {
fn from(value: CoreSkillInterface) -> Self {
Self {
display_name: value.display_name,
short_description: value.short_description,
brand_color: value.brand_color,
default_prompt: value.default_prompt,
icon_small: value.icon_small,
icon_large: value.icon_large,
}
}
}
@@ -1316,6 +1401,9 @@ pub struct Thread {
/// Unix timestamp (in seconds) when the thread was created.
#[ts(type = "number")]
pub created_at: i64,
/// Unix timestamp (in seconds) when the thread was last updated.
#[ts(type = "number")]
pub updated_at: i64,
/// [UNSTABLE] Path to the thread on disk.
pub path: PathBuf,
/// Working directory captured for the thread.
@@ -1465,6 +1553,10 @@ pub struct TurnStartParams {
pub summary: Option<ReasoningSummary>,
/// Optional JSON Schema used to constrain the final assistant message for this turn.
pub output_schema: Option<JsonValue>,
/// EXPERIMENTAL - set a pre-set collaboration mode.
/// Takes precedence over model, reasoning_effort, and developer instructions if set.
pub collaboration_mode: Option<CollaborationMode>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
@@ -1540,21 +1632,107 @@ pub struct TurnInterruptParams {
pub struct TurnInterruptResponse {}
// User input types
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Eq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct ByteRange {
pub start: usize,
pub end: usize,
}
impl From<CoreByteRange> for ByteRange {
fn from(value: CoreByteRange) -> Self {
Self {
start: value.start,
end: value.end,
}
}
}
impl From<ByteRange> for CoreByteRange {
fn from(value: ByteRange) -> Self {
Self {
start: value.start,
end: value.end,
}
}
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct TextElement {
/// Byte range in the parent `text` buffer that this element occupies.
pub byte_range: ByteRange,
/// Optional human-readable placeholder for the element, displayed in the UI.
placeholder: Option<String>,
}
impl TextElement {
pub fn new(byte_range: ByteRange, placeholder: Option<String>) -> Self {
Self {
byte_range,
placeholder,
}
}
pub fn set_placeholder(&mut self, placeholder: Option<String>) {
self.placeholder = placeholder;
}
pub fn placeholder(&self) -> Option<&str> {
self.placeholder.as_deref()
}
}
impl From<CoreTextElement> for TextElement {
fn from(value: CoreTextElement) -> Self {
Self::new(
value.byte_range.into(),
value._placeholder_for_conversion_only().map(str::to_string),
)
}
}
impl From<TextElement> for CoreTextElement {
fn from(value: TextElement) -> Self {
Self::new(value.byte_range.into(), value.placeholder)
}
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(tag = "type", rename_all = "camelCase")]
#[ts(tag = "type")]
#[ts(export_to = "v2/")]
pub enum UserInput {
Text { text: String },
Image { url: String },
LocalImage { path: PathBuf },
Skill { name: String, path: PathBuf },
Text {
text: String,
/// UI-defined spans within `text` used to render or persist special elements.
#[serde(default)]
text_elements: Vec<TextElement>,
},
Image {
url: String,
},
LocalImage {
path: PathBuf,
},
Skill {
name: String,
path: PathBuf,
},
}
impl UserInput {
pub fn into_core(self) -> CoreUserInput {
match self {
UserInput::Text { text } => CoreUserInput::Text { text },
UserInput::Text {
text,
text_elements,
} => CoreUserInput::Text {
text,
text_elements: text_elements.into_iter().map(Into::into).collect(),
},
UserInput::Image { url } => CoreUserInput::Image { image_url: url },
UserInput::LocalImage { path } => CoreUserInput::LocalImage { path },
UserInput::Skill { name, path } => CoreUserInput::Skill { name, path },
@@ -1565,7 +1743,13 @@ impl UserInput {
impl From<CoreUserInput> for UserInput {
fn from(value: CoreUserInput) -> Self {
match value {
CoreUserInput::Text { text } => UserInput::Text { text },
CoreUserInput::Text {
text,
text_elements,
} => UserInput::Text {
text,
text_elements: text_elements.into_iter().map(Into::into).collect(),
},
CoreUserInput::Image { image_url } => UserInput::Image { url: image_url },
CoreUserInput::LocalImage { path } => UserInput::LocalImage { path },
CoreUserInput::Skill { name, path } => UserInput::Skill { name, path },
@@ -1640,6 +1824,25 @@ pub enum ThreadItem {
},
#[serde(rename_all = "camelCase")]
#[ts(rename_all = "camelCase")]
CollabAgentToolCall {
/// Unique identifier for this collab tool call.
id: String,
/// Name of the collab tool that was invoked.
tool: CollabAgentTool,
/// Current status of the collab tool call.
status: CollabAgentToolCallStatus,
/// Thread ID of the agent issuing the collab request.
sender_thread_id: String,
/// Thread ID of the receiving agent, when applicable. In case of spawn operation,
/// this corresponds to the newly spawned agent.
receiver_thread_ids: Vec<String>,
/// Prompt text sent as part of the collab tool call, when available.
prompt: Option<String>,
/// Last known status of the target agents, when available.
agents_states: HashMap<String, CollabAgentState>,
},
#[serde(rename_all = "camelCase")]
#[ts(rename_all = "camelCase")]
WebSearch { id: String, query: String },
#[serde(rename_all = "camelCase")]
#[ts(rename_all = "camelCase")]
@@ -1692,6 +1895,16 @@ pub enum CommandExecutionStatus {
Declined,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub enum CollabAgentTool {
SpawnAgent,
SendInput,
Wait,
CloseAgent,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
@@ -1730,6 +1943,66 @@ pub enum McpToolCallStatus {
Failed,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Eq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub enum CollabAgentToolCallStatus {
InProgress,
Completed,
Failed,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Eq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub enum CollabAgentStatus {
PendingInit,
Running,
Completed,
Errored,
Shutdown,
NotFound,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Eq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct CollabAgentState {
pub status: CollabAgentStatus,
pub message: Option<String>,
}
impl From<CoreAgentStatus> for CollabAgentState {
fn from(value: CoreAgentStatus) -> Self {
match value {
CoreAgentStatus::PendingInit => Self {
status: CollabAgentStatus::PendingInit,
message: None,
},
CoreAgentStatus::Running => Self {
status: CollabAgentStatus::Running,
message: None,
},
CoreAgentStatus::Completed(message) => Self {
status: CollabAgentStatus::Completed,
message,
},
CoreAgentStatus::Errored(message) => Self {
status: CollabAgentStatus::Errored,
message: Some(message),
},
CoreAgentStatus::Shutdown => Self {
status: CollabAgentStatus::Shutdown,
message: None,
},
CoreAgentStatus::NotFound => Self {
status: CollabAgentStatus::NotFound,
message: None,
},
}
}
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
@@ -2018,6 +2291,54 @@ pub struct FileChangeRequestApprovalResponse {
pub decision: FileChangeApprovalDecision,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
/// EXPERIMENTAL. Defines a single selectable option for request_user_input.
pub struct ToolRequestUserInputOption {
pub label: String,
pub description: String,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
/// EXPERIMENTAL. Represents one request_user_input question and its optional options.
pub struct ToolRequestUserInputQuestion {
pub id: String,
pub header: String,
pub question: String,
pub options: Option<Vec<ToolRequestUserInputOption>>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
/// EXPERIMENTAL. Params sent with a request_user_input event.
pub struct ToolRequestUserInputParams {
pub thread_id: String,
pub turn_id: String,
pub item_id: String,
pub questions: Vec<ToolRequestUserInputQuestion>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
/// EXPERIMENTAL. Captures a user's answer to a request_user_input question.
pub struct ToolRequestUserInputAnswer {
pub selected: Vec<String>,
pub other: Option<String>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
/// EXPERIMENTAL. Response payload mapping question ids to answers.
pub struct ToolRequestUserInputResponse {
pub answers: HashMap<String, ToolRequestUserInputAnswer>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
@@ -2107,6 +2428,16 @@ pub struct DeprecationNoticeNotification {
pub details: Option<String>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, JsonSchema, TS)]
#[serde(rename_all = "camelCase")]
#[ts(export_to = "v2/")]
pub struct ConfigWarningNotification {
/// Concise summary of the warning.
pub summary: String,
/// Optional extra guidance or error details.
pub details: Option<String>,
}
#[cfg(test)]
mod tests {
use super::*;
@@ -2147,6 +2478,7 @@ mod tests {
content: vec![
CoreUserInput::Text {
text: "hello".to_string(),
text_elements: Vec::new(),
},
CoreUserInput::Image {
image_url: "https://example.com/image.png".to_string(),
@@ -2168,6 +2500,7 @@ mod tests {
content: vec![
UserInput::Text {
text: "hello".to_string(),
text_elements: Vec::new(),
},
UserInput::Image {
url: "https://example.com/image.png".to_string(),

View File

@@ -256,7 +256,11 @@ fn send_message_v2_with_policies(
println!("< thread/start response: {thread_response:?}");
let mut turn_params = TurnStartParams {
thread_id: thread_response.thread.id.clone(),
input: vec![V2UserInput::Text { text: user_message }],
input: vec![V2UserInput::Text {
text: user_message,
// Test client sends plain text without UI element ranges.
text_elements: Vec::new(),
}],
..Default::default()
};
turn_params.approval_policy = approval_policy;
@@ -288,6 +292,8 @@ fn send_follow_up_v2(
thread_id: thread_response.thread.id.clone(),
input: vec![V2UserInput::Text {
text: first_message,
// Test client sends plain text without UI element ranges.
text_elements: Vec::new(),
}],
..Default::default()
};
@@ -299,6 +305,8 @@ fn send_follow_up_v2(
thread_id: thread_response.thread.id.clone(),
input: vec![V2UserInput::Text {
text: follow_up_message,
// Test client sends plain text without UI element ranges.
text_elements: Vec::new(),
}],
..Default::default()
};
@@ -471,6 +479,8 @@ impl CodexClient {
conversation_id: *conversation_id,
items: vec![InputItem::Text {
text: message.to_string(),
// Test client sends plain text without UI element ranges.
text_elements: Vec::new(),
}],
},
};

View File

@@ -86,8 +86,11 @@ Example (from OpenAI's official VSCode extension):
- `review/start` — kick off Codexs automated reviewer for a thread; responds like `turn/start` and emits `item/started`/`item/completed` notifications with `enteredReviewMode` and `exitedReviewMode` items, plus a final assistant `agentMessage` containing the review.
- `command/exec` — run a single command under the server sandbox without starting a thread/turn (handy for utilities and validation).
- `model/list` — list available models (with reasoning effort options).
- `collaborationMode/list` — list available collaboration mode presets (experimental, no pagination).
- `skills/list` — list skills for one or more `cwd` values (optional `forceReload`).
- `skills/config/write` — write user-level skill config by path.
- `mcpServer/oauth/login` — start an OAuth login for a configured MCP server; returns an `authorization_url` and later emits `mcpServer/oauthLogin/completed` once the browser flow finishes.
- `tool/requestUserInput` — prompt the user with 13 short questions for a tool call and return their answers (experimental).
- `config/mcpServer/reload` — reload MCP server config from disk and queue a refresh for loaded threads (applied on each thread's next active turn); returns `{}`. Use this after editing `config.toml` without restarting the server.
- `mcpServerStatus/list` — enumerate configured MCP servers with their tools, resources, resource templates, and auth status; supports cursor+limit pagination.
- `feedback/upload` — submit a feedback report (classification + optional reason/logs and conversation_id); returns the tracking thread id.
@@ -138,10 +141,11 @@ To branch from a stored session, call `thread/fork` with the `thread.id`. This c
### Example: List threads (with pagination & filters)
`thread/list` lets you render a history UI. Pass any combination of:
`thread/list` lets you render a history UI. Results default to `createdAt` (newest first) descending. Pass any combination of:
- `cursor` — opaque string from a prior response; omit for the first page.
- `limit` — server defaults to a reasonable page size if unset.
- `sortKey``created_at` (default) or `updated_at`.
- `modelProviders` — restrict results to specific providers; unset, null, or an empty array will include all providers.
Example:
@@ -150,11 +154,12 @@ Example:
{ "method": "thread/list", "id": 20, "params": {
"cursor": null,
"limit": 25,
"sortKey": "created_at"
} }
{ "id": 20, "result": {
"data": [
{ "id": "thr_a", "preview": "Create a TUI", "modelProvider": "openai", "createdAt": 1730831111 },
{ "id": "thr_b", "preview": "Fix tests", "modelProvider": "openai", "createdAt": 1730750000 }
{ "id": "thr_a", "preview": "Create a TUI", "modelProvider": "openai", "createdAt": 1730831111, "updatedAt": 1730831111 },
{ "id": "thr_b", "preview": "Fix tests", "modelProvider": "openai", "createdAt": 1730750000, "updatedAt": 1730750000 }
],
"nextCursor": "opaque-token-or-null"
} }
@@ -375,6 +380,7 @@ Today both notifications carry an empty `items` array even when item events were
- `commandExecution``{id, command, cwd, status, commandActions, aggregatedOutput?, exitCode?, durationMs?}` for sandboxed commands; `status` is `inProgress`, `completed`, `failed`, or `declined`.
- `fileChange``{id, changes, status}` describing proposed edits; `changes` list `{path, kind, diff}` and `status` is `inProgress`, `completed`, `failed`, or `declined`.
- `mcpToolCall``{id, server, tool, status, arguments, result?, error?}` describing MCP calls; `status` is `inProgress`, `completed`, or `failed`.
- `collabToolCall``{id, tool, status, senderThreadId, receiverThreadId?, newThreadId?, prompt?, agentStatus?}` describing collab tool calls (`spawn_agent`, `send_input`, `wait`, `close_agent`); `status` is `inProgress`, `completed`, or `failed`.
- `webSearch``{id, query}` for a web search request issued by the agent.
- `imageView``{id, path}` emitted when the agent invokes the image viewer tool.
- `enteredReviewMode``{id, review}` sent when the reviewer starts; `review` is a short user-facing label such as `"current changes"` or the requested target description.
@@ -465,8 +471,15 @@ Invoke a skill by including `$<skill-name>` in the text input. Add a `skill` inp
"params": {
"threadId": "thread-1",
"input": [
{ "type": "text", "text": "$skill-creator Add a new skill for triaging flaky CI." },
{ "type": "skill", "name": "skill-creator", "path": "/Users/me/.codex/skills/skill-creator/SKILL.md" }
{
"type": "text",
"text": "$skill-creator Add a new skill for triaging flaky CI."
},
{
"type": "skill",
"name": "skill-creator",
"path": "/Users/me/.codex/skills/skill-creator/SKILL.md"
}
]
}
}
@@ -480,20 +493,37 @@ Example:
$skill-creator Add a new skill for triaging flaky CI and include step-by-step usage.
```
Use `skills/list` to fetch the available skills (optionally scoped by `cwd` and/or with `forceReload`).
Use `skills/list` to fetch the available skills (optionally scoped by `cwds`, with `forceReload`).
```json
{ "method": "skills/list", "id": 25, "params": {
"cwd": "/Users/me/project",
"cwds": ["/Users/me/project"],
"forceReload": false
} }
{ "id": 25, "result": {
"skills": [
{ "name": "skill-creator", "description": "Create or update a Codex skill" }
]
"data": [{
"cwd": "/Users/me/project",
"skills": [
{ "name": "skill-creator", "description": "Create or update a Codex skill", "enabled": true }
],
"errors": []
}]
} }
```
To enable or disable a skill by path:
```json
{
"method": "skills/config/write",
"id": 26,
"params": {
"path": "/Users/me/.codex/skills/skill-creator/SKILL.md",
"enabled": false
}
}
```
## Auth endpoints
The 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.

View File

@@ -14,6 +14,9 @@ use codex_app_server_protocol::AgentMessageDeltaNotification;
use codex_app_server_protocol::ApplyPatchApprovalParams;
use codex_app_server_protocol::ApplyPatchApprovalResponse;
use codex_app_server_protocol::CodexErrorInfo as V2CodexErrorInfo;
use codex_app_server_protocol::CollabAgentState as V2CollabAgentStatus;
use codex_app_server_protocol::CollabAgentTool;
use codex_app_server_protocol::CollabAgentToolCallStatus as V2CollabToolCallStatus;
use codex_app_server_protocol::CommandAction as V2ParsedCommand;
use codex_app_server_protocol::CommandExecutionApprovalDecision;
use codex_app_server_protocol::CommandExecutionOutputDeltaNotification;
@@ -51,6 +54,10 @@ use codex_app_server_protocol::ThreadItem;
use codex_app_server_protocol::ThreadRollbackResponse;
use codex_app_server_protocol::ThreadTokenUsage;
use codex_app_server_protocol::ThreadTokenUsageUpdatedNotification;
use codex_app_server_protocol::ToolRequestUserInputOption;
use codex_app_server_protocol::ToolRequestUserInputParams;
use codex_app_server_protocol::ToolRequestUserInputQuestion;
use codex_app_server_protocol::ToolRequestUserInputResponse;
use codex_app_server_protocol::Turn;
use codex_app_server_protocol::TurnCompletedNotification;
use codex_app_server_protocol::TurnDiffUpdatedNotification;
@@ -80,6 +87,8 @@ use codex_core::review_prompts;
use codex_protocol::ThreadId;
use codex_protocol::plan_tool::UpdatePlanArgs;
use codex_protocol::protocol::ReviewOutputEvent;
use codex_protocol::request_user_input::RequestUserInputAnswer as CoreRequestUserInputAnswer;
use codex_protocol::request_user_input::RequestUserInputResponse as CoreRequestUserInputResponse;
use std::collections::HashMap;
use std::convert::TryFrom;
use std::path::PathBuf;
@@ -255,6 +264,57 @@ pub(crate) async fn apply_bespoke_event_handling(
});
}
},
EventMsg::RequestUserInput(request) => {
if matches!(api_version, ApiVersion::V2) {
let questions = request
.questions
.into_iter()
.map(|question| ToolRequestUserInputQuestion {
id: question.id,
header: question.header,
question: question.question,
options: question.options.map(|options| {
options
.into_iter()
.map(|option| ToolRequestUserInputOption {
label: option.label,
description: option.description,
})
.collect()
}),
})
.collect();
let params = ToolRequestUserInputParams {
thread_id: conversation_id.to_string(),
turn_id: request.turn_id,
item_id: request.call_id,
questions,
};
let rx = outgoing
.send_request(ServerRequestPayload::ToolRequestUserInput(params))
.await;
tokio::spawn(async move {
on_request_user_input_response(event_turn_id, rx, conversation).await;
});
} else {
error!(
"request_user_input is only supported on api v2 (call_id: {})",
request.call_id
);
let empty = CoreRequestUserInputResponse {
answers: HashMap::new(),
};
if let Err(err) = conversation
.submit(Op::UserInputAnswer {
id: event_turn_id,
response: empty,
})
.await
{
error!("failed to submit UserInputAnswer: {err}");
}
}
}
// TODO(celia): properly construct McpToolCall TurnItem in core.
EventMsg::McpToolCallBegin(begin_event) => {
let notification = construct_mcp_tool_call_notification(
@@ -278,6 +338,218 @@ pub(crate) async fn apply_bespoke_event_handling(
.send_server_notification(ServerNotification::ItemCompleted(notification))
.await;
}
EventMsg::CollabAgentSpawnBegin(begin_event) => {
let item = ThreadItem::CollabAgentToolCall {
id: begin_event.call_id,
tool: CollabAgentTool::SpawnAgent,
status: V2CollabToolCallStatus::InProgress,
sender_thread_id: begin_event.sender_thread_id.to_string(),
receiver_thread_ids: Vec::new(),
prompt: Some(begin_event.prompt),
agents_states: HashMap::new(),
};
let notification = ItemStartedNotification {
thread_id: conversation_id.to_string(),
turn_id: event_turn_id.clone(),
item,
};
outgoing
.send_server_notification(ServerNotification::ItemStarted(notification))
.await;
}
EventMsg::CollabAgentSpawnEnd(end_event) => {
let has_receiver = end_event.new_thread_id.is_some();
let status = match &end_event.status {
codex_protocol::protocol::AgentStatus::Errored(_)
| codex_protocol::protocol::AgentStatus::NotFound => V2CollabToolCallStatus::Failed,
_ if has_receiver => V2CollabToolCallStatus::Completed,
_ => V2CollabToolCallStatus::Failed,
};
let (receiver_thread_ids, agents_states) = match end_event.new_thread_id {
Some(id) => {
let receiver_id = id.to_string();
let received_status = V2CollabAgentStatus::from(end_event.status.clone());
(
vec![receiver_id.clone()],
[(receiver_id, received_status)].into_iter().collect(),
)
}
None => (Vec::new(), HashMap::new()),
};
let item = ThreadItem::CollabAgentToolCall {
id: end_event.call_id,
tool: CollabAgentTool::SpawnAgent,
status,
sender_thread_id: end_event.sender_thread_id.to_string(),
receiver_thread_ids,
prompt: Some(end_event.prompt),
agents_states,
};
let notification = ItemCompletedNotification {
thread_id: conversation_id.to_string(),
turn_id: event_turn_id.clone(),
item,
};
outgoing
.send_server_notification(ServerNotification::ItemCompleted(notification))
.await;
}
EventMsg::CollabAgentInteractionBegin(begin_event) => {
let receiver_thread_ids = vec![begin_event.receiver_thread_id.to_string()];
let item = ThreadItem::CollabAgentToolCall {
id: begin_event.call_id,
tool: CollabAgentTool::SendInput,
status: V2CollabToolCallStatus::InProgress,
sender_thread_id: begin_event.sender_thread_id.to_string(),
receiver_thread_ids,
prompt: Some(begin_event.prompt),
agents_states: HashMap::new(),
};
let notification = ItemStartedNotification {
thread_id: conversation_id.to_string(),
turn_id: event_turn_id.clone(),
item,
};
outgoing
.send_server_notification(ServerNotification::ItemStarted(notification))
.await;
}
EventMsg::CollabAgentInteractionEnd(end_event) => {
let status = match &end_event.status {
codex_protocol::protocol::AgentStatus::Errored(_)
| codex_protocol::protocol::AgentStatus::NotFound => V2CollabToolCallStatus::Failed,
_ => V2CollabToolCallStatus::Completed,
};
let receiver_id = end_event.receiver_thread_id.to_string();
let received_status = V2CollabAgentStatus::from(end_event.status);
let item = ThreadItem::CollabAgentToolCall {
id: end_event.call_id,
tool: CollabAgentTool::SendInput,
status,
sender_thread_id: end_event.sender_thread_id.to_string(),
receiver_thread_ids: vec![receiver_id.clone()],
prompt: Some(end_event.prompt),
agents_states: [(receiver_id, received_status)].into_iter().collect(),
};
let notification = ItemCompletedNotification {
thread_id: conversation_id.to_string(),
turn_id: event_turn_id.clone(),
item,
};
outgoing
.send_server_notification(ServerNotification::ItemCompleted(notification))
.await;
}
EventMsg::CollabWaitingBegin(begin_event) => {
let receiver_thread_ids = begin_event
.receiver_thread_ids
.iter()
.map(ToString::to_string)
.collect();
let item = ThreadItem::CollabAgentToolCall {
id: begin_event.call_id,
tool: CollabAgentTool::Wait,
status: V2CollabToolCallStatus::InProgress,
sender_thread_id: begin_event.sender_thread_id.to_string(),
receiver_thread_ids,
prompt: None,
agents_states: HashMap::new(),
};
let notification = ItemStartedNotification {
thread_id: conversation_id.to_string(),
turn_id: event_turn_id.clone(),
item,
};
outgoing
.send_server_notification(ServerNotification::ItemStarted(notification))
.await;
}
EventMsg::CollabWaitingEnd(end_event) => {
let status = if end_event.statuses.values().any(|status| {
matches!(
status,
codex_protocol::protocol::AgentStatus::Errored(_)
| codex_protocol::protocol::AgentStatus::NotFound
)
}) {
V2CollabToolCallStatus::Failed
} else {
V2CollabToolCallStatus::Completed
};
let receiver_thread_ids = end_event.statuses.keys().map(ToString::to_string).collect();
let agents_states = end_event
.statuses
.iter()
.map(|(id, status)| (id.to_string(), V2CollabAgentStatus::from(status.clone())))
.collect();
let item = ThreadItem::CollabAgentToolCall {
id: end_event.call_id,
tool: CollabAgentTool::Wait,
status,
sender_thread_id: end_event.sender_thread_id.to_string(),
receiver_thread_ids,
prompt: None,
agents_states,
};
let notification = ItemCompletedNotification {
thread_id: conversation_id.to_string(),
turn_id: event_turn_id.clone(),
item,
};
outgoing
.send_server_notification(ServerNotification::ItemCompleted(notification))
.await;
}
EventMsg::CollabCloseBegin(begin_event) => {
let item = ThreadItem::CollabAgentToolCall {
id: begin_event.call_id,
tool: CollabAgentTool::CloseAgent,
status: V2CollabToolCallStatus::InProgress,
sender_thread_id: begin_event.sender_thread_id.to_string(),
receiver_thread_ids: vec![begin_event.receiver_thread_id.to_string()],
prompt: None,
agents_states: HashMap::new(),
};
let notification = ItemStartedNotification {
thread_id: conversation_id.to_string(),
turn_id: event_turn_id.clone(),
item,
};
outgoing
.send_server_notification(ServerNotification::ItemStarted(notification))
.await;
}
EventMsg::CollabCloseEnd(end_event) => {
let status = match &end_event.status {
codex_protocol::protocol::AgentStatus::Errored(_)
| codex_protocol::protocol::AgentStatus::NotFound => V2CollabToolCallStatus::Failed,
_ => V2CollabToolCallStatus::Completed,
};
let receiver_id = end_event.receiver_thread_id.to_string();
let agents_states = [(
receiver_id.clone(),
V2CollabAgentStatus::from(end_event.status),
)]
.into_iter()
.collect();
let item = ThreadItem::CollabAgentToolCall {
id: end_event.call_id,
tool: CollabAgentTool::CloseAgent,
status,
sender_thread_id: end_event.sender_thread_id.to_string(),
receiver_thread_ids: vec![receiver_id],
prompt: None,
agents_states,
};
let notification = ItemCompletedNotification {
thread_id: conversation_id.to_string(),
turn_id: event_turn_id.clone(),
item,
};
outgoing
.send_server_notification(ServerNotification::ItemCompleted(notification))
.await;
}
EventMsg::AgentMessageContentDelta(event) => {
let notification = AgentMessageDeltaNotification {
thread_id: conversation_id.to_string(),
@@ -1132,6 +1404,66 @@ async fn on_exec_approval_response(
}
}
async fn on_request_user_input_response(
event_turn_id: String,
receiver: oneshot::Receiver<JsonValue>,
conversation: Arc<CodexThread>,
) {
let response = receiver.await;
let value = match response {
Ok(value) => value,
Err(err) => {
error!("request failed: {err:?}");
let empty = CoreRequestUserInputResponse {
answers: HashMap::new(),
};
if let Err(err) = conversation
.submit(Op::UserInputAnswer {
id: event_turn_id,
response: empty,
})
.await
{
error!("failed to submit UserInputAnswer: {err}");
}
return;
}
};
let response =
serde_json::from_value::<ToolRequestUserInputResponse>(value).unwrap_or_else(|err| {
error!("failed to deserialize ToolRequestUserInputResponse: {err}");
ToolRequestUserInputResponse {
answers: HashMap::new(),
}
});
let response = CoreRequestUserInputResponse {
answers: response
.answers
.into_iter()
.map(|(id, answer)| {
(
id,
CoreRequestUserInputAnswer {
selected: answer.selected,
other: answer.other,
},
)
})
.collect(),
};
if let Err(err) = conversation
.submit(Op::UserInputAnswer {
id: event_turn_id,
response,
})
.await
{
error!("failed to submit UserInputAnswer: {err}");
}
}
const REVIEW_FALLBACK_MESSAGE: &str = "Reviewer failed to output a response.";
fn render_review_output_text(output: &ReviewOutputEvent) -> String {

View File

@@ -6,6 +6,7 @@ use crate::models::supported_models;
use crate::outgoing_message::OutgoingMessageSender;
use crate::outgoing_message::OutgoingNotification;
use chrono::DateTime;
use chrono::SecondsFormat;
use chrono::Utc;
use codex_app_server_protocol::Account;
use codex_app_server_protocol::AccountLoginCompletedNotification;
@@ -22,6 +23,8 @@ use codex_app_server_protocol::CancelLoginAccountResponse;
use codex_app_server_protocol::CancelLoginAccountStatus;
use codex_app_server_protocol::CancelLoginChatGptResponse;
use codex_app_server_protocol::ClientRequest;
use codex_app_server_protocol::CollaborationModeListParams;
use codex_app_server_protocol::CollaborationModeListResponse;
use codex_app_server_protocol::CommandExecParams;
use codex_app_server_protocol::ConversationGitInfo;
use codex_app_server_protocol::ConversationSummary;
@@ -84,6 +87,8 @@ use codex_app_server_protocol::ServerNotification;
use codex_app_server_protocol::SessionConfiguredNotification;
use codex_app_server_protocol::SetDefaultModelParams;
use codex_app_server_protocol::SetDefaultModelResponse;
use codex_app_server_protocol::SkillsConfigWriteParams;
use codex_app_server_protocol::SkillsConfigWriteResponse;
use codex_app_server_protocol::SkillsListParams;
use codex_app_server_protocol::SkillsListResponse;
use codex_app_server_protocol::Thread;
@@ -99,6 +104,7 @@ use codex_app_server_protocol::ThreadLoadedListResponse;
use codex_app_server_protocol::ThreadResumeParams;
use codex_app_server_protocol::ThreadResumeResponse;
use codex_app_server_protocol::ThreadRollbackParams;
use codex_app_server_protocol::ThreadSortKey;
use codex_app_server_protocol::ThreadStartParams;
use codex_app_server_protocol::ThreadStartResponse;
use codex_app_server_protocol::ThreadStartedNotification;
@@ -123,11 +129,13 @@ use codex_core::NewThread;
use codex_core::RolloutRecorder;
use codex_core::SessionMeta;
use codex_core::ThreadManager;
use codex_core::ThreadSortKey as CoreThreadSortKey;
use codex_core::auth::CLIENT_ID;
use codex_core::auth::login_with_api_key;
use codex_core::config::Config;
use codex_core::config::ConfigOverrides;
use codex_core::config::ConfigService;
use codex_core::config::edit::ConfigEdit;
use codex_core::config::edit::ConfigEditsBuilder;
use codex_core::config::types::McpServerTransportConfig;
use codex_core::default_client::get_codex_user_agent;
@@ -178,6 +186,7 @@ use std::sync::atomic::Ordering;
use std::time::Duration;
use tokio::select;
use tokio::sync::Mutex;
use tokio::sync::broadcast;
use tokio::sync::oneshot;
use toml::Value as TomlValue;
use tracing::error;
@@ -229,6 +238,7 @@ pub(crate) struct CodexMessageProcessor {
config: Arc<Config>,
cli_overrides: Vec<(String, TomlValue)>,
conversation_listeners: HashMap<Uuid, oneshot::Sender<()>>,
listener_thread_ids_by_subscription: HashMap<Uuid, ThreadId>,
active_login: Arc<Mutex<Option<ActiveLogin>>>,
// Queue of pending interrupt requests per conversation. We reply when TurnAborted arrives.
pending_interrupts: PendingInterrupts,
@@ -286,6 +296,7 @@ impl CodexMessageProcessor {
config,
cli_overrides,
conversation_listeners: HashMap::new(),
listener_thread_ids_by_subscription: HashMap::new(),
active_login: Arc::new(Mutex::new(None)),
pending_interrupts: Arc::new(Mutex::new(HashMap::new())),
pending_rollbacks: Arc::new(Mutex::new(HashMap::new())),
@@ -394,6 +405,9 @@ impl CodexMessageProcessor {
ClientRequest::SkillsList { request_id, params } => {
self.skills_list(request_id, params).await;
}
ClientRequest::SkillsConfigWrite { request_id, params } => {
self.skills_config_write(request_id, params).await;
}
ClientRequest::TurnStart { request_id, params } => {
self.turn_start(request_id, params).await;
}
@@ -424,6 +438,15 @@ impl CodexMessageProcessor {
Self::list_models(outgoing, thread_manager, config, request_id, params).await;
});
}
ClientRequest::CollaborationModeList { request_id, params } => {
let outgoing = self.outgoing.clone();
let thread_manager = self.thread_manager.clone();
tokio::spawn(async move {
Self::list_collaboration_modes(outgoing, thread_manager, request_id, params)
.await;
});
}
ClientRequest::McpServerOauthLogin { request_id, params } => {
self.mcp_server_oauth_login(request_id, params).await;
}
@@ -1271,7 +1294,11 @@ impl CodexMessageProcessor {
});
}
async fn process_new_conversation(&self, request_id: RequestId, params: NewConversationParams) {
async fn process_new_conversation(
&mut self,
request_id: RequestId,
params: NewConversationParams,
) {
let NewConversationParams {
model,
model_provider,
@@ -1591,6 +1618,7 @@ impl CodexMessageProcessor {
let ThreadListParams {
cursor,
limit,
sort_key,
model_providers,
} = params;
@@ -1598,8 +1626,12 @@ impl CodexMessageProcessor {
.map(|value| value as usize)
.unwrap_or(THREAD_LIST_DEFAULT_LIMIT)
.clamp(1, THREAD_LIST_MAX_LIMIT);
let core_sort_key = match sort_key.unwrap_or(ThreadSortKey::CreatedAt) {
ThreadSortKey::CreatedAt => CoreThreadSortKey::CreatedAt,
ThreadSortKey::UpdatedAt => CoreThreadSortKey::UpdatedAt,
};
let (summaries, next_cursor) = match self
.list_threads_common(requested_page_size, cursor, model_providers)
.list_threads_common(requested_page_size, cursor, model_providers, core_sort_key)
.await
{
Ok(r) => r,
@@ -1669,6 +1701,31 @@ impl CodexMessageProcessor {
self.outgoing.send_response(request_id, response).await;
}
pub(crate) fn thread_created_receiver(&self) -> broadcast::Receiver<ThreadId> {
self.thread_manager.subscribe_thread_created()
}
/// Best-effort: attach a listener for thread_id if missing.
pub(crate) async fn try_attach_thread_listener(&mut self, thread_id: ThreadId) {
if self
.listener_thread_ids_by_subscription
.values()
.any(|entry| *entry == thread_id)
{
return;
}
if let Err(err) = self
.attach_conversation_listener(thread_id, false, ApiVersion::V2)
.await
{
warn!(
"failed to attach listener for thread {thread_id}: {message}",
message = err.message
);
}
}
async fn thread_resume(&mut self, request_id: RequestId, params: ThreadResumeParams) {
let ThreadResumeParams {
thread_id,
@@ -2139,7 +2196,12 @@ impl CodexMessageProcessor {
.clamp(1, THREAD_LIST_MAX_LIMIT);
match self
.list_threads_common(requested_page_size, cursor, model_providers)
.list_threads_common(
requested_page_size,
cursor,
model_providers,
CoreThreadSortKey::UpdatedAt,
)
.await
{
Ok((items, next_cursor)) => {
@@ -2157,8 +2219,18 @@ impl CodexMessageProcessor {
requested_page_size: usize,
cursor: Option<String>,
model_providers: Option<Vec<String>>,
sort_key: CoreThreadSortKey,
) -> Result<(Vec<ConversationSummary>, Option<String>), JSONRPCErrorError> {
let mut cursor_obj: Option<RolloutCursor> = cursor.as_ref().and_then(|s| parse_cursor(s));
let mut cursor_obj: Option<RolloutCursor> = match cursor.as_ref() {
Some(cursor_str) => {
Some(parse_cursor(cursor_str).ok_or_else(|| JSONRPCErrorError {
code: INVALID_REQUEST_ERROR_CODE,
message: format!("invalid cursor: {cursor_str}"),
data: None,
})?)
}
None => None,
};
let mut last_cursor = cursor_obj.clone();
let mut remaining = requested_page_size;
let mut items = Vec::with_capacity(requested_page_size);
@@ -2182,6 +2254,7 @@ impl CodexMessageProcessor {
&self.config.codex_home,
page_size,
cursor_obj.as_ref(),
sort_key,
INTERACTIVE_SESSION_SOURCES,
model_provider_filter.as_deref(),
fallback_provider.as_str(),
@@ -2197,6 +2270,7 @@ impl CodexMessageProcessor {
.items
.into_iter()
.filter_map(|it| {
let updated_at = it.updated_at.clone();
let session_meta_line = it.head.first().and_then(|first| {
serde_json::from_value::<SessionMetaLine>(first.clone()).ok()
})?;
@@ -2206,6 +2280,7 @@ impl CodexMessageProcessor {
&session_meta_line.meta,
session_meta_line.git.as_ref(),
fallback_provider.as_str(),
updated_at,
)
})
.collect::<Vec<_>>();
@@ -2307,6 +2382,18 @@ impl CodexMessageProcessor {
outgoing.send_response(request_id, response).await;
}
async fn list_collaboration_modes(
outgoing: Arc<OutgoingMessageSender>,
thread_manager: Arc<ThreadManager>,
request_id: RequestId,
params: CollaborationModeListParams,
) {
let CollaborationModeListParams {} = params;
let items = thread_manager.list_collaboration_modes();
let response = CollaborationModeListResponse { data: items };
outgoing.send_response(request_id, response).await;
}
async fn mcp_server_refresh(&self, request_id: RequestId, _params: Option<()>) {
let config = match self.load_latest_config().await {
Ok(config) => config,
@@ -3093,7 +3180,13 @@ impl CodexMessageProcessor {
let mapped_items: Vec<CoreInputItem> = items
.into_iter()
.map(|item| match item {
WireInputItem::Text { text } => CoreInputItem::Text { text },
WireInputItem::Text {
text,
text_elements,
} => CoreInputItem::Text {
text,
text_elements: text_elements.into_iter().map(Into::into).collect(),
},
WireInputItem::Image { image_url } => CoreInputItem::Image { image_url },
WireInputItem::LocalImage { path } => CoreInputItem::LocalImage { path },
})
@@ -3139,7 +3232,13 @@ impl CodexMessageProcessor {
let mapped_items: Vec<CoreInputItem> = items
.into_iter()
.map(|item| match item {
WireInputItem::Text { text } => CoreInputItem::Text { text },
WireInputItem::Text {
text,
text_elements,
} => CoreInputItem::Text {
text,
text_elements: text_elements.into_iter().map(Into::into).collect(),
},
WireInputItem::Image { image_url } => CoreInputItem::Image { image_url },
WireInputItem::LocalImage { path } => CoreInputItem::LocalImage { path },
})
@@ -3155,6 +3254,7 @@ impl CodexMessageProcessor {
effort,
summary,
final_output_json_schema: output_schema,
collaboration_mode: None,
})
.await;
@@ -3176,7 +3276,7 @@ impl CodexMessageProcessor {
for cwd in cwds {
let outcome = skills_manager.skills_for_cwd(&cwd, force_reload).await;
let errors = errors_to_info(&outcome.errors);
let skills = skills_to_info(&outcome.skills);
let skills = skills_to_info(&outcome.skills, &outcome.disabled_paths);
data.push(codex_app_server_protocol::SkillsListEntry {
cwd,
skills,
@@ -3188,6 +3288,37 @@ impl CodexMessageProcessor {
.await;
}
async fn skills_config_write(&self, request_id: RequestId, params: SkillsConfigWriteParams) {
let SkillsConfigWriteParams { path, enabled } = params;
let edits = vec![ConfigEdit::SetSkillConfig { path, enabled }];
let result = ConfigEditsBuilder::new(&self.config.codex_home)
.with_edits(edits)
.apply()
.await;
match result {
Ok(()) => {
self.thread_manager.skills_manager().clear_cache();
self.outgoing
.send_response(
request_id,
SkillsConfigWriteResponse {
effective_enabled: enabled,
},
)
.await;
}
Err(err) => {
let error = JSONRPCErrorError {
code: INTERNAL_ERROR_CODE,
message: format!("failed to update skill settings: {err}"),
data: None,
};
self.outgoing.send_error(request_id, error).await;
}
}
}
async fn interrupt_conversation(
&mut self,
request_id: RequestId,
@@ -3237,7 +3368,8 @@ impl CodexMessageProcessor {
|| params.sandbox_policy.is_some()
|| params.model.is_some()
|| params.effort.is_some()
|| params.summary.is_some();
|| params.summary.is_some()
|| params.collaboration_mode.is_some();
// If any overrides are provided, update the session turn context first.
if has_any_overrides {
@@ -3249,6 +3381,7 @@ impl CodexMessageProcessor {
model: params.model,
effort: params.effort.map(Some),
summary: params.summary,
collaboration_mode: params.collaboration_mode,
})
.await;
}
@@ -3301,6 +3434,8 @@ impl CodexMessageProcessor {
id: turn_id.clone(),
content: vec![V2UserInput::Text {
text: display_text.to_string(),
// Review prompt display text is synthesized; no UI element ranges to preserve.
text_elements: Vec::new(),
}],
}]
};
@@ -3389,7 +3524,9 @@ impl CodexMessageProcessor {
})?;
let mut config = self.config.as_ref().clone();
config.model = Some(self.config.review_model.clone());
if let Some(review_model) = &config.review_model {
config.model = Some(review_model.clone());
}
let NewThread {
thread_id,
@@ -3563,6 +3700,12 @@ impl CodexMessageProcessor {
Some(sender) => {
// Signal the spawned task to exit and acknowledge.
let _ = sender.send(());
if let Some(thread_id) = self
.listener_thread_ids_by_subscription
.remove(&subscription_id)
{
info!("removed listener for thread {thread_id}");
}
let response = RemoveConversationSubscriptionResponse {};
self.outgoing.send_response(request_id, response).await;
}
@@ -3598,6 +3741,8 @@ impl CodexMessageProcessor {
let (cancel_tx, mut cancel_rx) = oneshot::channel();
self.conversation_listeners
.insert(subscription_id, cancel_tx);
self.listener_thread_ids_by_subscription
.insert(subscription_id, conversation_id);
let outgoing_for_task = self.outgoing.clone();
let pending_interrupts = self.pending_interrupts.clone();
@@ -3836,15 +3981,30 @@ impl CodexMessageProcessor {
fn skills_to_info(
skills: &[codex_core::skills::SkillMetadata],
disabled_paths: &std::collections::HashSet<PathBuf>,
) -> Vec<codex_app_server_protocol::SkillMetadata> {
skills
.iter()
.map(|skill| codex_app_server_protocol::SkillMetadata {
name: skill.name.clone(),
description: skill.description.clone(),
short_description: skill.short_description.clone(),
path: skill.path.clone(),
scope: skill.scope.into(),
.map(|skill| {
let enabled = !disabled_paths.contains(&skill.path);
codex_app_server_protocol::SkillMetadata {
name: skill.name.clone(),
description: skill.description.clone(),
short_description: skill.short_description.clone(),
interface: skill.interface.clone().map(|interface| {
codex_app_server_protocol::SkillInterface {
display_name: interface.display_name,
short_description: interface.short_description,
icon_small: interface.icon_small,
icon_large: interface.icon_large,
brand_color: interface.brand_color,
default_prompt: interface.default_prompt,
}
}),
path: skill.path.clone(),
scope: skill.scope.into(),
enabled,
}
})
.collect()
}
@@ -3916,12 +4076,19 @@ pub(crate) async fn read_summary_from_rollout(
git,
} = session_meta_line;
let created_at = if session_meta.timestamp.is_empty() {
None
} else {
Some(session_meta.timestamp.as_str())
};
let updated_at = read_updated_at(path, created_at).await;
if let Some(summary) = extract_conversation_summary(
path.to_path_buf(),
&head,
&session_meta,
git.as_ref(),
fallback_provider,
updated_at.clone(),
) {
return Ok(summary);
}
@@ -3936,10 +4103,12 @@ pub(crate) async fn read_summary_from_rollout(
.clone()
.unwrap_or_else(|| fallback_provider.to_string());
let git_info = git.as_ref().map(map_git_info);
let updated_at = updated_at.or_else(|| timestamp.clone());
Ok(ConversationSummary {
conversation_id: session_meta.id,
timestamp,
updated_at,
path: path.to_path_buf(),
preview: String::new(),
model_provider,
@@ -3974,6 +4143,7 @@ fn extract_conversation_summary(
session_meta: &SessionMeta,
git: Option<&CoreGitInfo>,
fallback_provider: &str,
updated_at: Option<String>,
) -> Option<ConversationSummary> {
let preview = head
.iter()
@@ -3999,10 +4169,12 @@ fn extract_conversation_summary(
.clone()
.unwrap_or_else(|| fallback_provider.to_string());
let git_info = git.map(map_git_info);
let updated_at = updated_at.or_else(|| timestamp.clone());
Some(ConversationSummary {
conversation_id,
timestamp,
updated_at,
path,
preview: preview.to_string(),
model_provider,
@@ -4029,12 +4201,25 @@ fn parse_datetime(timestamp: Option<&str>) -> Option<DateTime<Utc>> {
})
}
async fn read_updated_at(path: &Path, created_at: Option<&str>) -> Option<String> {
let updated_at = tokio::fs::metadata(path)
.await
.ok()
.and_then(|meta| meta.modified().ok())
.map(|modified| {
let updated_at: DateTime<Utc> = modified.into();
updated_at.to_rfc3339_opts(SecondsFormat::Secs, true)
});
updated_at.or_else(|| created_at.map(str::to_string))
}
pub(crate) fn summary_to_thread(summary: ConversationSummary) -> Thread {
let ConversationSummary {
conversation_id,
path,
preview,
timestamp,
updated_at,
model_provider,
cwd,
cli_version,
@@ -4043,6 +4228,7 @@ pub(crate) fn summary_to_thread(summary: ConversationSummary) -> Thread {
} = summary;
let created_at = parse_datetime(timestamp.as_deref());
let updated_at = parse_datetime(updated_at.as_deref()).or(created_at);
let git_info = git_info.map(|info| ApiGitInfo {
sha: info.sha,
branch: info.branch,
@@ -4054,6 +4240,7 @@ pub(crate) fn summary_to_thread(summary: ConversationSummary) -> Thread {
preview,
model_provider,
created_at: created_at.map(|dt| dt.timestamp()).unwrap_or(0),
updated_at: updated_at.map(|dt| dt.timestamp()).unwrap_or(0),
path,
cwd,
cli_version,
@@ -4085,7 +4272,6 @@ mod tests {
"cwd": "/",
"originator": "codex",
"cli_version": "0.0.0",
"instructions": null,
"model_provider": "test-provider"
}),
json!({
@@ -4108,13 +4294,20 @@ mod tests {
let session_meta = serde_json::from_value::<SessionMeta>(head[0].clone())?;
let summary =
extract_conversation_summary(path.clone(), &head, &session_meta, None, "test-provider")
.expect("summary");
let summary = extract_conversation_summary(
path.clone(),
&head,
&session_meta,
None,
"test-provider",
timestamp.clone(),
)
.expect("summary");
let expected = ConversationSummary {
conversation_id,
timestamp,
timestamp: timestamp.clone(),
updated_at: timestamp,
path,
preview: "Count to 5".to_string(),
model_provider: "test-provider".to_string(),
@@ -4134,6 +4327,7 @@ mod tests {
use codex_protocol::protocol::RolloutLine;
use codex_protocol::protocol::SessionMetaLine;
use std::fs;
use std::fs::FileTimes;
let temp_dir = TempDir::new()?;
let path = temp_dir.path().join("rollout.jsonl");
@@ -4157,12 +4351,19 @@ mod tests {
};
fs::write(&path, format!("{}\n", serde_json::to_string(&line)?))?;
let parsed = chrono::DateTime::parse_from_rfc3339(&timestamp)?.with_timezone(&Utc);
let times = FileTimes::new().set_modified(parsed.into());
std::fs::OpenOptions::new()
.append(true)
.open(&path)?
.set_times(times)?;
let summary = read_summary_from_rollout(path.as_path(), "fallback").await?;
let expected = ConversationSummary {
conversation_id,
timestamp: Some(timestamp),
timestamp: Some(timestamp.clone()),
updated_at: Some("2025-09-05T16:53:11Z".to_string()),
path: path.clone(),
preview: String::new(),
model_provider: "fallback".to_string(),

View File

@@ -135,7 +135,7 @@ mod tests {
CoreSandboxModeRequirement::ReadOnly,
CoreSandboxModeRequirement::ExternalSandbox,
]),
mcp_server_requirements: None,
mcp_servers: None,
};
let mapped = map_requirements_toml_to_api(requirements);

View File

@@ -1,6 +1,7 @@
#![deny(clippy::print_stdout, clippy::print_stderr)]
use codex_common::CliConfigOverrides;
use codex_core::config::Config;
use codex_core::config::ConfigBuilder;
use codex_core::config_loader::LoaderOverrides;
use std::io::ErrorKind;
@@ -10,7 +11,9 @@ use std::path::PathBuf;
use crate::message_processor::MessageProcessor;
use crate::outgoing_message::OutgoingMessage;
use crate::outgoing_message::OutgoingMessageSender;
use codex_app_server_protocol::ConfigWarningNotification;
use codex_app_server_protocol::JSONRPCMessage;
use codex_core::check_execpolicy_for_warnings;
use codex_feedback::CodexFeedback;
use tokio::io::AsyncBufReadExt;
use tokio::io::AsyncWriteExt;
@@ -21,6 +24,7 @@ use toml::Value as TomlValue;
use tracing::debug;
use tracing::error;
use tracing::info;
use tracing::warn;
use tracing_subscriber::EnvFilter;
use tracing_subscriber::Layer;
use tracing_subscriber::layer::SubscriberExt;
@@ -82,14 +86,38 @@ pub async fn run_main(
)
})?;
let loader_overrides_for_config_api = loader_overrides.clone();
let config = ConfigBuilder::default()
let mut config_warnings = Vec::new();
let config = match ConfigBuilder::default()
.cli_overrides(cli_kv_overrides.clone())
.loader_overrides(loader_overrides)
.build()
.await
.map_err(|e| {
std::io::Error::new(ErrorKind::InvalidData, format!("error loading config: {e}"))
})?;
{
Ok(config) => config,
Err(err) => {
let message = ConfigWarningNotification {
summary: "Invalid configuration; using defaults.".to_string(),
details: Some(err.to_string()),
};
config_warnings.push(message);
Config::load_default_with_cli_overrides(cli_kv_overrides.clone()).map_err(|e| {
std::io::Error::new(
ErrorKind::InvalidData,
format!("error loading default config after config error: {e}"),
)
})?
}
};
if let Ok(Some(err)) =
check_execpolicy_for_warnings(&config.features, &config.config_layer_stack).await
{
let message = ConfigWarningNotification {
summary: "Error parsing rules; custom rules not applied.".to_string(),
details: Some(err.to_string()),
};
config_warnings.push(message);
}
let feedback = CodexFeedback::new();
@@ -127,6 +155,12 @@ pub async fn run_main(
.with(otel_logger_layer)
.with(otel_tracing_layer)
.try_init();
for warning in &config_warnings {
match &warning.details {
Some(details) => error!("{} {}", warning.summary, details),
None => error!("{}", warning.summary),
}
}
// Task: process incoming messages.
let processor_handle = tokio::spawn({
@@ -140,14 +174,41 @@ pub async fn run_main(
cli_overrides,
loader_overrides,
feedback.clone(),
config_warnings,
);
let mut thread_created_rx = processor.thread_created_receiver();
async move {
while let Some(msg) = incoming_rx.recv().await {
match msg {
JSONRPCMessage::Request(r) => processor.process_request(r).await,
JSONRPCMessage::Response(r) => processor.process_response(r).await,
JSONRPCMessage::Notification(n) => processor.process_notification(n).await,
JSONRPCMessage::Error(e) => processor.process_error(e),
let mut listen_for_threads = true;
loop {
tokio::select! {
msg = incoming_rx.recv() => {
let Some(msg) = msg else {
break;
};
match msg {
JSONRPCMessage::Request(r) => processor.process_request(r).await,
JSONRPCMessage::Response(r) => processor.process_response(r).await,
JSONRPCMessage::Notification(n) => processor.process_notification(n).await,
JSONRPCMessage::Error(e) => processor.process_error(e),
}
}
created = thread_created_rx.recv(), if listen_for_threads => {
match created {
Ok(thread_id) => {
processor.try_attach_thread_listener(thread_id).await;
}
Err(tokio::sync::broadcast::error::RecvError::Lagged(_)) => {
// TODO(jif) handle lag.
// Assumes thread creation volume is low enough that lag never happens.
// If it does, we log and continue without resyncing to avoid attaching
// listeners for threads that should remain unsubscribed.
warn!("thread_created receiver lagged; skipping resync");
}
Err(tokio::sync::broadcast::error::RecvError::Closed) => {
listen_for_threads = false;
}
}
}
}
}

View File

@@ -10,6 +10,7 @@ use codex_app_server_protocol::ClientRequest;
use codex_app_server_protocol::ConfigBatchWriteParams;
use codex_app_server_protocol::ConfigReadParams;
use codex_app_server_protocol::ConfigValueWriteParams;
use codex_app_server_protocol::ConfigWarningNotification;
use codex_app_server_protocol::InitializeResponse;
use codex_app_server_protocol::JSONRPCError;
use codex_app_server_protocol::JSONRPCErrorError;
@@ -17,6 +18,7 @@ use codex_app_server_protocol::JSONRPCNotification;
use codex_app_server_protocol::JSONRPCRequest;
use codex_app_server_protocol::JSONRPCResponse;
use codex_app_server_protocol::RequestId;
use codex_app_server_protocol::ServerNotification;
use codex_core::AuthManager;
use codex_core::ThreadManager;
use codex_core::config::Config;
@@ -26,7 +28,9 @@ use codex_core::default_client::USER_AGENT_SUFFIX;
use codex_core::default_client::get_codex_user_agent;
use codex_core::default_client::set_default_originator;
use codex_feedback::CodexFeedback;
use codex_protocol::ThreadId;
use codex_protocol::protocol::SessionSource;
use tokio::sync::broadcast;
use toml::Value as TomlValue;
pub(crate) struct MessageProcessor {
@@ -34,6 +38,7 @@ pub(crate) struct MessageProcessor {
codex_message_processor: CodexMessageProcessor,
config_api: ConfigApi,
initialized: bool,
config_warnings: Vec<ConfigWarningNotification>,
}
impl MessageProcessor {
@@ -46,6 +51,7 @@ impl MessageProcessor {
cli_overrides: Vec<(String, TomlValue)>,
loader_overrides: LoaderOverrides,
feedback: CodexFeedback,
config_warnings: Vec<ConfigWarningNotification>,
) -> Self {
let outgoing = Arc::new(outgoing);
let auth_manager = AuthManager::shared(
@@ -74,6 +80,7 @@ impl MessageProcessor {
codex_message_processor,
config_api,
initialized: false,
config_warnings,
}
}
@@ -154,6 +161,15 @@ impl MessageProcessor {
self.outgoing.send_response(request_id, response).await;
self.initialized = true;
if !self.config_warnings.is_empty() {
for notification in self.config_warnings.drain(..) {
self.outgoing
.send_server_notification(ServerNotification::ConfigWarning(
notification,
))
.await;
}
}
return;
}
@@ -199,6 +215,19 @@ impl MessageProcessor {
tracing::info!("<- notification: {:?}", notification);
}
pub(crate) fn thread_created_receiver(&self) -> broadcast::Receiver<ThreadId> {
self.codex_message_processor.thread_created_receiver()
}
pub(crate) async fn try_attach_thread_listener(&mut self, thread_id: ThreadId) {
if !self.initialized {
return;
}
self.codex_message_processor
.try_attach_thread_listener(thread_id)
.await;
}
/// Handle a standalone JSON-RPC response originating from the peer.
pub(crate) async fn process_response(&mut self, response: JSONRPCResponse) {
tracing::info!("<- response: {:?}", response);

View File

@@ -4,12 +4,13 @@ use codex_app_server_protocol::Model;
use codex_app_server_protocol::ReasoningEffortOption;
use codex_core::ThreadManager;
use codex_core::config::Config;
use codex_core::models_manager::manager::RefreshStrategy;
use codex_protocol::openai_models::ModelPreset;
use codex_protocol::openai_models::ReasoningEffortPreset;
pub async fn supported_models(thread_manager: Arc<ThreadManager>, config: &Config) -> Vec<Model> {
thread_manager
.list_models(config)
.list_models(config, RefreshStrategy::OnlineIfUncached)
.await
.into_iter()
.filter(|preset| preset.show_in_picker)

View File

@@ -162,6 +162,7 @@ mod tests {
use codex_app_server_protocol::AccountRateLimitsUpdatedNotification;
use codex_app_server_protocol::AccountUpdatedNotification;
use codex_app_server_protocol::AuthMode;
use codex_app_server_protocol::ConfigWarningNotification;
use codex_app_server_protocol::LoginChatGptCompleteNotification;
use codex_app_server_protocol::RateLimitSnapshot;
use codex_app_server_protocol::RateLimitWindow;
@@ -279,4 +280,26 @@ mod tests {
"ensure the notification serializes correctly"
);
}
#[test]
fn verify_config_warning_notification_serialization() {
let notification = ServerNotification::ConfigWarning(ConfigWarningNotification {
summary: "Config error: using defaults".to_string(),
details: Some("error loading config: bad config".to_string()),
});
let jsonrpc_notification = OutgoingMessage::AppServerNotification(notification);
assert_eq!(
json!( {
"method": "configWarning",
"params": {
"summary": "Config error: using defaults",
"details": "error loading config: bad config",
},
}),
serde_json::to_value(jsonrpc_notification)
.expect("ensure the notification serializes correctly"),
"ensure the notification serializes correctly"
);
}
}

View File

@@ -27,8 +27,11 @@ pub use models_cache::write_models_cache_with_models;
pub use responses::create_apply_patch_sse_response;
pub use responses::create_exec_command_sse_response;
pub use responses::create_final_assistant_message_sse_response;
pub use responses::create_request_user_input_sse_response;
pub use responses::create_shell_command_sse_response;
pub use rollout::create_fake_rollout;
pub use rollout::create_fake_rollout_with_text_elements;
pub use rollout::rollout_path;
use serde::de::DeserializeOwned;
pub fn to_response<T: DeserializeOwned>(response: JSONRPCResponse) -> anyhow::Result<T> {

View File

@@ -17,6 +17,7 @@ use codex_app_server_protocol::CancelLoginAccountParams;
use codex_app_server_protocol::CancelLoginChatGptParams;
use codex_app_server_protocol::ClientInfo;
use codex_app_server_protocol::ClientNotification;
use codex_app_server_protocol::CollaborationModeListParams;
use codex_app_server_protocol::ConfigBatchWriteParams;
use codex_app_server_protocol::ConfigReadParams;
use codex_app_server_protocol::ConfigValueWriteParams;
@@ -396,6 +397,15 @@ impl McpProcess {
self.send_request("model/list", params).await
}
/// Send a `collaborationMode/list` JSON-RPC request.
pub async fn send_list_collaboration_modes_request(
&mut self,
params: CollaborationModeListParams,
) -> anyhow::Result<i64> {
let params = Some(serde_json::to_value(params)?);
self.send_request("collaborationMode/list", params).await
}
/// Send a `resumeConversation` JSON-RPC request.
pub async fn send_resume_conversation_request(
&mut self,

View File

@@ -25,7 +25,7 @@ fn preset_to_info(preset: &ModelPreset, priority: i32) -> ModelInfo {
},
supported_in_api: true,
priority,
upgrade: preset.upgrade.as_ref().map(|u| u.id.clone()),
upgrade: preset.upgrade.as_ref().map(|u| u.into()),
base_instructions: "base instructions".to_string(),
supports_reasoning_summaries: false,
support_verbosity: false,

View File

@@ -60,3 +60,26 @@ pub fn create_exec_command_sse_response(call_id: &str) -> anyhow::Result<String>
responses::ev_completed("resp-1"),
]))
}
pub fn create_request_user_input_sse_response(call_id: &str) -> anyhow::Result<String> {
let tool_call_arguments = serde_json::to_string(&json!({
"questions": [{
"id": "confirm_path",
"header": "Confirm",
"question": "Proceed with the plan?",
"options": [{
"label": "Yes (Recommended)",
"description": "Continue the current plan."
}, {
"label": "No",
"description": "Stop and revisit the approach."
}]
}]
}))?;
Ok(responses::sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_function_call(call_id, "request_user_input", &tool_call_arguments),
responses::ev_completed("resp-1"),
]))
}

View File

@@ -6,10 +6,23 @@ use codex_protocol::protocol::SessionMetaLine;
use codex_protocol::protocol::SessionSource;
use serde_json::json;
use std::fs;
use std::fs::FileTimes;
use std::path::Path;
use std::path::PathBuf;
use uuid::Uuid;
pub fn rollout_path(codex_home: &Path, filename_ts: &str, thread_id: &str) -> PathBuf {
let year = &filename_ts[0..4];
let month = &filename_ts[5..7];
let day = &filename_ts[8..10];
codex_home
.join("sessions")
.join(year)
.join(month)
.join(day)
.join(format!("rollout-{filename_ts}-{thread_id}.jsonl"))
}
/// Create a minimal rollout file under `CODEX_HOME/sessions/YYYY/MM/DD/`.
///
/// - `filename_ts` is the filename timestamp component in `YYYY-MM-DDThh-mm-ss` format.
@@ -30,25 +43,23 @@ pub fn create_fake_rollout(
let uuid_str = uuid.to_string();
let conversation_id = ThreadId::from_string(&uuid_str)?;
// sessions/YYYY/MM/DD derived from filename_ts (YYYY-MM-DDThh-mm-ss)
let year = &filename_ts[0..4];
let month = &filename_ts[5..7];
let day = &filename_ts[8..10];
let dir = codex_home.join("sessions").join(year).join(month).join(day);
fs::create_dir_all(&dir)?;
let file_path = dir.join(format!("rollout-{filename_ts}-{uuid}.jsonl"));
let file_path = rollout_path(codex_home, filename_ts, &uuid_str);
let dir = file_path
.parent()
.ok_or_else(|| anyhow::anyhow!("missing rollout parent directory"))?;
fs::create_dir_all(dir)?;
// Build JSONL lines
let meta = SessionMeta {
id: conversation_id,
forked_from_id: None,
timestamp: meta_rfc3339.to_string(),
cwd: PathBuf::from("/"),
originator: "codex".to_string(),
cli_version: "0.0.0".to_string(),
instructions: None,
source: SessionSource::Cli,
model_provider: model_provider.map(str::to_string),
base_instructions: None,
};
let payload = serde_json::to_value(SessionMetaLine {
meta,
@@ -84,6 +95,85 @@ pub fn create_fake_rollout(
.to_string(),
];
fs::write(&file_path, lines.join("\n") + "\n")?;
let parsed = chrono::DateTime::parse_from_rfc3339(meta_rfc3339)?.with_timezone(&chrono::Utc);
let times = FileTimes::new().set_modified(parsed.into());
std::fs::OpenOptions::new()
.append(true)
.open(&file_path)?
.set_times(times)?;
Ok(uuid_str)
}
pub fn create_fake_rollout_with_text_elements(
codex_home: &Path,
filename_ts: &str,
meta_rfc3339: &str,
preview: &str,
text_elements: Vec<serde_json::Value>,
model_provider: Option<&str>,
git_info: Option<GitInfo>,
) -> Result<String> {
let uuid = Uuid::new_v4();
let uuid_str = uuid.to_string();
let conversation_id = ThreadId::from_string(&uuid_str)?;
// sessions/YYYY/MM/DD derived from filename_ts (YYYY-MM-DDThh-mm-ss)
let year = &filename_ts[0..4];
let month = &filename_ts[5..7];
let day = &filename_ts[8..10];
let dir = codex_home.join("sessions").join(year).join(month).join(day);
fs::create_dir_all(&dir)?;
let file_path = dir.join(format!("rollout-{filename_ts}-{uuid}.jsonl"));
// Build JSONL lines
let meta = SessionMeta {
id: conversation_id,
forked_from_id: None,
timestamp: meta_rfc3339.to_string(),
cwd: PathBuf::from("/"),
originator: "codex".to_string(),
cli_version: "0.0.0".to_string(),
source: SessionSource::Cli,
model_provider: model_provider.map(str::to_string),
base_instructions: None,
};
let payload = serde_json::to_value(SessionMetaLine {
meta,
git: git_info,
})?;
let lines = [
json!( {
"timestamp": meta_rfc3339,
"type": "session_meta",
"payload": payload
})
.to_string(),
json!( {
"timestamp": meta_rfc3339,
"type":"response_item",
"payload": {
"type":"message",
"role":"user",
"content":[{"type":"input_text","text": preview}]
}
})
.to_string(),
json!( {
"timestamp": meta_rfc3339,
"type":"event_msg",
"payload": {
"type":"user_message",
"message": preview,
"text_elements": text_elements,
"local_images": []
}
})
.to_string(),
];
fs::write(file_path, lines.join("\n") + "\n")?;
Ok(uuid_str)
}

View File

@@ -114,6 +114,7 @@ async fn test_codex_jsonrpc_conversation_flow() -> Result<()> {
conversation_id,
items: vec![codex_app_server_protocol::InputItem::Text {
text: "text".to_string(),
text_elements: Vec::new(),
}],
})
.await?;
@@ -241,6 +242,7 @@ async fn test_send_user_turn_changes_approval_policy_behavior() -> Result<()> {
conversation_id,
items: vec![codex_app_server_protocol::InputItem::Text {
text: "run python".to_string(),
text_elements: Vec::new(),
}],
})
.await?;
@@ -296,6 +298,7 @@ async fn test_send_user_turn_changes_approval_policy_behavior() -> Result<()> {
conversation_id,
items: vec![codex_app_server_protocol::InputItem::Text {
text: "run python again".to_string(),
text_elements: Vec::new(),
}],
cwd: working_directory.clone(),
approval_policy: AskForApproval::Never,
@@ -405,6 +408,7 @@ async fn test_send_user_turn_updates_sandbox_and_cwd_between_turns() -> Result<(
conversation_id,
items: vec![InputItem::Text {
text: "first turn".to_string(),
text_elements: Vec::new(),
}],
cwd: first_cwd.clone(),
approval_policy: AskForApproval::Never,
@@ -437,6 +441,7 @@ async fn test_send_user_turn_updates_sandbox_and_cwd_between_turns() -> Result<(
conversation_id,
items: vec![InputItem::Text {
text: "second turn".to_string(),
text_elements: Vec::new(),
}],
cwd: second_cwd.clone(),
approval_policy: AskForApproval::Never,

View File

@@ -77,6 +77,7 @@ async fn test_conversation_create_and_send_message_ok() -> Result<()> {
conversation_id,
items: vec![InputItem::Text {
text: "Hello".to_string(),
text_elements: Vec::new(),
}],
})
.await?;

View File

@@ -105,6 +105,7 @@ async fn shell_command_interruption() -> anyhow::Result<()> {
conversation_id,
items: vec![codex_app_server_protocol::InputItem::Text {
text: "run first sleep command".to_string(),
text_elements: Vec::new(),
}],
})
.await?;

View File

@@ -80,6 +80,7 @@ async fn send_user_turn_accepts_output_schema_v1() -> Result<()> {
conversation_id,
items: vec![InputItem::Text {
text: "Hello".to_string(),
text_elements: Vec::new(),
}],
cwd: codex_home.path().to_path_buf(),
approval_policy: AskForApproval::Never,
@@ -181,6 +182,7 @@ async fn send_user_turn_output_schema_is_per_turn_v1() -> Result<()> {
conversation_id,
items: vec![InputItem::Text {
text: "Hello".to_string(),
text_elements: Vec::new(),
}],
cwd: codex_home.path().to_path_buf(),
approval_policy: AskForApproval::Never,
@@ -228,6 +230,7 @@ async fn send_user_turn_output_schema_is_per_turn_v1() -> Result<()> {
conversation_id,
items: vec![InputItem::Text {
text: "Hello again".to_string(),
text_elements: Vec::new(),
}],
cwd: codex_home.path().to_path_buf(),
approval_policy: AskForApproval::Never,

View File

@@ -101,6 +101,7 @@ async fn send_message(
conversation_id,
items: vec![InputItem::Text {
text: message.to_string(),
text_elements: Vec::new(),
}],
})
.await?;
@@ -194,6 +195,7 @@ async fn test_send_message_raw_notifications_opt_in() -> Result<()> {
conversation_id,
items: vec![InputItem::Text {
text: "Hello".to_string(),
text_elements: Vec::new(),
}],
})
.await?;
@@ -245,6 +247,7 @@ async fn test_send_message_session_not_found() -> Result<()> {
conversation_id: unknown,
items: vec![InputItem::Text {
text: "ping".to_string(),
text_elements: Vec::new(),
}],
})
.await?;
@@ -425,7 +428,7 @@ fn content_texts(content: &[ContentItem]) -> Vec<&str> {
content
.iter()
.filter_map(|item| match item {
ContentItem::InputText { text } | ContentItem::OutputText { text } => {
ContentItem::InputText { text, .. } | ContentItem::OutputText { text } => {
Some(text.as_str())
}
_ => None,

View File

@@ -0,0 +1,86 @@
//! Validates that the collaboration mode list endpoint returns the expected default presets.
//!
//! The test drives the app server through the MCP harness and asserts that the list response
//! includes the plan, pair programming, and execute modes with their default model and reasoning
//! effort settings, which keeps the API contract visible in one place.
#![allow(clippy::unwrap_used)]
use std::time::Duration;
use anyhow::Result;
use app_test_support::McpProcess;
use app_test_support::to_response;
use codex_app_server_protocol::CollaborationModeListParams;
use codex_app_server_protocol::CollaborationModeListResponse;
use codex_app_server_protocol::JSONRPCResponse;
use codex_app_server_protocol::RequestId;
use codex_core::models_manager::test_builtin_collaboration_mode_presets;
use codex_protocol::config_types::CollaborationMode;
use pretty_assertions::assert_eq;
use tempfile::TempDir;
use tokio::time::timeout;
const DEFAULT_TIMEOUT: Duration = Duration::from_secs(10);
/// Confirms the server returns the default collaboration mode presets in a stable order.
#[tokio::test]
async fn list_collaboration_modes_returns_presets() -> Result<()> {
let codex_home = TempDir::new()?;
let mut mcp = McpProcess::new(codex_home.path()).await?;
timeout(DEFAULT_TIMEOUT, mcp.initialize()).await??;
let request_id = mcp
.send_list_collaboration_modes_request(CollaborationModeListParams {})
.await?;
let response: JSONRPCResponse = timeout(
DEFAULT_TIMEOUT,
mcp.read_stream_until_response_message(RequestId::Integer(request_id)),
)
.await??;
let CollaborationModeListResponse { data: items } =
to_response::<CollaborationModeListResponse>(response)?;
let expected = vec![plan_preset(), pair_programming_preset(), execute_preset()];
assert_eq!(expected, items);
Ok(())
}
/// Builds the plan preset that the list response is expected to return.
///
/// If the defaults change in the app server, this helper should be updated alongside the
/// contract, or the test will fail in ways that imply a regression in the API.
fn plan_preset() -> CollaborationMode {
let presets = test_builtin_collaboration_mode_presets();
presets
.into_iter()
.find(|p| matches!(p, CollaborationMode::Plan(_)))
.unwrap()
}
/// Builds the pair programming preset that the list response is expected to return.
///
/// The helper keeps the expected model and reasoning defaults co-located with the test
/// so that mismatches point directly at the API contract being exercised.
fn pair_programming_preset() -> CollaborationMode {
let presets = test_builtin_collaboration_mode_presets();
presets
.into_iter()
.find(|p| matches!(p, CollaborationMode::PairProgramming(_)))
.unwrap()
}
/// Builds the execute preset that the list response is expected to return.
///
/// The execute preset uses a different reasoning effort to capture the higher-effort
/// execution contract the server currently exposes.
fn execute_preset() -> CollaborationMode {
let presets = test_builtin_collaboration_mode_presets();
presets
.into_iter()
.find(|p| matches!(p, CollaborationMode::Execute(_)))
.unwrap()
}

View File

@@ -1,10 +1,12 @@
mod account;
mod analytics;
mod collaboration_mode_list;
mod config_rpc;
mod initialize;
mod model_list;
mod output_schema;
mod rate_limits;
mod request_user_input;
mod review;
mod thread_archive;
mod thread_fork;

View File

@@ -61,6 +61,7 @@ async fn turn_start_accepts_output_schema_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "Hello".to_string(),
text_elements: Vec::new(),
}],
output_schema: Some(output_schema.clone()),
..Default::default()
@@ -142,6 +143,7 @@ async fn turn_start_output_schema_is_per_turn_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "Hello".to_string(),
text_elements: Vec::new(),
}],
output_schema: Some(output_schema.clone()),
..Default::default()
@@ -183,6 +185,7 @@ async fn turn_start_output_schema_is_per_turn_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "Hello again".to_string(),
text_elements: Vec::new(),
}],
output_schema: None,
..Default::default()

View File

@@ -0,0 +1,134 @@
use anyhow::Result;
use app_test_support::McpProcess;
use app_test_support::create_final_assistant_message_sse_response;
use app_test_support::create_mock_responses_server_sequence;
use app_test_support::create_request_user_input_sse_response;
use app_test_support::to_response;
use codex_app_server_protocol::JSONRPCResponse;
use codex_app_server_protocol::RequestId;
use codex_app_server_protocol::ServerRequest;
use codex_app_server_protocol::ThreadStartParams;
use codex_app_server_protocol::ThreadStartResponse;
use codex_app_server_protocol::TurnStartParams;
use codex_app_server_protocol::TurnStartResponse;
use codex_app_server_protocol::UserInput as V2UserInput;
use codex_protocol::config_types::CollaborationMode;
use codex_protocol::config_types::Settings;
use codex_protocol::openai_models::ReasoningEffort;
use tokio::time::timeout;
const DEFAULT_READ_TIMEOUT: std::time::Duration = std::time::Duration::from_secs(10);
#[tokio::test(flavor = "multi_thread", worker_threads = 4)]
async fn request_user_input_round_trip() -> Result<()> {
let codex_home = tempfile::TempDir::new()?;
let responses = vec![
create_request_user_input_sse_response("call1")?,
create_final_assistant_message_sse_response("done")?,
];
let server = create_mock_responses_server_sequence(responses).await;
create_config_toml(codex_home.path(), &server.uri())?;
let mut mcp = McpProcess::new(codex_home.path()).await?;
timeout(DEFAULT_READ_TIMEOUT, mcp.initialize()).await??;
let thread_start_id = mcp
.send_thread_start_request(ThreadStartParams {
model: Some("mock-model".to_string()),
..Default::default()
})
.await?;
let thread_start_resp: JSONRPCResponse = timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_response_message(RequestId::Integer(thread_start_id)),
)
.await??;
let ThreadStartResponse { thread, .. } = to_response(thread_start_resp)?;
let turn_start_id = mcp
.send_turn_start_request(TurnStartParams {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "ask something".to_string(),
text_elements: Vec::new(),
}],
model: Some("mock-model".to_string()),
effort: Some(ReasoningEffort::Medium),
collaboration_mode: Some(CollaborationMode::Plan(Settings {
model: "mock-model".to_string(),
reasoning_effort: Some(ReasoningEffort::Medium),
developer_instructions: None,
})),
..Default::default()
})
.await?;
let turn_start_resp: JSONRPCResponse = timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_response_message(RequestId::Integer(turn_start_id)),
)
.await??;
let TurnStartResponse { turn, .. } = to_response(turn_start_resp)?;
let server_req = timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_request_message(),
)
.await??;
let ServerRequest::ToolRequestUserInput { request_id, params } = server_req else {
panic!("expected ToolRequestUserInput request, got: {server_req:?}");
};
assert_eq!(params.thread_id, thread.id);
assert_eq!(params.turn_id, turn.id);
assert_eq!(params.item_id, "call1");
assert_eq!(params.questions.len(), 1);
mcp.send_response(
request_id,
serde_json::json!({
"answers": {
"confirm_path": { "selected": ["yes"], "other": serde_json::Value::Null }
}
}),
)
.await?;
timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_notification_message("codex/event/task_complete"),
)
.await??;
timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_notification_message("turn/completed"),
)
.await??;
Ok(())
}
fn create_config_toml(codex_home: &std::path::Path, server_uri: &str) -> std::io::Result<()> {
let config_toml = codex_home.join("config.toml");
std::fs::write(
config_toml,
format!(
r#"
model = "mock-model"
approval_policy = "untrusted"
sandbox_mode = "read-only"
model_provider = "mock_provider"
[features]
collaboration_modes = true
[model_providers.mock_provider]
name = "Mock provider for test"
base_url = "{server_uri}/v1"
wire_api = "responses"
request_max_retries = 0
stream_max_retries = 0
"#
),
)
}

View File

@@ -95,7 +95,8 @@ async fn thread_fork_creates_new_thread_and_emits_started() -> Result<()> {
assert_eq!(
content,
&vec![UserInput::Text {
text: preview.to_string()
text: preview.to_string(),
text_elements: Vec::new(),
}]
);
}

View File

@@ -1,17 +1,27 @@
use anyhow::Result;
use app_test_support::McpProcess;
use app_test_support::create_fake_rollout;
use app_test_support::rollout_path;
use app_test_support::to_response;
use chrono::DateTime;
use chrono::Utc;
use codex_app_server_protocol::GitInfo as ApiGitInfo;
use codex_app_server_protocol::JSONRPCError;
use codex_app_server_protocol::JSONRPCResponse;
use codex_app_server_protocol::RequestId;
use codex_app_server_protocol::SessionSource;
use codex_app_server_protocol::ThreadListResponse;
use codex_app_server_protocol::ThreadSortKey;
use codex_protocol::protocol::GitInfo as CoreGitInfo;
use pretty_assertions::assert_eq;
use std::cmp::Reverse;
use std::fs::FileTimes;
use std::fs::OpenOptions;
use std::path::Path;
use std::path::PathBuf;
use tempfile::TempDir;
use tokio::time::timeout;
use uuid::Uuid;
const DEFAULT_READ_TIMEOUT: std::time::Duration = std::time::Duration::from_secs(10);
@@ -26,11 +36,22 @@ async fn list_threads(
cursor: Option<String>,
limit: Option<u32>,
providers: Option<Vec<String>>,
) -> Result<ThreadListResponse> {
list_threads_with_sort(mcp, cursor, limit, providers, None).await
}
async fn list_threads_with_sort(
mcp: &mut McpProcess,
cursor: Option<String>,
limit: Option<u32>,
providers: Option<Vec<String>>,
sort_key: Option<ThreadSortKey>,
) -> Result<ThreadListResponse> {
let request_id = mcp
.send_thread_list_request(codex_app_server_protocol::ThreadListParams {
cursor,
limit,
sort_key,
model_providers: providers,
})
.await?;
@@ -82,6 +103,16 @@ fn timestamp_at(
)
}
fn set_rollout_mtime(path: &Path, updated_at_rfc3339: &str) -> Result<()> {
let parsed = DateTime::parse_from_rfc3339(updated_at_rfc3339)?.with_timezone(&Utc);
let times = FileTimes::new().set_modified(parsed.into());
OpenOptions::new()
.append(true)
.open(path)?
.set_times(times)?;
Ok(())
}
#[tokio::test]
async fn thread_list_basic_empty() -> Result<()> {
let codex_home = TempDir::new()?;
@@ -163,6 +194,7 @@ async fn thread_list_pagination_next_cursor_none_on_last_page() -> Result<()> {
assert_eq!(thread.preview, "Hello");
assert_eq!(thread.model_provider, "mock_provider");
assert!(thread.created_at > 0);
assert_eq!(thread.updated_at, thread.created_at);
assert_eq!(thread.cwd, PathBuf::from("/"));
assert_eq!(thread.cli_version, "0.0.0");
assert_eq!(thread.source, SessionSource::Cli);
@@ -186,6 +218,7 @@ async fn thread_list_pagination_next_cursor_none_on_last_page() -> Result<()> {
assert_eq!(thread.preview, "Hello");
assert_eq!(thread.model_provider, "mock_provider");
assert!(thread.created_at > 0);
assert_eq!(thread.updated_at, thread.created_at);
assert_eq!(thread.cwd, PathBuf::from("/"));
assert_eq!(thread.cli_version, "0.0.0");
assert_eq!(thread.source, SessionSource::Cli);
@@ -236,6 +269,7 @@ async fn thread_list_respects_provider_filter() -> Result<()> {
assert_eq!(thread.model_provider, "other_provider");
let expected_ts = chrono::DateTime::parse_from_rfc3339("2025-01-02T11:00:00Z")?.timestamp();
assert_eq!(thread.created_at, expected_ts);
assert_eq!(thread.updated_at, expected_ts);
assert_eq!(thread.cwd, PathBuf::from("/"));
assert_eq!(thread.cli_version, "0.0.0");
assert_eq!(thread.source, SessionSource::Cli);
@@ -429,3 +463,351 @@ async fn thread_list_includes_git_info() -> Result<()> {
Ok(())
}
#[tokio::test]
async fn thread_list_default_sorts_by_created_at() -> Result<()> {
let codex_home = TempDir::new()?;
create_minimal_config(codex_home.path())?;
let id_a = create_fake_rollout(
codex_home.path(),
"2025-01-02T12-00-00",
"2025-01-02T12:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let id_b = create_fake_rollout(
codex_home.path(),
"2025-01-01T13-00-00",
"2025-01-01T13:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let id_c = create_fake_rollout(
codex_home.path(),
"2025-01-01T12-00-00",
"2025-01-01T12:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let mut mcp = init_mcp(codex_home.path()).await?;
let ThreadListResponse { data, .. } = list_threads_with_sort(
&mut mcp,
None,
Some(10),
Some(vec!["mock_provider".to_string()]),
None,
)
.await?;
let ids: Vec<_> = data.iter().map(|thread| thread.id.as_str()).collect();
assert_eq!(ids, vec![id_a.as_str(), id_b.as_str(), id_c.as_str()]);
Ok(())
}
#[tokio::test]
async fn thread_list_sort_updated_at_orders_by_mtime() -> Result<()> {
let codex_home = TempDir::new()?;
create_minimal_config(codex_home.path())?;
let id_old = create_fake_rollout(
codex_home.path(),
"2025-01-01T10-00-00",
"2025-01-01T10:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let id_mid = create_fake_rollout(
codex_home.path(),
"2025-01-01T11-00-00",
"2025-01-01T11:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let id_new = create_fake_rollout(
codex_home.path(),
"2025-01-01T12-00-00",
"2025-01-01T12:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
set_rollout_mtime(
rollout_path(codex_home.path(), "2025-01-01T10-00-00", &id_old).as_path(),
"2025-01-03T00:00:00Z",
)?;
set_rollout_mtime(
rollout_path(codex_home.path(), "2025-01-01T11-00-00", &id_mid).as_path(),
"2025-01-02T00:00:00Z",
)?;
set_rollout_mtime(
rollout_path(codex_home.path(), "2025-01-01T12-00-00", &id_new).as_path(),
"2025-01-01T00:00:00Z",
)?;
let mut mcp = init_mcp(codex_home.path()).await?;
let ThreadListResponse { data, .. } = list_threads_with_sort(
&mut mcp,
None,
Some(10),
Some(vec!["mock_provider".to_string()]),
Some(ThreadSortKey::UpdatedAt),
)
.await?;
let ids: Vec<_> = data.iter().map(|thread| thread.id.as_str()).collect();
assert_eq!(ids, vec![id_old.as_str(), id_mid.as_str(), id_new.as_str()]);
Ok(())
}
#[tokio::test]
async fn thread_list_updated_at_paginates_with_cursor() -> Result<()> {
let codex_home = TempDir::new()?;
create_minimal_config(codex_home.path())?;
let id_a = create_fake_rollout(
codex_home.path(),
"2025-02-01T10-00-00",
"2025-02-01T10:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let id_b = create_fake_rollout(
codex_home.path(),
"2025-02-01T11-00-00",
"2025-02-01T11:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let id_c = create_fake_rollout(
codex_home.path(),
"2025-02-01T12-00-00",
"2025-02-01T12:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
set_rollout_mtime(
rollout_path(codex_home.path(), "2025-02-01T10-00-00", &id_a).as_path(),
"2025-02-03T00:00:00Z",
)?;
set_rollout_mtime(
rollout_path(codex_home.path(), "2025-02-01T11-00-00", &id_b).as_path(),
"2025-02-02T00:00:00Z",
)?;
set_rollout_mtime(
rollout_path(codex_home.path(), "2025-02-01T12-00-00", &id_c).as_path(),
"2025-02-01T00:00:00Z",
)?;
let mut mcp = init_mcp(codex_home.path()).await?;
let ThreadListResponse {
data: page1,
next_cursor: cursor1,
} = list_threads_with_sort(
&mut mcp,
None,
Some(2),
Some(vec!["mock_provider".to_string()]),
Some(ThreadSortKey::UpdatedAt),
)
.await?;
let ids_page1: Vec<_> = page1.iter().map(|thread| thread.id.as_str()).collect();
assert_eq!(ids_page1, vec![id_a.as_str(), id_b.as_str()]);
let cursor1 = cursor1.expect("expected nextCursor on first page");
let ThreadListResponse {
data: page2,
next_cursor: cursor2,
} = list_threads_with_sort(
&mut mcp,
Some(cursor1),
Some(2),
Some(vec!["mock_provider".to_string()]),
Some(ThreadSortKey::UpdatedAt),
)
.await?;
let ids_page2: Vec<_> = page2.iter().map(|thread| thread.id.as_str()).collect();
assert_eq!(ids_page2, vec![id_c.as_str()]);
assert_eq!(cursor2, None);
Ok(())
}
#[tokio::test]
async fn thread_list_created_at_tie_breaks_by_uuid() -> Result<()> {
let codex_home = TempDir::new()?;
create_minimal_config(codex_home.path())?;
let id_a = create_fake_rollout(
codex_home.path(),
"2025-02-01T10-00-00",
"2025-02-01T10:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let id_b = create_fake_rollout(
codex_home.path(),
"2025-02-01T10-00-00",
"2025-02-01T10:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let mut mcp = init_mcp(codex_home.path()).await?;
let ThreadListResponse { data, .. } = list_threads(
&mut mcp,
None,
Some(10),
Some(vec!["mock_provider".to_string()]),
)
.await?;
let ids: Vec<_> = data.iter().map(|thread| thread.id.as_str()).collect();
let mut expected = [id_a, id_b];
expected.sort_by_key(|id| Reverse(Uuid::parse_str(id).expect("uuid should parse")));
let expected: Vec<_> = expected.iter().map(String::as_str).collect();
assert_eq!(ids, expected);
Ok(())
}
#[tokio::test]
async fn thread_list_updated_at_tie_breaks_by_uuid() -> Result<()> {
let codex_home = TempDir::new()?;
create_minimal_config(codex_home.path())?;
let id_a = create_fake_rollout(
codex_home.path(),
"2025-02-01T10-00-00",
"2025-02-01T10:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let id_b = create_fake_rollout(
codex_home.path(),
"2025-02-01T11-00-00",
"2025-02-01T11:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
let updated_at = "2025-02-03T00:00:00Z";
set_rollout_mtime(
rollout_path(codex_home.path(), "2025-02-01T10-00-00", &id_a).as_path(),
updated_at,
)?;
set_rollout_mtime(
rollout_path(codex_home.path(), "2025-02-01T11-00-00", &id_b).as_path(),
updated_at,
)?;
let mut mcp = init_mcp(codex_home.path()).await?;
let ThreadListResponse { data, .. } = list_threads_with_sort(
&mut mcp,
None,
Some(10),
Some(vec!["mock_provider".to_string()]),
Some(ThreadSortKey::UpdatedAt),
)
.await?;
let ids: Vec<_> = data.iter().map(|thread| thread.id.as_str()).collect();
let mut expected = [id_a, id_b];
expected.sort_by_key(|id| Reverse(Uuid::parse_str(id).expect("uuid should parse")));
let expected: Vec<_> = expected.iter().map(String::as_str).collect();
assert_eq!(ids, expected);
Ok(())
}
#[tokio::test]
async fn thread_list_updated_at_uses_mtime() -> Result<()> {
let codex_home = TempDir::new()?;
create_minimal_config(codex_home.path())?;
let thread_id = create_fake_rollout(
codex_home.path(),
"2025-02-01T10-00-00",
"2025-02-01T10:00:00Z",
"Hello",
Some("mock_provider"),
None,
)?;
set_rollout_mtime(
rollout_path(codex_home.path(), "2025-02-01T10-00-00", &thread_id).as_path(),
"2025-02-05T00:00:00Z",
)?;
let mut mcp = init_mcp(codex_home.path()).await?;
let ThreadListResponse { data, .. } = list_threads_with_sort(
&mut mcp,
None,
Some(10),
Some(vec!["mock_provider".to_string()]),
Some(ThreadSortKey::UpdatedAt),
)
.await?;
let thread = data
.iter()
.find(|item| item.id == thread_id)
.expect("expected thread for created rollout");
let expected_created =
chrono::DateTime::parse_from_rfc3339("2025-02-01T10:00:00Z")?.timestamp();
let expected_updated =
chrono::DateTime::parse_from_rfc3339("2025-02-05T00:00:00Z")?.timestamp();
assert_eq!(thread.created_at, expected_created);
assert_eq!(thread.updated_at, expected_updated);
Ok(())
}
#[tokio::test]
async fn thread_list_invalid_cursor_returns_error() -> Result<()> {
let codex_home = TempDir::new()?;
create_minimal_config(codex_home.path())?;
let mut mcp = init_mcp(codex_home.path()).await?;
let request_id = mcp
.send_thread_list_request(codex_app_server_protocol::ThreadListParams {
cursor: Some("not-a-cursor".to_string()),
limit: Some(2),
sort_key: None,
model_providers: Some(vec!["mock_provider".to_string()]),
})
.await?;
let error: JSONRPCError = timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_error_message(RequestId::Integer(request_id)),
)
.await??;
assert_eq!(error.error.code, -32600);
assert_eq!(error.error.message, "invalid cursor: not-a-cursor");
Ok(())
}

View File

@@ -1,6 +1,6 @@
use anyhow::Result;
use app_test_support::McpProcess;
use app_test_support::create_fake_rollout;
use app_test_support::create_fake_rollout_with_text_elements;
use app_test_support::create_mock_responses_server_repeating_assistant;
use app_test_support::to_response;
use codex_app_server_protocol::JSONRPCResponse;
@@ -15,6 +15,9 @@ use codex_app_server_protocol::TurnStatus;
use codex_app_server_protocol::UserInput;
use codex_protocol::models::ContentItem;
use codex_protocol::models::ResponseItem;
use codex_protocol::user_input::ByteRange;
use codex_protocol::user_input::TextElement;
use pretty_assertions::assert_eq;
use std::path::PathBuf;
use tempfile::TempDir;
use tokio::time::timeout;
@@ -59,7 +62,9 @@ async fn thread_resume_returns_original_thread() -> Result<()> {
let ThreadResumeResponse {
thread: resumed, ..
} = to_response::<ThreadResumeResponse>(resume_resp)?;
assert_eq!(resumed, thread);
let mut expected = thread;
expected.updated_at = resumed.updated_at;
assert_eq!(resumed, expected);
Ok(())
}
@@ -71,11 +76,19 @@ async fn thread_resume_returns_rollout_history() -> Result<()> {
create_config_toml(codex_home.path(), &server.uri())?;
let preview = "Saved user message";
let conversation_id = create_fake_rollout(
let text_elements = vec![TextElement::new(
ByteRange { start: 0, end: 5 },
Some("<note>".into()),
)];
let conversation_id = create_fake_rollout_with_text_elements(
codex_home.path(),
"2025-01-05T12-00-00",
"2025-01-05T12:00:00Z",
preview,
text_elements
.iter()
.map(|elem| serde_json::to_value(elem).expect("serialize text element"))
.collect(),
Some("mock_provider"),
None,
)?;
@@ -118,7 +131,8 @@ async fn thread_resume_returns_rollout_history() -> Result<()> {
assert_eq!(
content,
&vec![UserInput::Text {
text: preview.to_string()
text: preview.to_string(),
text_elements: text_elements.clone().into_iter().map(Into::into).collect(),
}]
);
}
@@ -167,7 +181,9 @@ async fn thread_resume_prefers_path_over_thread_id() -> Result<()> {
let ThreadResumeResponse {
thread: resumed, ..
} = to_response::<ThreadResumeResponse>(resume_resp)?;
assert_eq!(resumed, thread);
let mut expected = thread;
expected.updated_at = resumed.updated_at;
assert_eq!(resumed, expected);
Ok(())
}

View File

@@ -57,6 +57,7 @@ async fn thread_rollback_drops_last_turns_and_persists_to_rollout() -> Result<()
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: first_text.to_string(),
text_elements: Vec::new(),
}],
..Default::default()
})
@@ -77,6 +78,7 @@ async fn thread_rollback_drops_last_turns_and_persists_to_rollout() -> Result<()
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "Second".to_string(),
text_elements: Vec::new(),
}],
..Default::default()
})
@@ -115,7 +117,8 @@ async fn thread_rollback_drops_last_turns_and_persists_to_rollout() -> Result<()
assert_eq!(
content,
&vec![V2UserInput::Text {
text: first_text.to_string()
text: first_text.to_string(),
text_elements: Vec::new(),
}]
);
}
@@ -143,7 +146,8 @@ async fn thread_rollback_drops_last_turns_and_persists_to_rollout() -> Result<()
assert_eq!(
content,
&vec![V2UserInput::Text {
text: first_text.to_string()
text: first_text.to_string(),
text_elements: Vec::new(),
}]
);
}

View File

@@ -73,6 +73,7 @@ async fn turn_interrupt_aborts_running_turn() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "run sleep".to_string(),
text_elements: Vec::new(),
}],
cwd: Some(working_directory.clone()),
..Default::default()

View File

@@ -8,6 +8,7 @@ use app_test_support::create_mock_responses_server_sequence_unchecked;
use app_test_support::create_shell_command_sse_response;
use app_test_support::format_with_current_shell_display;
use app_test_support::to_response;
use codex_app_server_protocol::ByteRange;
use codex_app_server_protocol::ClientInfo;
use codex_app_server_protocol::CommandExecutionApprovalDecision;
use codex_app_server_protocol::CommandExecutionRequestApprovalResponse;
@@ -23,6 +24,7 @@ use codex_app_server_protocol::PatchApplyStatus;
use codex_app_server_protocol::PatchChangeKind;
use codex_app_server_protocol::RequestId;
use codex_app_server_protocol::ServerRequest;
use codex_app_server_protocol::TextElement;
use codex_app_server_protocol::ThreadItem;
use codex_app_server_protocol::ThreadStartParams;
use codex_app_server_protocol::ThreadStartResponse;
@@ -33,7 +35,10 @@ use codex_app_server_protocol::TurnStartedNotification;
use codex_app_server_protocol::TurnStatus;
use codex_app_server_protocol::UserInput as V2UserInput;
use codex_core::protocol_config_types::ReasoningSummary;
use codex_protocol::config_types::CollaborationMode;
use codex_protocol::config_types::Settings;
use codex_protocol::openai_models::ReasoningEffort;
use core_test_support::responses;
use core_test_support::skip_if_no_network;
use pretty_assertions::assert_eq;
use std::path::Path;
@@ -80,6 +85,7 @@ async fn turn_start_sends_originator_header() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "Hello".to_string(),
text_elements: Vec::new(),
}],
..Default::default()
})
@@ -112,6 +118,87 @@ async fn turn_start_sends_originator_header() -> Result<()> {
Ok(())
}
#[tokio::test]
async fn turn_start_emits_user_message_item_with_text_elements() -> Result<()> {
let responses = vec![create_final_assistant_message_sse_response("Done")?];
let server = create_mock_responses_server_sequence_unchecked(responses).await;
let codex_home = TempDir::new()?;
create_config_toml(codex_home.path(), &server.uri(), "never")?;
let mut mcp = McpProcess::new(codex_home.path()).await?;
timeout(DEFAULT_READ_TIMEOUT, mcp.initialize()).await??;
let thread_req = mcp
.send_thread_start_request(ThreadStartParams {
model: Some("mock-model".to_string()),
..Default::default()
})
.await?;
let thread_resp: JSONRPCResponse = timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_response_message(RequestId::Integer(thread_req)),
)
.await??;
let ThreadStartResponse { thread, .. } = to_response::<ThreadStartResponse>(thread_resp)?;
let text_elements = vec![TextElement::new(
ByteRange { start: 0, end: 5 },
Some("<note>".to_string()),
)];
let turn_req = mcp
.send_turn_start_request(TurnStartParams {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "Hello".to_string(),
text_elements: text_elements.clone(),
}],
..Default::default()
})
.await?;
timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_response_message(RequestId::Integer(turn_req)),
)
.await??;
let user_message_item = timeout(DEFAULT_READ_TIMEOUT, async {
loop {
let notification = mcp
.read_stream_until_notification_message("item/started")
.await?;
let params = notification.params.expect("item/started params");
let item_started: ItemStartedNotification =
serde_json::from_value(params).expect("deserialize item/started notification");
if let ThreadItem::UserMessage { .. } = item_started.item {
return Ok::<ThreadItem, anyhow::Error>(item_started.item);
}
}
})
.await??;
match user_message_item {
ThreadItem::UserMessage { content, .. } => {
assert_eq!(
content,
vec![V2UserInput::Text {
text: "Hello".to_string(),
text_elements,
}]
);
}
other => panic!("expected user message item, got {other:?}"),
}
timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_notification_message("turn/completed"),
)
.await??;
Ok(())
}
#[tokio::test]
async fn turn_start_emits_notifications_and_accepts_model_override() -> Result<()> {
// Provide a mock server and config so model wiring is valid.
@@ -149,6 +236,7 @@ async fn turn_start_emits_notifications_and_accepts_model_override() -> Result<(
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "Hello".to_string(),
text_elements: Vec::new(),
}],
..Default::default()
})
@@ -181,6 +269,7 @@ async fn turn_start_emits_notifications_and_accepts_model_override() -> Result<(
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "Second".to_string(),
text_elements: Vec::new(),
}],
model: Some("mock-model-override".to_string()),
..Default::default()
@@ -219,6 +308,77 @@ async fn turn_start_emits_notifications_and_accepts_model_override() -> Result<(
Ok(())
}
#[tokio::test]
async fn turn_start_accepts_collaboration_mode_override_v2() -> Result<()> {
skip_if_no_network!(Ok(()));
let server = responses::start_mock_server().await;
let body = responses::sse(vec![
responses::ev_response_created("resp-1"),
responses::ev_assistant_message("msg-1", "Done"),
responses::ev_completed("resp-1"),
]);
let response_mock = responses::mount_sse_once(&server, body).await;
let codex_home = TempDir::new()?;
create_config_toml(codex_home.path(), &server.uri(), "never")?;
let mut mcp = McpProcess::new(codex_home.path()).await?;
timeout(DEFAULT_READ_TIMEOUT, mcp.initialize()).await??;
let thread_req = mcp
.send_thread_start_request(ThreadStartParams {
..Default::default()
})
.await?;
let thread_resp: JSONRPCResponse = timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_response_message(RequestId::Integer(thread_req)),
)
.await??;
let ThreadStartResponse { thread, .. } = to_response::<ThreadStartResponse>(thread_resp)?;
let collaboration_mode = CollaborationMode::Custom(Settings {
model: "mock-model-collab".to_string(),
reasoning_effort: Some(ReasoningEffort::High),
developer_instructions: None,
});
let turn_req = mcp
.send_turn_start_request(TurnStartParams {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "Hello".to_string(),
text_elements: Vec::new(),
}],
model: Some("mock-model-override".to_string()),
effort: Some(ReasoningEffort::Low),
summary: Some(ReasoningSummary::Auto),
output_schema: None,
collaboration_mode: Some(collaboration_mode),
..Default::default()
})
.await?;
let turn_resp: JSONRPCResponse = timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_response_message(RequestId::Integer(turn_req)),
)
.await??;
let _turn: TurnStartResponse = to_response::<TurnStartResponse>(turn_resp)?;
timeout(
DEFAULT_READ_TIMEOUT,
mcp.read_stream_until_notification_message("turn/completed"),
)
.await??;
let request = response_mock.single_request();
let payload = request.body_json();
assert_eq!(payload["model"].as_str(), Some("mock-model-collab"));
Ok(())
}
#[tokio::test]
async fn turn_start_accepts_local_image_input() -> Result<()> {
// Two Codex turns hit the mock model (session start + turn/start).
@@ -331,6 +491,7 @@ async fn turn_start_exec_approval_toggle_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "run python".to_string(),
text_elements: Vec::new(),
}],
..Default::default()
})
@@ -376,6 +537,7 @@ async fn turn_start_exec_approval_toggle_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "run python again".to_string(),
text_elements: Vec::new(),
}],
approval_policy: Some(codex_app_server_protocol::AskForApproval::Never),
sandbox_policy: Some(codex_app_server_protocol::SandboxPolicy::DangerFullAccess),
@@ -452,6 +614,7 @@ async fn turn_start_exec_approval_decline_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "run python".to_string(),
text_elements: Vec::new(),
}],
cwd: Some(workspace.clone()),
..Default::default()
@@ -600,6 +763,7 @@ async fn turn_start_updates_sandbox_and_cwd_between_turns_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "first turn".to_string(),
text_elements: Vec::new(),
}],
cwd: Some(first_cwd.clone()),
approval_policy: Some(codex_app_server_protocol::AskForApproval::Never),
@@ -613,6 +777,7 @@ async fn turn_start_updates_sandbox_and_cwd_between_turns_v2() -> Result<()> {
effort: Some(ReasoningEffort::Medium),
summary: Some(ReasoningSummary::Auto),
output_schema: None,
collaboration_mode: None,
})
.await?;
timeout(
@@ -633,6 +798,7 @@ async fn turn_start_updates_sandbox_and_cwd_between_turns_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "second turn".to_string(),
text_elements: Vec::new(),
}],
cwd: Some(second_cwd.clone()),
approval_policy: Some(codex_app_server_protocol::AskForApproval::Never),
@@ -641,6 +807,7 @@ async fn turn_start_updates_sandbox_and_cwd_between_turns_v2() -> Result<()> {
effort: Some(ReasoningEffort::Medium),
summary: Some(ReasoningSummary::Auto),
output_schema: None,
collaboration_mode: None,
})
.await?;
timeout(
@@ -733,6 +900,7 @@ async fn turn_start_file_change_approval_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "apply patch".into(),
text_elements: Vec::new(),
}],
cwd: Some(workspace.clone()),
..Default::default()
@@ -910,6 +1078,7 @@ async fn turn_start_file_change_approval_accept_for_session_persists_v2() -> Res
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "apply patch 1".into(),
text_elements: Vec::new(),
}],
cwd: Some(workspace.clone()),
..Default::default()
@@ -986,6 +1155,7 @@ async fn turn_start_file_change_approval_accept_for_session_persists_v2() -> Res
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "apply patch 2".into(),
text_elements: Vec::new(),
}],
cwd: Some(workspace.clone()),
..Default::default()
@@ -1083,6 +1253,7 @@ async fn turn_start_file_change_approval_decline_v2() -> Result<()> {
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "apply patch".into(),
text_elements: Vec::new(),
}],
cwd: Some(workspace.clone()),
..Default::default()
@@ -1230,6 +1401,7 @@ unified_exec = true
thread_id: thread.id.clone(),
input: vec![V2UserInput::Text {
text: "run a command".to_string(),
text_elements: Vec::new(),
}],
..Default::default()
})
@@ -1299,8 +1471,18 @@ unified_exec = true
unreachable!("loop ensures we break on command execution items");
};
assert_eq!(completed_id, "uexec-1");
assert_eq!(completed_status, CommandExecutionStatus::Completed);
assert_eq!(exit_code, Some(0));
assert!(
matches!(
completed_status,
CommandExecutionStatus::Completed | CommandExecutionStatus::Failed
),
"unexpected command execution status: {completed_status:?}"
);
if completed_status == CommandExecutionStatus::Completed {
assert_eq!(exit_code, Some(0));
} else {
assert!(exit_code.is_some(), "expected exit_code for failed command");
}
assert_eq!(
completed_process_id.as_deref(),
Some(started_process_id.as_str())

View File

@@ -12,7 +12,7 @@ path = "src/lib.rs"
anyhow = "1"
serde = { version = "1", features = ["derive"] }
serde_json = "1"
reqwest = { version = "0.12", default-features = false, features = ["json", "rustls-tls"] }
reqwest = { version = "0.12", default-features = false, features = ["json", "rustls-tls-webpki-roots", "rustls-tls-native-roots"] }
codex-backend-openapi-models = { path = "../codex-backend-openapi-models" }
codex-protocol = { workspace = true }
codex-core = { workspace = true }

View File

@@ -174,6 +174,7 @@ impl Client {
limit: Option<i32>,
task_filter: Option<&str>,
environment_id: Option<&str>,
cursor: Option<&str>,
) -> Result<PaginatedListTaskListItem> {
let url = match self.path_style {
PathStyle::CodexApi => format!("{}/api/codex/tasks/list", self.base_url),
@@ -190,6 +191,11 @@ impl Client {
} else {
req
};
let req = if let Some(c) = cursor {
req.query(&[("cursor", c)])
} else {
req
};
let req = if let Some(id) = environment_id {
req.query(&[("environment_id", id)])
} else {

View File

@@ -14,11 +14,9 @@ use codex_cli::login::run_login_status;
use codex_cli::login::run_login_with_api_key;
use codex_cli::login::run_login_with_chatgpt;
use codex_cli::login::run_login_with_device_code;
use codex_cli::login::run_login_with_device_code_fallback_to_browser;
use codex_cli::login::run_logout;
use codex_cloud_tasks::Cli as CloudTasksCli;
use codex_common::CliConfigOverrides;
use codex_core::env::is_headless_environment;
use codex_exec::Cli as ExecCli;
use codex_exec::Command as ExecCommand;
use codex_exec::ReviewArgs;
@@ -30,6 +28,7 @@ use codex_tui::ExitReason;
use codex_tui::update_action::UpdateAction;
use codex_tui2 as tui2;
use owo_colors::OwoColorize;
use std::io::IsTerminal;
use std::path::PathBuf;
use supports_color::Stream;
@@ -47,6 +46,7 @@ use codex_core::features::Feature;
use codex_core::features::FeatureOverrides;
use codex_core::features::Features;
use codex_core::features::is_known_feature_key;
use codex_core::terminal::TerminalName;
use codex_utils_absolute_path::AbsolutePathBuf;
/// Codex CLI
@@ -461,8 +461,8 @@ enum FeaturesSubcommand {
fn stage_str(stage: codex_core::features::Stage) -> &'static str {
use codex_core::features::Stage;
match stage {
Stage::Experimental => "experimental",
Stage::Beta { .. } => "beta",
Stage::Beta => "experimental",
Stage::Experimental { .. } => "beta",
Stage::Stable => "stable",
Stage::Deprecated => "deprecated",
Stage::Removed => "removed",
@@ -600,13 +600,6 @@ async fn cli_main(codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()
} else if login_cli.with_api_key {
let api_key = read_api_key_from_stdin();
run_login_with_api_key(login_cli.config_overrides, api_key).await;
} else if is_headless_environment() {
run_login_with_device_code_fallback_to_browser(
login_cli.config_overrides,
login_cli.issuer_base_url,
login_cli.client_id,
)
.await;
} else {
run_login_with_chatgpt(login_cli.config_overrides).await;
}
@@ -691,11 +684,11 @@ async fn cli_main(codex_linux_sandbox_exe: Option<PathBuf>) -> anyhow::Result<()
.parse_overrides()
.map_err(anyhow::Error::msg)?;
// Honor `--search` via the new feature toggle.
// Honor `--search` via the canonical web_search mode.
if interactive.web_search {
cli_kv_overrides.push((
"features.web_search_request".to_string(),
toml::Value::Boolean(true),
"web_search".to_string(),
toml::Value::String("live".to_string()),
));
}
@@ -737,9 +730,32 @@ fn prepend_config_flags(
/// Run the interactive Codex TUI, dispatching to either the legacy implementation or the
/// experimental TUI v2 shim based on feature flags resolved from config.
async fn run_interactive_tui(
interactive: TuiCli,
mut interactive: TuiCli,
codex_linux_sandbox_exe: Option<PathBuf>,
) -> std::io::Result<AppExitInfo> {
if let Some(prompt) = interactive.prompt.take() {
// Normalize CRLF/CR to LF so CLI-provided text can't leak `\r` into TUI state.
interactive.prompt = Some(prompt.replace("\r\n", "\n").replace('\r', "\n"));
}
let terminal_info = codex_core::terminal::terminal_info();
if terminal_info.name == TerminalName::Dumb {
if !(std::io::stdin().is_terminal() && std::io::stderr().is_terminal()) {
return Ok(AppExitInfo::fatal(
"TERM is set to \"dumb\". Refusing to start the interactive TUI because no terminal is available for a confirmation prompt (stdin/stderr is not a TTY). Run in a supported terminal or unset TERM.",
));
}
eprintln!(
"WARNING: TERM is set to \"dumb\". Codex's interactive TUI may not work in this terminal."
);
if !confirm("Continue anyway? [y/N]: ")? {
return Ok(AppExitInfo::fatal(
"Refusing to start the interactive TUI because TERM is set to \"dumb\". Run in a supported terminal or unset TERM.",
));
}
}
if is_tui2_enabled(&interactive).await? {
let result = tui2::run_main(interactive.into(), codex_linux_sandbox_exe).await?;
Ok(result.into())
@@ -748,6 +764,15 @@ async fn run_interactive_tui(
}
}
fn confirm(prompt: &str) -> std::io::Result<bool> {
eprintln!("{prompt}");
let mut input = String::new();
std::io::stdin().read_line(&mut input)?;
let answer = input.trim();
Ok(answer.eq_ignore_ascii_case("y") || answer.eq_ignore_ascii_case("yes"))
}
/// Returns `Ok(true)` when the resolved configuration enables the `tui2` feature flag.
///
/// This performs a lightweight config load (honoring the same precedence as the lower-level TUI
@@ -864,7 +889,8 @@ fn merge_interactive_cli_flags(interactive: &mut TuiCli, subcommand_cli: TuiCli)
interactive.add_dir.extend(subcommand_cli.add_dir);
}
if let Some(prompt) = subcommand_cli.prompt {
interactive.prompt = Some(prompt);
// Normalize CRLF/CR to LF so CLI-provided text can't leak `\r` into TUI state.
interactive.prompt = Some(prompt.replace("\r\n", "\n").replace('\r', "\n"));
}
interactive

View File

@@ -241,6 +241,7 @@ async fn run_add(config_overrides: &CliConfigOverrides, add_args: AddArgs) -> Re
let new_entry = McpServerConfig {
transport: transport.clone(),
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -448,6 +449,7 @@ async fn run_list(config_overrides: &CliConfigOverrides, list_args: ListArgs) ->
serde_json::json!({
"name": name,
"enabled": cfg.enabled,
"disabled_reason": cfg.disabled_reason.as_ref().map(ToString::to_string),
"transport": transport,
"startup_timeout_sec": cfg
.startup_timeout_sec
@@ -492,11 +494,7 @@ async fn run_list(config_overrides: &CliConfigOverrides, list_args: ListArgs) ->
.map(|path| path.display().to_string())
.filter(|value| !value.is_empty())
.unwrap_or_else(|| "-".to_string());
let status = if cfg.enabled {
"enabled".to_string()
} else {
"disabled".to_string()
};
let status = format_mcp_status(cfg);
let auth_status = auth_statuses
.get(name.as_str())
.map(|entry| entry.auth_status)
@@ -517,11 +515,7 @@ async fn run_list(config_overrides: &CliConfigOverrides, list_args: ListArgs) ->
bearer_token_env_var,
..
} => {
let status = if cfg.enabled {
"enabled".to_string()
} else {
"disabled".to_string()
};
let status = format_mcp_status(cfg);
let auth_status = auth_statuses
.get(name.as_str())
.map(|entry| entry.auth_status)
@@ -691,6 +685,7 @@ async fn run_get(config_overrides: &CliConfigOverrides, get_args: GetArgs) -> Re
let output = serde_json::to_string_pretty(&serde_json::json!({
"name": get_args.name,
"enabled": server.enabled,
"disabled_reason": server.disabled_reason.as_ref().map(ToString::to_string),
"transport": transport,
"enabled_tools": server.enabled_tools.clone(),
"disabled_tools": server.disabled_tools.clone(),
@@ -706,7 +701,11 @@ async fn run_get(config_overrides: &CliConfigOverrides, get_args: GetArgs) -> Re
}
if !server.enabled {
println!("{} (disabled)", get_args.name);
if let Some(reason) = server.disabled_reason.as_ref() {
println!("{name} (disabled: {reason})", name = get_args.name);
} else {
println!("{name} (disabled)", name = get_args.name);
}
return Ok(());
}
@@ -828,3 +827,13 @@ fn validate_server_name(name: &str) -> Result<()> {
bail!("invalid server name '{name}' (use letters, numbers, '-', '_')");
}
}
fn format_mcp_status(config: &McpServerConfig) -> String {
if config.enabled {
"enabled".to_string()
} else if let Some(reason) = config.disabled_reason.as_ref() {
format!("disabled: {reason}")
} else {
"disabled".to_string()
}
}

View File

@@ -89,6 +89,7 @@ async fn list_and_get_render_expected_output() -> Result<()> {
{
"name": "docs",
"enabled": true,
"disabled_reason": null,
"transport": {
"type": "stdio",
"command": "docs-server",

View File

@@ -94,6 +94,12 @@ pub struct CreatedTask {
pub id: TaskId,
}
#[derive(Clone, Debug, PartialEq, Eq)]
pub struct TaskListPage {
pub tasks: Vec<TaskSummary>,
pub cursor: Option<String>,
}
#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize, Default)]
pub struct DiffSummary {
pub files_changed: usize,
@@ -126,7 +132,12 @@ impl Default for TaskText {
#[async_trait::async_trait]
pub trait CloudBackend: Send + Sync {
async fn list_tasks(&self, env: Option<&str>) -> Result<Vec<TaskSummary>>;
async fn list_tasks(
&self,
env: Option<&str>,
limit: Option<i64>,
cursor: Option<&str>,
) -> Result<TaskListPage>;
async fn get_task_summary(&self, id: TaskId) -> Result<TaskSummary>;
async fn get_task_diff(&self, id: TaskId) -> Result<Option<String>>;
/// Return assistant output messages (no diff) when available.

View File

@@ -6,6 +6,7 @@ use crate::CloudTaskError;
use crate::DiffSummary;
use crate::Result;
use crate::TaskId;
use crate::TaskListPage;
use crate::TaskStatus;
use crate::TaskSummary;
use crate::TurnAttempt;
@@ -59,8 +60,13 @@ impl HttpClient {
#[async_trait::async_trait]
impl CloudBackend for HttpClient {
async fn list_tasks(&self, env: Option<&str>) -> Result<Vec<TaskSummary>> {
self.tasks_api().list(env).await
async fn list_tasks(
&self,
env: Option<&str>,
limit: Option<i64>,
cursor: Option<&str>,
) -> Result<TaskListPage> {
self.tasks_api().list(env, limit, cursor).await
}
async fn get_task_summary(&self, id: TaskId) -> Result<TaskSummary> {
@@ -132,10 +138,16 @@ mod api {
}
}
pub(crate) async fn list(&self, env: Option<&str>) -> Result<Vec<TaskSummary>> {
pub(crate) async fn list(
&self,
env: Option<&str>,
limit: Option<i64>,
cursor: Option<&str>,
) -> Result<TaskListPage> {
let limit_i32 = limit.and_then(|lim| i32::try_from(lim).ok());
let resp = self
.backend
.list_tasks(Some(20), Some("current"), env)
.list_tasks(limit_i32, Some("current"), env, cursor)
.await
.map_err(|e| CloudTaskError::Http(format!("list_tasks failed: {e}")))?;
@@ -146,11 +158,19 @@ mod api {
.collect();
append_error_log(&format!(
"http.list_tasks: env={} items={}",
"http.list_tasks: env={} limit={} cursor_in={} cursor_out={} items={}",
env.unwrap_or("<all>"),
limit_i32
.map(|v| v.to_string())
.unwrap_or_else(|| "<default>".to_string()),
cursor.unwrap_or("<none>"),
resp.cursor.as_deref().unwrap_or("<none>"),
tasks.len()
));
Ok(tasks)
Ok(TaskListPage {
tasks,
cursor: resp.cursor,
})
}
pub(crate) async fn summary(&self, id: TaskId) -> Result<TaskSummary> {

View File

@@ -9,6 +9,7 @@ pub use api::CreatedTask;
pub use api::DiffSummary;
pub use api::Result;
pub use api::TaskId;
pub use api::TaskListPage;
pub use api::TaskStatus;
pub use api::TaskSummary;
pub use api::TaskText;

View File

@@ -16,7 +16,12 @@ pub struct MockClient;
#[async_trait::async_trait]
impl CloudBackend for MockClient {
async fn list_tasks(&self, _env: Option<&str>) -> Result<Vec<TaskSummary>> {
async fn list_tasks(
&self,
_env: Option<&str>,
_limit: Option<i64>,
_cursor: Option<&str>,
) -> Result<crate::TaskListPage> {
// Slightly vary content by env to aid tests that rely on the mock
let rows = match _env {
Some("env-A") => vec![("T-2000", "A: First", TaskStatus::Ready)],
@@ -58,11 +63,14 @@ impl CloudBackend for MockClient {
attempt_total: Some(if id_str == "T-1000" { 2 } else { 1 }),
});
}
Ok(out)
Ok(crate::TaskListPage {
tasks: out,
cursor: None,
})
}
async fn get_task_summary(&self, id: TaskId) -> Result<TaskSummary> {
let tasks = self.list_tasks(None).await?;
let tasks = self.list_tasks(None, None, None).await?.tasks;
tasks
.into_iter()
.find(|t| t.id == id)

View File

@@ -123,9 +123,13 @@ pub async fn load_tasks(
env: Option<&str>,
) -> anyhow::Result<Vec<TaskSummary>> {
// In later milestones, add a small debounce, spinner, and error display.
let tasks = tokio::time::timeout(Duration::from_secs(5), backend.list_tasks(env)).await??;
let tasks = tokio::time::timeout(
Duration::from_secs(5),
backend.list_tasks(env, Some(20), None),
)
.await??;
// Hide review-only tasks from the main list.
let filtered: Vec<TaskSummary> = tasks.into_iter().filter(|t| !t.is_review).collect();
let filtered: Vec<TaskSummary> = tasks.tasks.into_iter().filter(|t| !t.is_review).collect();
Ok(filtered)
}
@@ -362,7 +366,9 @@ mod tests {
async fn list_tasks(
&self,
env: Option<&str>,
) -> codex_cloud_tasks_client::Result<Vec<TaskSummary>> {
limit: Option<i64>,
cursor: Option<&str>,
) -> codex_cloud_tasks_client::Result<codex_cloud_tasks_client::TaskListPage> {
let key = env.map(str::to_string);
let titles = self
.by_env
@@ -383,15 +389,28 @@ mod tests {
attempt_total: Some(1),
});
}
Ok(out)
let max = limit.unwrap_or(i64::MAX);
let max = max.min(20);
let mut limited = Vec::new();
for task in out {
if (limited.len() as i64) >= max {
break;
}
limited.push(task);
}
Ok(codex_cloud_tasks_client::TaskListPage {
tasks: limited,
cursor: cursor.map(str::to_string),
})
}
async fn get_task_summary(
&self,
id: TaskId,
) -> codex_cloud_tasks_client::Result<TaskSummary> {
self.list_tasks(None)
self.list_tasks(None, None, None)
.await?
.tasks
.into_iter()
.find(|t| t.id == id)
.ok_or_else(|| CloudTaskError::Msg(format!("Task {} not found", id.0)))

View File

@@ -18,6 +18,8 @@ pub enum Command {
Exec(ExecCommand),
/// Show the status of a Codex Cloud task.
Status(StatusCommand),
/// List Codex Cloud tasks.
List(ListCommand),
/// Apply the diff for a Codex Cloud task locally.
Apply(ApplyCommand),
/// Show the unified diff for a Codex Cloud task.
@@ -58,6 +60,17 @@ fn parse_attempts(input: &str) -> Result<usize, String> {
}
}
fn parse_limit(input: &str) -> Result<i64, String> {
let value: i64 = input
.parse()
.map_err(|_| "limit must be an integer between 1 and 20".to_string())?;
if (1..=20).contains(&value) {
Ok(value)
} else {
Err("limit must be between 1 and 20".to_string())
}
}
#[derive(Debug, Args)]
pub struct StatusCommand {
/// Codex Cloud task identifier to inspect.
@@ -65,6 +78,25 @@ pub struct StatusCommand {
pub task_id: String,
}
#[derive(Debug, Args)]
pub struct ListCommand {
/// Filter tasks by environment identifier.
#[arg(long = "env", value_name = "ENV_ID")]
pub environment: Option<String>,
/// Maximum number of tasks to return (1-20).
#[arg(long = "limit", default_value_t = 20, value_parser = parse_limit, value_name = "N")]
pub limit: i64,
/// Pagination cursor returned by a previous call.
#[arg(long = "cursor", value_name = "CURSOR")]
pub cursor: Option<String>,
/// Emit JSON instead of plain text.
#[arg(long = "json", default_value_t = false)]
pub json: bool,
}
#[derive(Debug, Args)]
pub struct ApplyCommand {
/// Codex Cloud task identifier to apply.

View File

@@ -393,11 +393,10 @@ fn summary_line(summary: &codex_cloud_tasks_client::DiffSummary, colorize: bool)
let bullet = ""
.if_supports_color(Stream::Stdout, |t| t.dimmed())
.to_string();
let file_label = "file"
let file_label = format!("file{}", if files == 1 { "" } else { "s" })
.if_supports_color(Stream::Stdout, |t| t.dimmed())
.to_string();
let plural = if files == 1 { "" } else { "s" };
format!("{adds_str}/{dels_str} {bullet} {files} {file_label}{plural}")
format!("{adds_str}/{dels_str} {bullet} {files} {file_label}")
} else {
format!(
"+{adds}/-{dels}{files} file{}",
@@ -473,6 +472,25 @@ fn format_task_status_lines(
lines
}
fn format_task_list_lines(
tasks: &[codex_cloud_tasks_client::TaskSummary],
base_url: &str,
now: chrono::DateTime<Utc>,
colorize: bool,
) -> Vec<String> {
let mut lines = Vec::new();
for (idx, task) in tasks.iter().enumerate() {
lines.push(util::task_url(base_url, &task.id.0));
for line in format_task_status_lines(task, now, colorize) {
lines.push(format!(" {line}"));
}
if idx + 1 < tasks.len() {
lines.push(String::new());
}
}
lines
}
async fn run_status_command(args: crate::cli::StatusCommand) -> anyhow::Result<()> {
let ctx = init_backend("codex_cloud_tasks_status").await?;
let task_id = parse_task_id(&args.task_id)?;
@@ -489,6 +507,73 @@ async fn run_status_command(args: crate::cli::StatusCommand) -> anyhow::Result<(
Ok(())
}
async fn run_list_command(args: crate::cli::ListCommand) -> anyhow::Result<()> {
let ctx = init_backend("codex_cloud_tasks_list").await?;
let env_filter = if let Some(env) = args.environment {
Some(resolve_environment_id(&ctx, &env).await?)
} else {
None
};
let page = codex_cloud_tasks_client::CloudBackend::list_tasks(
&*ctx.backend,
env_filter.as_deref(),
Some(args.limit),
args.cursor.as_deref(),
)
.await?;
if args.json {
let tasks: Vec<_> = page
.tasks
.iter()
.map(|task| {
serde_json::json!({
"id": task.id.0,
"url": util::task_url(&ctx.base_url, &task.id.0),
"title": task.title,
"status": task.status,
"updated_at": task.updated_at,
"environment_id": task.environment_id,
"environment_label": task.environment_label,
"summary": {
"files_changed": task.summary.files_changed,
"lines_added": task.summary.lines_added,
"lines_removed": task.summary.lines_removed,
},
"is_review": task.is_review,
"attempt_total": task.attempt_total,
})
})
.collect();
let payload = serde_json::json!({
"tasks": tasks,
"cursor": page.cursor,
});
println!("{}", serde_json::to_string_pretty(&payload)?);
return Ok(());
}
if page.tasks.is_empty() {
println!("No tasks found.");
return Ok(());
}
let now = Utc::now();
let colorize = supports_color::on(SupportStream::Stdout).is_some();
for line in format_task_list_lines(&page.tasks, &ctx.base_url, now, colorize) {
println!("{line}");
}
if let Some(cursor) = page.cursor {
let command = format!("codex cloud list --cursor='{cursor}'");
if colorize {
println!(
"\nTo fetch the next page, run {}",
command.if_supports_color(Stream::Stdout, |text| text.cyan())
);
} else {
println!("\nTo fetch the next page, run {command}");
}
}
Ok(())
}
async fn run_diff_command(args: crate::cli::DiffCommand) -> anyhow::Result<()> {
let ctx = init_backend("codex_cloud_tasks_diff").await?;
let task_id = parse_task_id(&args.task_id)?;
@@ -649,6 +734,7 @@ pub async fn run_main(cli: Cli, _codex_linux_sandbox_exe: Option<PathBuf>) -> an
return match command {
crate::cli::Command::Exec(args) => run_exec_command(args).await,
crate::cli::Command::Status(args) => run_status_command(args).await,
crate::cli::Command::List(args) => run_list_command(args).await,
crate::cli::Command::Apply(args) => run_apply_command(args).await,
crate::cli::Command::Diff(args) => run_diff_command(args).await,
};
@@ -2181,6 +2267,54 @@ mod tests {
);
}
#[test]
fn format_task_list_lines_formats_urls() {
let now = Utc::now();
let tasks = vec![
TaskSummary {
id: TaskId("task_1".to_string()),
title: "Example task".to_string(),
status: TaskStatus::Ready,
updated_at: now,
environment_id: Some("env-1".to_string()),
environment_label: Some("Env".to_string()),
summary: DiffSummary {
files_changed: 3,
lines_added: 5,
lines_removed: 2,
},
is_review: false,
attempt_total: None,
},
TaskSummary {
id: TaskId("task_2".to_string()),
title: "No diff task".to_string(),
status: TaskStatus::Pending,
updated_at: now,
environment_id: Some("env-2".to_string()),
environment_label: None,
summary: DiffSummary::default(),
is_review: false,
attempt_total: Some(1),
},
];
let lines = format_task_list_lines(&tasks, "https://chatgpt.com/backend-api", now, false);
assert_eq!(
lines,
vec![
"https://chatgpt.com/codex/tasks/task_1".to_string(),
" [READY] Example task".to_string(),
" Env • 0s ago".to_string(),
" +5/-2 • 3 files".to_string(),
String::new(),
"https://chatgpt.com/codex/tasks/task_2".to_string(),
" [PENDING] No diff task".to_string(),
" env-2 • 0s ago".to_string(),
" no diff".to_string(),
]
);
}
#[tokio::test]
async fn collect_attempt_diffs_includes_sibling_attempts() {
let backend = MockClient;

View File

@@ -5,18 +5,23 @@ use codex_cloud_tasks_client::MockClient;
async fn mock_backend_varies_by_env() {
let client = MockClient;
let root = CloudBackend::list_tasks(&client, None).await.unwrap();
let root = CloudBackend::list_tasks(&client, None, None, None)
.await
.unwrap()
.tasks;
assert!(root.iter().any(|t| t.title.contains("Update README")));
let a = CloudBackend::list_tasks(&client, Some("env-A"))
let a = CloudBackend::list_tasks(&client, Some("env-A"), None, None)
.await
.unwrap();
.unwrap()
.tasks;
assert_eq!(a.len(), 1);
assert_eq!(a[0].title, "A: First");
let b = CloudBackend::list_tasks(&client, Some("env-B"))
let b = CloudBackend::list_tasks(&client, Some("env-B"), None, None)
.await
.unwrap();
.unwrap()
.tasks;
assert_eq!(b.len(), 2);
assert!(b[0].title.starts_with("B: "));
}

View File

@@ -42,6 +42,10 @@ pub enum ResponseEvent {
Created,
OutputItemDone(ResponseItem),
OutputItemAdded(ResponseItem),
/// Emitted when `X-Reasoning-Included: true` is present on the response,
/// meaning the server already accounted for past reasoning tokens and the
/// client should not re-estimate them.
ServerReasoningIncluded(bool),
Completed {
response_id: String,
token_usage: Option<TokenUsage>,

View File

@@ -87,6 +87,7 @@ impl<T: HttpTransport, A: AuthProvider> ChatClient<T, A> {
extra_headers,
RequestCompression::None,
spawn_chat_stream,
None,
)
.await
}
@@ -156,6 +157,9 @@ impl Stream for AggregatedStream {
return Poll::Ready(Some(Ok(ResponseEvent::OutputItemDone(item))));
}
Poll::Ready(Some(Ok(ResponseEvent::ServerReasoningIncluded(included)))) => {
return Poll::Ready(Some(Ok(ResponseEvent::ServerReasoningIncluded(included))));
}
Poll::Ready(Some(Ok(ResponseEvent::RateLimits(snapshot)))) => {
return Poll::Ready(Some(Ok(ResponseEvent::RateLimits(snapshot))));
}

View File

@@ -19,6 +19,7 @@ use codex_protocol::protocol::SessionSource;
use http::HeaderMap;
use serde_json::Value;
use std::sync::Arc;
use std::sync::OnceLock;
use tracing::instrument;
pub struct ResponsesClient<T: HttpTransport, A: AuthProvider> {
@@ -36,6 +37,7 @@ pub struct ResponsesOptions {
pub session_source: Option<SessionSource>,
pub extra_headers: HeaderMap,
pub compression: Compression,
pub turn_state: Option<Arc<OnceLock<String>>>,
}
impl<T: HttpTransport, A: AuthProvider> ResponsesClient<T, A> {
@@ -58,9 +60,15 @@ impl<T: HttpTransport, A: AuthProvider> ResponsesClient<T, A> {
pub async fn stream_request(
&self,
request: ResponsesRequest,
turn_state: Option<Arc<OnceLock<String>>>,
) -> Result<ResponseStream, ApiError> {
self.stream(request.body, request.headers, request.compression)
.await
self.stream(
request.body,
request.headers,
request.compression,
turn_state,
)
.await
}
#[instrument(level = "trace", skip_all, err)]
@@ -80,6 +88,7 @@ impl<T: HttpTransport, A: AuthProvider> ResponsesClient<T, A> {
session_source,
extra_headers,
compression,
turn_state,
} = options;
let request = ResponsesRequestBuilder::new(model, &prompt.instructions, &prompt.input)
@@ -96,7 +105,7 @@ impl<T: HttpTransport, A: AuthProvider> ResponsesClient<T, A> {
.compression(compression)
.build(self.streaming.provider())?;
self.stream_request(request).await
self.stream_request(request, turn_state).await
}
fn path(&self) -> &'static str {
@@ -111,6 +120,7 @@ impl<T: HttpTransport, A: AuthProvider> ResponsesClient<T, A> {
body: Value,
extra_headers: HeaderMap,
compression: Compression,
turn_state: Option<Arc<OnceLock<String>>>,
) -> Result<ResponseStream, ApiError> {
let compression = match compression {
Compression::None => RequestCompression::None,
@@ -124,6 +134,7 @@ impl<T: HttpTransport, A: AuthProvider> ResponsesClient<T, A> {
extra_headers,
compression,
spawn_response_stream,
turn_state,
)
.await
}

View File

@@ -13,6 +13,7 @@ use http::HeaderMap;
use http::HeaderValue;
use serde_json::Value;
use std::sync::Arc;
use std::sync::OnceLock;
use std::time::Duration;
use tokio::net::TcpStream;
use tokio::sync::Mutex;
@@ -27,18 +28,22 @@ use tracing::trace;
use url::Url;
type WsStream = WebSocketStream<MaybeTlsStream<TcpStream>>;
const X_CODEX_TURN_STATE_HEADER: &str = "x-codex-turn-state";
const X_REASONING_INCLUDED_HEADER: &str = "x-reasoning-included";
pub struct ResponsesWebsocketConnection {
stream: Arc<Mutex<Option<WsStream>>>,
// TODO (pakrym): is this the right place for timeout?
idle_timeout: Duration,
server_reasoning_included: bool,
}
impl ResponsesWebsocketConnection {
fn new(stream: WsStream, idle_timeout: Duration) -> Self {
fn new(stream: WsStream, idle_timeout: Duration, server_reasoning_included: bool) -> Self {
Self {
stream: Arc::new(Mutex::new(Some(stream))),
idle_timeout,
server_reasoning_included,
}
}
@@ -54,11 +59,17 @@ impl ResponsesWebsocketConnection {
mpsc::channel::<std::result::Result<ResponseEvent, ApiError>>(1600);
let stream = Arc::clone(&self.stream);
let idle_timeout = self.idle_timeout;
let server_reasoning_included = self.server_reasoning_included;
let request_body = serde_json::to_value(&request).map_err(|err| {
ApiError::Stream(format!("failed to encode websocket request: {err}"))
})?;
tokio::spawn(async move {
if server_reasoning_included {
let _ = tx_event
.send(Ok(ResponseEvent::ServerReasoningIncluded(true)))
.await;
}
let mut guard = stream.lock().await;
let Some(ws_stream) = guard.as_mut() else {
let _ = tx_event
@@ -100,18 +111,23 @@ impl<A: AuthProvider> ResponsesWebsocketClient<A> {
pub async fn connect(
&self,
extra_headers: HeaderMap,
turn_state: Option<Arc<OnceLock<String>>>,
) -> Result<ResponsesWebsocketConnection, ApiError> {
let ws_url = Url::parse(&self.provider.url_for_path("responses"))
let ws_url = self
.provider
.websocket_url_for_path("responses")
.map_err(|err| ApiError::Stream(format!("failed to build websocket URL: {err}")))?;
let mut headers = self.provider.headers.clone();
headers.extend(extra_headers);
apply_auth_headers(&mut headers, &self.auth);
let stream = connect_websocket(ws_url, headers).await?;
let (stream, server_reasoning_included) =
connect_websocket(ws_url, headers, turn_state).await?;
Ok(ResponsesWebsocketConnection::new(
stream,
self.provider.stream_idle_timeout,
server_reasoning_included,
))
}
}
@@ -130,17 +146,30 @@ fn apply_auth_headers(headers: &mut HeaderMap, auth: &impl AuthProvider) {
}
}
async fn connect_websocket(url: Url, headers: HeaderMap) -> Result<WsStream, ApiError> {
async fn connect_websocket(
url: Url,
headers: HeaderMap,
turn_state: Option<Arc<OnceLock<String>>>,
) -> Result<(WsStream, bool), ApiError> {
let mut request = url
.clone()
.as_str()
.into_client_request()
.map_err(|err| ApiError::Stream(format!("failed to build websocket request: {err}")))?;
request.headers_mut().extend(headers);
let (stream, _) = tokio_tungstenite::connect_async(request)
let (stream, response) = tokio_tungstenite::connect_async(request)
.await
.map_err(|err| map_ws_error(err, &url))?;
Ok(stream)
let reasoning_included = response.headers().contains_key(X_REASONING_INCLUDED_HEADER);
if let Some(turn_state) = turn_state
&& let Some(header_value) = response
.headers()
.get(X_CODEX_TURN_STATE_HEADER)
.and_then(|value| value.to_str().ok())
{
let _ = turn_state.set(header_value.to_string());
}
Ok((stream, reasoning_included))
}
fn map_ws_error(err: WsError, url: &Url) -> ApiError {
@@ -182,7 +211,7 @@ async fn run_websocket_response_stream(
}
};
if let Err(err) = ws_stream.send(Message::Text(request_text)).await {
if let Err(err) = ws_stream.send(Message::Text(request_text.into())).await {
return Err(ApiError::Stream(format!(
"failed to send websocket request: {err}"
)));

View File

@@ -13,6 +13,7 @@ use http::HeaderMap;
use http::Method;
use serde_json::Value;
use std::sync::Arc;
use std::sync::OnceLock;
use std::time::Duration;
pub(crate) struct StreamingClient<T: HttpTransport, A: AuthProvider> {
@@ -23,6 +24,13 @@ pub(crate) struct StreamingClient<T: HttpTransport, A: AuthProvider> {
sse_telemetry: Option<Arc<dyn SseTelemetry>>,
}
type StreamSpawner = fn(
StreamResponse,
Duration,
Option<Arc<dyn SseTelemetry>>,
Option<Arc<OnceLock<String>>>,
) -> ResponseStream;
impl<T: HttpTransport, A: AuthProvider> StreamingClient<T, A> {
pub(crate) fn new(transport: T, provider: Provider, auth: A) -> Self {
Self {
@@ -54,7 +62,8 @@ impl<T: HttpTransport, A: AuthProvider> StreamingClient<T, A> {
body: Value,
extra_headers: HeaderMap,
compression: RequestCompression,
spawner: fn(StreamResponse, Duration, Option<Arc<dyn SseTelemetry>>) -> ResponseStream,
spawner: StreamSpawner,
turn_state: Option<Arc<OnceLock<String>>>,
) -> Result<ResponseStream, ApiError> {
let builder = || {
let mut req = self.provider.build_request(Method::POST, path);
@@ -80,6 +89,7 @@ impl<T: HttpTransport, A: AuthProvider> StreamingClient<T, A> {
stream_response,
self.provider.stream_idle_timeout,
self.sse_telemetry.clone(),
turn_state,
))
}
}

View File

@@ -25,6 +25,8 @@ pub enum ApiError {
},
#[error("rate limit: {0}")]
RateLimit(String),
#[error("invalid request: {message}")]
InvalidRequest { message: String },
}
impl From<RateLimitError> for ApiError {

View File

@@ -6,6 +6,7 @@ use http::Method;
use http::header::HeaderMap;
use std::collections::HashMap;
use std::time::Duration;
use url::Url;
/// Wire-level APIs supported by a `Provider`.
#[derive(Debug, Clone, PartialEq, Eq)]
@@ -105,6 +106,19 @@ impl Provider {
self.base_url.to_ascii_lowercase().contains("openai.azure.")
|| matches_azure_responses_base_url(&self.base_url)
}
pub fn websocket_url_for_path(&self, path: &str) -> Result<Url, url::ParseError> {
let mut url = Url::parse(&self.url_for_path(path))?;
let scheme = match url.scheme() {
"http" => "ws",
"https" => "wss",
"ws" | "wss" => return Ok(url),
_ => return Ok(url),
};
let _ = url.set_scheme(scheme);
Ok(url)
}
}
fn matches_azure_responses_base_url(base_url: &str) -> bool {

View File

@@ -11,6 +11,8 @@ use futures::Stream;
use futures::StreamExt;
use std::collections::HashMap;
use std::collections::HashSet;
use std::sync::Arc;
use std::sync::OnceLock;
use std::time::Duration;
use tokio::sync::mpsc;
use tokio::time::Instant;
@@ -21,7 +23,8 @@ use tracing::trace;
pub(crate) fn spawn_chat_stream(
stream_response: StreamResponse,
idle_timeout: Duration,
telemetry: Option<std::sync::Arc<dyn SseTelemetry>>,
telemetry: Option<Arc<dyn SseTelemetry>>,
_turn_state: Option<Arc<OnceLock<String>>>,
) -> ResponseStream {
let (tx_event, rx_event) = mpsc::channel::<Result<ResponseEvent, ApiError>>(1600);
tokio::spawn(async move {

View File

@@ -16,6 +16,7 @@ use serde_json::Value;
use std::io::BufRead;
use std::path::Path;
use std::sync::Arc;
use std::sync::OnceLock;
use std::time::Duration;
use tokio::sync::mpsc;
use tokio::time::Instant;
@@ -24,6 +25,8 @@ use tokio_util::io::ReaderStream;
use tracing::debug;
use tracing::trace;
const X_REASONING_INCLUDED_HEADER: &str = "x-reasoning-included";
/// Streams SSE events from an on-disk fixture for tests.
pub fn stream_from_fixture(
path: impl AsRef<Path>,
@@ -49,6 +52,7 @@ pub fn spawn_response_stream(
stream_response: StreamResponse,
idle_timeout: Duration,
telemetry: Option<Arc<dyn SseTelemetry>>,
turn_state: Option<Arc<OnceLock<String>>>,
) -> ResponseStream {
let rate_limits = parse_rate_limit(&stream_response.headers);
let models_etag = stream_response
@@ -56,6 +60,18 @@ pub fn spawn_response_stream(
.get("X-Models-Etag")
.and_then(|v| v.to_str().ok())
.map(ToString::to_string);
let reasoning_included = stream_response
.headers
.get(X_REASONING_INCLUDED_HEADER)
.is_some();
if let Some(turn_state) = turn_state.as_ref()
&& let Some(header_value) = stream_response
.headers
.get("x-codex-turn-state")
.and_then(|v| v.to_str().ok())
{
let _ = turn_state.set(header_value.to_string());
}
let (tx_event, rx_event) = mpsc::channel::<Result<ResponseEvent, ApiError>>(1600);
tokio::spawn(async move {
if let Some(snapshot) = rate_limits {
@@ -64,6 +80,11 @@ pub fn spawn_response_stream(
if let Some(etag) = models_etag {
let _ = tx_event.send(Ok(ResponseEvent::ModelsEtag(etag))).await;
}
if reasoning_included {
let _ = tx_event
.send(Ok(ResponseEvent::ServerReasoningIncluded(true)))
.await;
}
process_sse(stream_response.bytes, tx_event, idle_timeout, telemetry).await;
});
@@ -207,6 +228,11 @@ pub fn process_responses_event(
response_error = ApiError::QuotaExceeded;
} else if is_usage_not_included(&error) {
response_error = ApiError::UsageNotIncluded;
} else if is_invalid_prompt_error(&error) {
let message = error
.message
.unwrap_or_else(|| "Invalid request.".to_string());
response_error = ApiError::InvalidRequest { message };
} else {
let delay = try_parse_retry_after(&error);
let message = error.message.unwrap_or_default();
@@ -290,7 +316,6 @@ pub async fn process_sse(
telemetry: Option<Arc<dyn SseTelemetry>>,
) {
let mut stream = stream.eventsource();
let mut response_completed: Option<ResponseEvent> = None;
let mut response_error: Option<ApiError> = None;
loop {
@@ -307,17 +332,10 @@ pub async fn process_sse(
return;
}
Ok(None) => {
match response_completed.take() {
Some(event) => {
let _ = tx_event.send(Ok(event)).await;
}
None => {
let error = response_error.unwrap_or(ApiError::Stream(
"stream closed before response.completed".into(),
));
let _ = tx_event.send(Err(error)).await;
}
}
let error = response_error.unwrap_or(ApiError::Stream(
"stream closed before response.completed".into(),
));
let _ = tx_event.send(Err(error)).await;
return;
}
Err(_) => {
@@ -328,8 +346,7 @@ pub async fn process_sse(
}
};
let raw = sse.data.clone();
trace!("SSE event: {raw}");
trace!("SSE event: {}", &sse.data);
let event: ResponsesStreamEvent = match serde_json::from_str(&sse.data) {
Ok(event) => event,
@@ -341,9 +358,11 @@ pub async fn process_sse(
match process_responses_event(event) {
Ok(Some(event)) => {
if matches!(event, ResponseEvent::Completed { .. }) {
response_completed = Some(event);
} else if tx_event.send(Ok(event)).await.is_err() {
let is_completed = matches!(event, ResponseEvent::Completed { .. });
if tx_event.send(Ok(event)).await.is_err() {
return;
}
if is_completed {
return;
}
}
@@ -393,6 +412,10 @@ fn is_usage_not_included(error: &Error) -> bool {
error.code.as_deref() == Some("usage_not_included")
}
fn is_invalid_prompt_error(error: &Error) -> bool {
error.code.as_deref() == Some("invalid_prompt")
}
fn rate_limit_regex() -> &'static regex_lite::Regex {
static RE: std::sync::OnceLock<regex_lite::Regex> = std::sync::OnceLock::new();
#[expect(clippy::unwrap_used)]
@@ -405,7 +428,9 @@ fn rate_limit_regex() -> &'static regex_lite::Regex {
mod tests {
use super::*;
use assert_matches::assert_matches;
use bytes::Bytes;
use codex_protocol::models::ResponseItem;
use futures::stream;
use pretty_assertions::assert_eq;
use serde_json::json;
use tokio::sync::mpsc;
@@ -607,6 +632,44 @@ mod tests {
}
}
#[tokio::test]
async fn emits_completed_without_stream_end() {
let completed = json!({
"type": "response.completed",
"response": { "id": "resp1" }
})
.to_string();
let sse1 = format!("event: response.completed\ndata: {completed}\n\n");
let stream = stream::iter(vec![Ok(Bytes::from(sse1))]).chain(stream::pending());
let stream: ByteStream = Box::pin(stream);
let (tx, mut rx) = mpsc::channel::<Result<ResponseEvent, ApiError>>(8);
tokio::spawn(process_sse(stream, tx, idle_timeout(), None));
let events = tokio::time::timeout(Duration::from_millis(1000), async {
let mut events = Vec::new();
while let Some(ev) = rx.recv().await {
events.push(ev);
}
events
})
.await
.expect("timed out collecting events");
assert_eq!(events.len(), 1);
match &events[0] {
Ok(ResponseEvent::Completed {
response_id,
token_usage,
}) => {
assert_eq!(response_id, "resp1");
assert!(token_usage.is_none());
}
other => panic!("unexpected event: {other:?}"),
}
}
#[tokio::test]
async fn error_when_error_event() {
let raw_error = r#"{"type":"response.failed","sequence_number":3,"response":{"id":"resp_689bcf18d7f08194bf3440ba62fe05d803fee0cdac429894","object":"response","created_at":1755041560,"status":"failed","background":false,"error":{"code":"rate_limit_exceeded","message":"Rate limit reached for gpt-5.1 in organization org-AAA on tokens per min (TPM): Limit 30000, Used 22999, Requested 12528. Please try again in 11.054s. Visit https://platform.openai.com/account/rate-limits to learn more."}, "usage":null,"user":null,"metadata":{}}}"#;
@@ -668,6 +731,27 @@ mod tests {
assert_matches!(events[0], Err(ApiError::QuotaExceeded));
}
#[tokio::test]
async fn invalid_prompt_without_type_is_invalid_request() {
let raw_error = r#"{"type":"response.failed","sequence_number":3,"response":{"id":"resp_invalid_prompt_no_type","object":"response","created_at":1759771628,"status":"failed","background":false,"error":{"code":"invalid_prompt","message":"Invalid prompt: we've limited access to this content for safety reasons."},"incomplete_details":null}}"#;
let sse1 = format!("event: response.failed\ndata: {raw_error}\n\n");
let events = collect_events(&[sse1.as_bytes()]).await;
assert_eq!(events.len(), 1);
match &events[0] {
Err(ApiError::InvalidRequest { message }) => {
assert_eq!(
message,
"Invalid prompt: we've limited access to this content for safety reasons."
);
}
other => panic!("unexpected event: {other:?}"),
}
}
#[tokio::test]
async fn table_driven_event_kinds() {
struct TestCase {

View File

@@ -231,7 +231,7 @@ async fn responses_client_uses_responses_path_for_responses_wire() -> Result<()>
let body = serde_json::json!({ "echo": true });
let _stream = client
.stream(body, HeaderMap::new(), Compression::None)
.stream(body, HeaderMap::new(), Compression::None, None)
.await?;
let requests = state.take_stream_requests();
@@ -247,7 +247,7 @@ async fn responses_client_uses_chat_path_for_chat_wire() -> Result<()> {
let body = serde_json::json!({ "echo": true });
let _stream = client
.stream(body, HeaderMap::new(), Compression::None)
.stream(body, HeaderMap::new(), Compression::None, None)
.await?;
let requests = state.take_stream_requests();
@@ -264,7 +264,7 @@ async fn streaming_client_adds_auth_headers() -> Result<()> {
let body = serde_json::json!({ "model": "gpt-test" });
let _stream = client
.stream(body, HeaderMap::new(), Compression::None)
.stream(body, HeaderMap::new(), Compression::None, None)
.await?;
let requests = state.take_stream_requests();

View File

@@ -129,6 +129,7 @@ async fn responses_stream_parses_items_and_completed_end_to_end() -> Result<()>
serde_json::json!({"echo": true}),
HeaderMap::new(),
Compression::None,
None,
)
.await?;
@@ -198,6 +199,7 @@ async fn responses_stream_aggregates_output_text_deltas() -> Result<()> {
serde_json::json!({"echo": true}),
HeaderMap::new(),
Compression::None,
None,
)
.await?;

View File

@@ -42,9 +42,12 @@ impl RateLimitStatusPayload {
}
}
#[derive(Clone, Copy, Debug, Eq, PartialEq, Ord, PartialOrd, Hash, Serialize, Deserialize)]
#[derive(
Clone, Copy, Debug, Eq, PartialEq, Ord, PartialOrd, Hash, Serialize, Deserialize, Default,
)]
pub enum PlanType {
#[serde(rename = "guest")]
#[default]
Guest,
#[serde(rename = "free")]
Free,
@@ -71,9 +74,3 @@ pub enum PlanType {
#[serde(rename = "edu")]
Edu,
}
impl Default for PlanType {
fn default() -> PlanType {
Self::Guest
}
}

View File

@@ -8,7 +8,6 @@ use reqwest::header::HeaderMap;
use reqwest::header::HeaderName;
use reqwest::header::HeaderValue;
use serde::Serialize;
use std::collections::HashMap;
use std::fmt::Display;
use std::time::Duration;
use tracing::Span;
@@ -116,12 +115,11 @@ impl CodexRequestBuilder {
match self.builder.headers(headers).send().await {
Ok(response) => {
let request_ids = Self::extract_request_ids(&response);
tracing::debug!(
method = %self.method,
url = %self.url,
status = %response.status(),
request_ids = ?request_ids,
headers = ?response.headers(),
version = ?response.version(),
"Request completed"
);
@@ -141,18 +139,6 @@ impl CodexRequestBuilder {
}
}
}
fn extract_request_ids(response: &Response) -> HashMap<String, String> {
["cf-ray", "x-request-id", "x-oai-request-id"]
.iter()
.filter_map(|&name| {
let header_name = HeaderName::from_static(name);
let value = response.headers().get(header_name)?;
let value = value.to_str().ok()?.to_owned();
Some((name.to_owned(), value))
})
.collect()
}
}
struct HeaderMapInjector<'a>(&'a mut HeaderMap);

View File

@@ -16,7 +16,7 @@ pub use sandbox_mode_cli_arg::SandboxModeCliArg;
#[cfg(feature = "cli")]
pub mod format_env_display;
#[cfg(any(feature = "cli", test))]
#[cfg(feature = "cli")]
mod config_override;
#[cfg(feature = "cli")]

View File

@@ -18,6 +18,7 @@ codex_rust_crate(
),
integration_compile_data_extra = [
"//codex-rs/apply-patch:apply_patch_tool_instructions.md",
"models.json",
"prompt.md",
],
test_data_extra = [

View File

@@ -18,7 +18,7 @@ workspace = true
[dependencies]
anyhow = { workspace = true }
arc-swap = "1.7.1"
arc-swap = "1.8.0"
async-channel = { workspace = true }
async-trait = { workspace = true }
base64 = { workspace = true }

View File

@@ -10,6 +10,10 @@ Note that `codex-core` makes some assumptions about certain helper utilities bei
Expects `/usr/bin/sandbox-exec` to be present.
When using the workspace-write sandbox policy, the Seatbelt profile allows
writes under the configured writable roots while keeping `.git` (directory or
pointer file), the resolved `gitdir:` target, and `.codex` read-only.
### Linux
Expects the binary containing `codex-core` to run the equivalent of `codex sandbox linux` (legacy alias: `codex debug landlock`) when `arg0` is `codex-linux-sandbox`. See the `codex-arg0` crate for details.

View File

@@ -53,14 +53,6 @@
"experimental_compact_prompt_file": {
"$ref": "#/definitions/AbsolutePathBuf"
},
"experimental_instructions_file": {
"description": "Legacy, now use features",
"allOf": [
{
"$ref": "#/definitions/AbsolutePathBuf"
}
]
},
"experimental_use_freeform_apply_patch": {
"type": "boolean"
},
@@ -75,9 +67,15 @@
"apply_patch_freeform": {
"type": "boolean"
},
"child_agents_md": {
"type": "boolean"
},
"collab": {
"type": "boolean"
},
"collaboration_modes": {
"type": "boolean"
},
"elevated_windows_sandbox": {
"type": "boolean"
},
@@ -99,9 +97,6 @@
"experimental_windows_sandbox": {
"type": "boolean"
},
"hierarchical_agents": {
"type": "boolean"
},
"include_apply_patch_tool": {
"type": "boolean"
},
@@ -114,6 +109,9 @@
"remote_models": {
"type": "boolean"
},
"responses_websockets": {
"type": "boolean"
},
"shell_snapshot": {
"type": "boolean"
},
@@ -237,6 +235,22 @@
"type": "integer",
"format": "int64"
},
"model_instructions_file": {
"description": "Optional path to a file containing model instructions that will override the built-in instructions for the selected model. Users are STRONGLY DISCOURAGED from using this field, as deviating from the instructions sanctioned by Codex will likely degrade model performance.",
"allOf": [
{
"$ref": "#/definitions/AbsolutePathBuf"
}
]
},
"model_personality": {
"description": "EXPERIMENTAL Optionally specify a personality for the model",
"allOf": [
{
"$ref": "#/definitions/Personality"
}
]
},
"model_provider": {
"description": "Provider to use from the model_providers map.",
"type": "string"
@@ -373,6 +387,14 @@
"description": "When set to `true`, `AgentReasoningRawContentEvent` events will be shown in the UI/output. Defaults to `false`.",
"type": "boolean"
},
"skills": {
"description": "User-level skill config entries keyed by SKILL.md path.",
"allOf": [
{
"$ref": "#/definitions/SkillsConfig"
}
]
},
"tool_output_token_limit": {
"description": "Token budget applied when storing tool/function outputs in the context manager.",
"type": "integer",
@@ -395,6 +417,14 @@
}
]
},
"web_search": {
"description": "Controls the web search tool mode: disabled, cached, or live.",
"allOf": [
{
"$ref": "#/definitions/WebSearchMode"
}
]
},
"windows_wsl_setup_acknowledged": {
"description": "Tracks whether the Windows onboarding screen has been acknowledged.",
"type": "boolean"
@@ -518,9 +548,6 @@
"experimental_compact_prompt_file": {
"$ref": "#/definitions/AbsolutePathBuf"
},
"experimental_instructions_file": {
"$ref": "#/definitions/AbsolutePathBuf"
},
"experimental_use_freeform_apply_patch": {
"type": "boolean"
},
@@ -535,9 +562,15 @@
"apply_patch_freeform": {
"type": "boolean"
},
"child_agents_md": {
"type": "boolean"
},
"collab": {
"type": "boolean"
},
"collaboration_modes": {
"type": "boolean"
},
"elevated_windows_sandbox": {
"type": "boolean"
},
@@ -559,9 +592,6 @@
"experimental_windows_sandbox": {
"type": "boolean"
},
"hierarchical_agents": {
"type": "boolean"
},
"include_apply_patch_tool": {
"type": "boolean"
},
@@ -574,6 +604,9 @@
"remote_models": {
"type": "boolean"
},
"responses_websockets": {
"type": "boolean"
},
"shell_snapshot": {
"type": "boolean"
},
@@ -610,6 +643,17 @@
"model": {
"type": "string"
},
"model_instructions_file": {
"description": "Optional path to a file containing model instructions.",
"allOf": [
{
"$ref": "#/definitions/AbsolutePathBuf"
}
]
},
"model_personality": {
"$ref": "#/definitions/Personality"
},
"model_provider": {
"description": "The key in the `model_providers` map identifying the [`ModelProviderInfo`] to use.",
"type": "string"
@@ -634,6 +678,9 @@
},
"tools_web_search": {
"type": "boolean"
},
"web_search": {
"$ref": "#/definitions/WebSearchMode"
}
},
"additionalProperties": false
@@ -1026,6 +1073,13 @@
},
"additionalProperties": false
},
"Personality": {
"type": "string",
"enum": [
"friendly",
"pragmatic"
]
},
"ProjectConfig": {
"type": "object",
"properties": {
@@ -1277,6 +1331,34 @@
},
"additionalProperties": false
},
"SkillConfig": {
"type": "object",
"required": [
"enabled",
"path"
],
"properties": {
"enabled": {
"type": "boolean"
},
"path": {
"$ref": "#/definitions/AbsolutePathBuf"
}
},
"additionalProperties": false
},
"SkillsConfig": {
"type": "object",
"properties": {
"config": {
"type": "array",
"items": {
"$ref": "#/definitions/SkillConfig"
}
}
},
"additionalProperties": false
},
"ToolsToml": {
"type": "object",
"properties": {
@@ -1420,6 +1502,14 @@
"high"
]
},
"WebSearchMode": {
"type": "string",
"enum": [
"disabled",
"cached",
"live"
]
},
"WireApi": {
"description": "Wire protocol that the provider speaks. Most third-party services only implement the classic OpenAI Chat Completions JSON schema, whereas OpenAI itself (and a handful of others) additionally expose the more modern *Responses* API. The two protocols use different request/response shapes and *cannot* be auto-detected at runtime, therefore each provider entry must declare which one it expects.",
"oneOf": [

File diff suppressed because one or more lines are too long

View File

@@ -1,10 +1,8 @@
use crate::CodexThread;
use crate::agent::AgentStatus;
use crate::error::CodexErr;
use crate::error::Result as CodexResult;
use crate::thread_manager::ThreadManagerState;
use codex_protocol::ThreadId;
use codex_protocol::protocol::EventMsg;
use codex_protocol::protocol::Op;
use codex_protocol::user_input::UserInput;
use std::sync::Arc;
@@ -29,21 +27,18 @@ impl AgentControl {
}
/// Spawn a new agent thread and submit the initial prompt.
///
/// If `headless` is true, a background drain task is spawned to prevent unbounded event growth
/// of the channel queue when there is no client actively reading the thread events.
pub(crate) async fn spawn_agent(
&self,
config: crate::config::Config,
prompt: String,
headless: bool,
) -> CodexResult<ThreadId> {
let state = self.upgrade()?;
let new_thread = state.spawn_new_thread(config, self.clone()).await?;
if headless {
spawn_headless_drain(Arc::clone(&new_thread.thread));
}
// Notify a new thread has been created. This notification will be processed by clients
// to subscribe or drain this newly created thread.
// TODO(jif) add helper for drain
state.notify_thread_created(new_thread.thread_id);
self.send_prompt(new_thread.thread_id, prompt).await?;
@@ -61,7 +56,11 @@ impl AgentControl {
.send_op(
agent_id,
Op::UserInput {
items: vec![UserInput::Text { text: prompt }],
items: vec![UserInput::Text {
text: prompt,
// Agent control prompts are plain text with no UI text elements.
text_elements: Vec::new(),
}],
final_output_json_schema: None,
},
)
@@ -72,6 +71,12 @@ impl AgentControl {
result
}
/// Interrupt the current task for an existing agent thread.
pub(crate) async fn interrupt_agent(&self, agent_id: ThreadId) -> CodexResult<String> {
let state = self.upgrade()?;
state.send_op(agent_id, Op::Interrupt).await
}
/// Submit a shutdown request to an existing agent thread.
pub(crate) async fn shutdown_agent(&self, agent_id: ThreadId) -> CodexResult<String> {
let state = self.upgrade()?;
@@ -80,7 +85,6 @@ impl AgentControl {
result
}
#[allow(dead_code)] // Will be used for collab tools.
/// Fetch the last known status for `agent_id`, returning `NotFound` when unavailable.
pub(crate) async fn get_status(&self, agent_id: ThreadId) -> AgentStatus {
let Ok(state) = self.upgrade() else {
@@ -110,38 +114,18 @@ impl AgentControl {
}
}
/// When an agent is spawned "headless" (no UI/view attached), there may be no consumer polling
/// `CodexThread::next_event()`. The underlying event channel is unbounded, so the producer can
/// accumulate events indefinitely. This drain task prevents that memory growth by polling and
/// discarding events until shutdown.
fn spawn_headless_drain(thread: Arc<CodexThread>) {
tokio::spawn(async move {
loop {
match thread.next_event().await {
Ok(event) => {
if matches!(event.msg, EventMsg::ShutdownComplete) {
break;
}
}
Err(err) => {
tracing::warn!("failed to receive event from agent: {err:?}");
break;
}
}
}
});
}
#[cfg(test)]
mod tests {
use super::*;
use crate::CodexAuth;
use crate::CodexThread;
use crate::ThreadManager;
use crate::agent::agent_status_from_event;
use crate::config::Config;
use crate::config::ConfigBuilder;
use assert_matches::assert_matches;
use codex_protocol::protocol::ErrorEvent;
use codex_protocol::protocol::EventMsg;
use codex_protocol::protocol::TurnAbortReason;
use codex_protocol::protocol::TurnAbortedEvent;
use codex_protocol::protocol::TurnCompleteEvent;
@@ -262,7 +246,7 @@ mod tests {
let control = AgentControl::default();
let (_home, config) = test_config().await;
let err = control
.spawn_agent(config, "hello".to_string(), false)
.spawn_agent(config, "hello".to_string())
.await
.expect_err("spawn_agent should fail without a manager");
assert_eq!(
@@ -346,6 +330,7 @@ mod tests {
Op::UserInput {
items: vec![UserInput::Text {
text: "hello from tests".to_string(),
text_elements: Vec::new(),
}],
final_output_json_schema: None,
},
@@ -363,7 +348,7 @@ mod tests {
let harness = AgentControlHarness::new().await;
let thread_id = harness
.control
.spawn_agent(harness.config.clone(), "spawned".to_string(), false)
.spawn_agent(harness.config.clone(), "spawned".to_string())
.await
.expect("spawn_agent should succeed");
let _thread = harness
@@ -376,6 +361,7 @@ mod tests {
Op::UserInput {
items: vec![UserInput::Text {
text: "spawned".to_string(),
text_elements: Vec::new(),
}],
final_output_json_schema: None,
},

View File

@@ -1,6 +1,8 @@
pub(crate) mod control;
pub(crate) mod role;
pub(crate) mod status;
pub(crate) use codex_protocol::protocol::AgentStatus;
pub(crate) use control::AgentControl;
pub(crate) use role::AgentRole;
pub(crate) use status::agent_status_from_event;

View File

@@ -0,0 +1,85 @@
use crate::config::Config;
use crate::protocol::SandboxPolicy;
use serde::Deserialize;
use serde::Serialize;
/// Base instructions for the orchestrator role.
const ORCHESTRATOR_PROMPT: &str = include_str!("../../templates/agents/orchestrator.md");
/// Base instructions for the worker role.
const WORKER_PROMPT: &str = include_str!("../../gpt-5.2-codex_prompt.md");
/// Default worker model override used by the worker role.
const WORKER_MODEL: &str = "gpt-5.2-codex";
/// Enumerated list of all supported agent roles.
const ALL_ROLES: [AgentRole; 3] = [
AgentRole::Default,
AgentRole::Orchestrator,
AgentRole::Worker,
];
/// Hard-coded agent role selection used when spawning sub-agents.
#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
#[serde(rename_all = "snake_case")]
pub enum AgentRole {
/// Inherit the parent agent's configuration unchanged.
Default,
/// Coordination-only agent that delegates to workers.
Orchestrator,
/// Task-executing agent with a fixed model override.
Worker,
}
/// Immutable profile data that drives per-agent configuration overrides.
#[derive(Debug, Clone, Copy, PartialEq, Eq, Default)]
pub struct AgentProfile {
/// Optional base instructions override.
pub base_instructions: Option<&'static str>,
/// Optional model override.
pub model: Option<&'static str>,
/// Whether to force a read-only sandbox policy.
pub read_only: bool,
}
impl AgentRole {
/// Returns the string values used by JSON schema enums.
pub fn enum_values() -> Vec<String> {
ALL_ROLES
.iter()
.filter_map(|role| serde_json::to_string(role).ok())
.collect()
}
/// Returns the hard-coded profile for this role.
pub fn profile(self) -> AgentProfile {
match self {
AgentRole::Default => AgentProfile::default(),
AgentRole::Orchestrator => AgentProfile {
base_instructions: Some(ORCHESTRATOR_PROMPT),
..Default::default()
},
AgentRole::Worker => AgentProfile {
base_instructions: Some(WORKER_PROMPT),
model: Some(WORKER_MODEL),
..Default::default()
},
}
}
/// Applies this role's profile onto the provided config.
pub fn apply_to_config(self, config: &mut Config) -> Result<(), String> {
let profile = self.profile();
if let Some(base_instructions) = profile.base_instructions {
config.base_instructions = Some(base_instructions.to_string());
}
if let Some(model) = profile.model {
config.model = Some(model.to_string());
}
if profile.read_only {
config
.sandbox_policy
.set(SandboxPolicy::new_read_only_policy())
.map_err(|err| format!("sandbox_policy is invalid: {err}"))?;
}
Ok(())
}
}

View File

@@ -28,6 +28,7 @@ pub(crate) fn map_api_error(err: ApiError) -> CodexErr {
url: None,
request_id: None,
}),
ApiError::InvalidRequest { message } => CodexErr::InvalidRequest(message),
ApiError::Transport(transport) => match transport {
TransportError::Http {
status,

View File

@@ -1,4 +1,5 @@
use std::sync::Arc;
use std::sync::OnceLock;
use crate::api_bridge::CoreAuthProvider;
use crate::api_bridge::auth_provider_from_auth;
@@ -31,6 +32,7 @@ use codex_otel::OtelManager;
use codex_protocol::ThreadId;
use codex_protocol::config_types::ReasoningSummary as ReasoningSummaryConfig;
use codex_protocol::config_types::WebSearchMode;
use codex_protocol::models::ResponseItem;
use codex_protocol::openai_models::ModelInfo;
use codex_protocol::openai_models::ReasoningEffort as ReasoningEffortConfig;
@@ -64,6 +66,9 @@ use crate::model_provider_info::WireApi;
use crate::tools::spec::create_tools_json_for_chat_completions_api;
use crate::tools::spec::create_tools_json_for_responses_api;
pub const WEB_SEARCH_ELIGIBLE_HEADER: &str = "x-oai-web-search-eligible";
pub const X_CODEX_TURN_STATE_HEADER: &str = "x-codex-turn-state";
#[derive(Debug)]
struct ModelClientState {
config: Arc<Config>,
@@ -86,6 +91,17 @@ pub struct ModelClientSession {
state: Arc<ModelClientState>,
connection: Option<ApiWebSocketConnection>,
websocket_last_items: Vec<ResponseItem>,
/// Turn state for sticky routing.
///
/// This is an `OnceLock` that stores the turn state value received from the server
/// on turn start via the `x-codex-turn-state` response header. Once set, this value
/// should be sent back to the server in the `x-codex-turn-state` request header for
/// all subsequent requests within the same turn to maintain sticky routing.
///
/// This is a contract between the client and server: we receive it at turn start,
/// keep sending it unchanged between turn requests (e.g., for retries, incremental
/// appends, or continuation requests), and must not send it between different turns.
turn_state: Arc<OnceLock<String>>,
}
#[allow(clippy::too_many_arguments)]
@@ -121,6 +137,7 @@ impl ModelClient {
state: Arc::clone(&self.state),
connection: None,
websocket_last_items: Vec::new(),
turn_state: Arc::new(OnceLock::new()),
}
}
}
@@ -200,9 +217,7 @@ impl ModelClient {
let client = ApiCompactClient::new(transport, api_provider, api_auth)
.with_telemetry(Some(request_telemetry));
let instructions = prompt
.get_full_instructions(&self.state.model_info)
.into_owned();
let instructions = prompt.base_instructions.text.clone();
let payload = ApiCompactionInput {
model: &self.state.model_info.slug,
input: &prompt.input,
@@ -223,7 +238,6 @@ impl ModelClient {
extra_headers.insert("x-openai-subagent", val);
}
}
client
.compact_input(&payload, extra_headers)
.await
@@ -260,8 +274,7 @@ impl ModelClientSession {
}
fn build_responses_request(&self, prompt: &Prompt) -> Result<ApiPrompt> {
let model_info = self.state.model_info.clone();
let instructions = prompt.get_full_instructions(&model_info).into_owned();
let instructions = prompt.base_instructions.text.clone();
let tools_json: Vec<Value> = create_tools_json_for_responses_api(&prompt.tools)?;
Ok(build_api_prompt(prompt, instructions, tools_json))
}
@@ -319,8 +332,9 @@ impl ModelClientSession {
store_override: None,
conversation_id: Some(conversation_id),
session_source: Some(self.state.session_source.clone()),
extra_headers: beta_feature_headers(&self.state.config),
extra_headers: build_responses_headers(&self.state.config, Some(&self.turn_state)),
compression,
turn_state: Some(Arc::clone(&self.turn_state)),
}
}
@@ -394,7 +408,7 @@ impl ModelClientSession {
headers.extend(build_conversation_headers(options.conversation_id.clone()));
let new_conn: ApiWebSocketConnection =
ApiWebSocketResponsesClient::new(api_provider, api_auth)
.connect(headers)
.connect(headers, options.turn_state.clone())
.await?;
self.connection = Some(new_conn);
}
@@ -431,8 +445,7 @@ impl ModelClientSession {
}
let auth_manager = self.state.auth_manager.clone();
let model_info = self.state.model_info.clone();
let instructions = prompt.get_full_instructions(&model_info).into_owned();
let instructions = prompt.base_instructions.text.clone();
let tools_json = create_tools_json_for_chat_completions_api(&prompt.tools)?;
let api_prompt = build_api_prompt(prompt, instructions, tools_json);
let conversation_id = self.state.conversation_id.to_string();
@@ -635,6 +648,30 @@ fn beta_feature_headers(config: &Config) -> ApiHeaderMap {
headers
}
fn build_responses_headers(
config: &Config,
turn_state: Option<&Arc<OnceLock<String>>>,
) -> ApiHeaderMap {
let mut headers = beta_feature_headers(config);
headers.insert(
WEB_SEARCH_ELIGIBLE_HEADER,
HeaderValue::from_static(
if matches!(config.web_search_mode, Some(WebSearchMode::Disabled)) {
"false"
} else {
"true"
},
),
);
if let Some(turn_state) = turn_state
&& let Some(state) = turn_state.get()
&& let Ok(header_value) = HeaderValue::from_str(state)
{
headers.insert(X_CODEX_TURN_STATE_HEADER, header_value);
}
headers
}
fn map_response_stream<S>(api_stream: S, otel_manager: OtelManager) -> ResponseStream
where
S: futures::Stream<Item = std::result::Result<ResponseEvent, ApiError>>

View File

@@ -1,12 +1,12 @@
use crate::client_common::tools::ToolSpec;
use crate::config::types::Personality;
use crate::error::Result;
pub use codex_api::common::ResponseEvent;
use codex_protocol::models::BaseInstructions;
use codex_protocol::models::ResponseItem;
use codex_protocol::openai_models::ModelInfo;
use futures::Stream;
use serde::Deserialize;
use serde_json::Value;
use std::borrow::Cow;
use std::collections::HashSet;
use std::pin::Pin;
use std::task::Context;
@@ -34,22 +34,16 @@ pub struct Prompt {
/// Whether parallel tool calls are permitted for this prompt.
pub(crate) parallel_tool_calls: bool,
/// Optional override for the built-in BASE_INSTRUCTIONS.
pub base_instructions_override: Option<String>,
pub base_instructions: BaseInstructions,
/// Optionally specify the personality of the model.
pub personality: Option<Personality>,
/// Optional the output schema for the model's response.
pub output_schema: Option<Value>,
}
impl Prompt {
pub(crate) fn get_full_instructions<'a>(&'a self, model: &'a ModelInfo) -> Cow<'a, str> {
Cow::Borrowed(
self.base_instructions_override
.as_deref()
.unwrap_or(model.base_instructions.as_str()),
)
}
pub(crate) fn get_formatted_input(&self) -> Vec<ResponseItem> {
let mut input = self.input.clone();
@@ -245,76 +239,8 @@ mod tests {
use codex_api::create_text_param_for_request;
use pretty_assertions::assert_eq;
use crate::config::test_config;
use crate::models_manager::manager::ModelsManager;
use super::*;
struct InstructionsTestCase {
pub slug: &'static str,
pub expects_apply_patch_instructions: bool,
}
#[test]
fn get_full_instructions_no_user_content() {
let prompt = Prompt {
..Default::default()
};
let prompt_with_apply_patch_instructions =
include_str!("../prompt_with_apply_patch_instructions.md");
let test_cases = vec![
InstructionsTestCase {
slug: "gpt-3.5",
expects_apply_patch_instructions: true,
},
InstructionsTestCase {
slug: "gpt-4.1",
expects_apply_patch_instructions: true,
},
InstructionsTestCase {
slug: "gpt-4o",
expects_apply_patch_instructions: true,
},
InstructionsTestCase {
slug: "gpt-5",
expects_apply_patch_instructions: true,
},
InstructionsTestCase {
slug: "gpt-5.1",
expects_apply_patch_instructions: false,
},
InstructionsTestCase {
slug: "codex-mini-latest",
expects_apply_patch_instructions: true,
},
InstructionsTestCase {
slug: "gpt-oss:120b",
expects_apply_patch_instructions: false,
},
InstructionsTestCase {
slug: "gpt-5.1-codex",
expects_apply_patch_instructions: false,
},
InstructionsTestCase {
slug: "gpt-5.1-codex-max",
expects_apply_patch_instructions: false,
},
];
for test_case in test_cases {
let config = test_config();
let model_info = ModelsManager::construct_model_info_offline(test_case.slug, &config);
if test_case.expects_apply_patch_instructions {
assert_eq!(
model_info.base_instructions.as_str(),
prompt_with_apply_patch_instructions
);
}
let expected = model_info.base_instructions.as_str();
let full = prompt.get_full_instructions(&model_info);
assert_eq!(full, expected);
}
}
#[test]
fn serializes_text_verbosity_when_set() {
let input: Vec<ResponseItem> = vec![];

File diff suppressed because it is too large Load Diff

View File

@@ -1,3 +1,4 @@
use std::collections::HashMap;
use std::sync::Arc;
use std::sync::atomic::AtomicU64;
@@ -9,9 +10,12 @@ use codex_protocol::protocol::Event;
use codex_protocol::protocol::EventMsg;
use codex_protocol::protocol::ExecApprovalRequestEvent;
use codex_protocol::protocol::Op;
use codex_protocol::protocol::RequestUserInputEvent;
use codex_protocol::protocol::SessionSource;
use codex_protocol::protocol::SubAgentSource;
use codex_protocol::protocol::Submission;
use codex_protocol::request_user_input::RequestUserInputArgs;
use codex_protocol::request_user_input::RequestUserInputResponse;
use codex_protocol::user_input::UserInput;
use std::time::Duration;
use tokio::time::timeout;
@@ -229,6 +233,20 @@ async fn forward_events(
)
.await;
}
Event {
id,
msg: EventMsg::RequestUserInput(event),
} => {
handle_request_user_input(
&codex,
id,
&parent_session,
&parent_ctx,
event,
&cancel_token,
)
.await;
}
other => {
match tx_sub.send(other).or_cancel(&cancel_token).await {
Ok(Ok(())) => {}
@@ -334,6 +352,55 @@ async fn handle_patch_approval(
let _ = codex.submit(Op::PatchApproval { id, decision }).await;
}
async fn handle_request_user_input(
codex: &Codex,
id: String,
parent_session: &Session,
parent_ctx: &TurnContext,
event: RequestUserInputEvent,
cancel_token: &CancellationToken,
) {
let args = RequestUserInputArgs {
questions: event.questions,
};
let response_fut =
parent_session.request_user_input(parent_ctx, parent_ctx.sub_id.clone(), args);
let response = await_user_input_with_cancel(
response_fut,
parent_session,
&parent_ctx.sub_id,
cancel_token,
)
.await;
let _ = codex.submit(Op::UserInputAnswer { id, response }).await;
}
async fn await_user_input_with_cancel<F>(
fut: F,
parent_session: &Session,
sub_id: &str,
cancel_token: &CancellationToken,
) -> RequestUserInputResponse
where
F: core::future::Future<Output = Option<RequestUserInputResponse>>,
{
tokio::select! {
biased;
_ = cancel_token.cancelled() => {
let empty = RequestUserInputResponse {
answers: HashMap::new(),
};
parent_session
.notify_user_input_response(sub_id, empty.clone())
.await;
empty
}
response = fut => response.unwrap_or_else(|| RequestUserInputResponse {
answers: HashMap::new(),
}),
}
}
/// Await an approval decision, aborting on cancellation.
async fn await_approval_with_cancel<F>(
fut: F,

View File

@@ -82,6 +82,11 @@ fn is_dangerous_powershell(command: &[String]) -> bool {
}
}
// Check for force delete operations (e.g., Remove-Item -Force)
if has_force_delete_cmdlet(&tokens_lc) {
return true;
}
false
}
@@ -107,15 +112,49 @@ fn is_dangerous_cmd(command: &[String]) -> bool {
}
}
let Some(first_cmd) = iter.next() else {
return false;
};
// Classic `cmd /c start https://...` ShellExecute path.
if !first_cmd.eq_ignore_ascii_case("start") {
let remaining: Vec<String> = iter.cloned().collect();
if remaining.is_empty() {
return false;
}
let remaining: Vec<String> = iter.cloned().collect();
args_have_url(&remaining)
let cmd_tokens: Vec<String> = match remaining.as_slice() {
[only] => shlex_split(only).unwrap_or_else(|| vec![only.clone()]),
_ => remaining,
};
// Refine tokens by splitting concatenated CMD operators (e.g. "echo hi&del")
let tokens: Vec<String> = cmd_tokens
.into_iter()
.flat_map(|t| split_embedded_cmd_operators(&t))
.collect();
const CMD_SEPARATORS: &[&str] = &["&", "&&", "|", "||"];
tokens
.split(|t| CMD_SEPARATORS.contains(&t.as_str()))
.any(|segment| {
let Some(cmd) = segment.first() else {
return false;
};
// Classic `cmd /c ... start https://...` ShellExecute path.
if cmd.eq_ignore_ascii_case("start") && args_have_url(segment) {
return true;
}
// Force delete: del /f, erase /f
if (cmd.eq_ignore_ascii_case("del") || cmd.eq_ignore_ascii_case("erase"))
&& has_force_flag_cmd(segment)
{
return true;
}
// Recursive directory removal: rd /s /q, rmdir /s /q
if (cmd.eq_ignore_ascii_case("rd") || cmd.eq_ignore_ascii_case("rmdir"))
&& has_recursive_flag_cmd(segment)
&& has_quiet_flag_cmd(segment)
{
return true;
}
false
})
}
fn is_direct_gui_launch(command: &[String]) -> bool {
@@ -149,6 +188,123 @@ fn is_direct_gui_launch(command: &[String]) -> bool {
false
}
fn split_embedded_cmd_operators(token: &str) -> Vec<String> {
// Split concatenated CMD operators so `echo hi&del` becomes `["echo hi", "&", "del"]`.
// Handles `&`, `&&`, `|`, `||`. Best-effort (CMD escaping is weird by nature).
let mut parts = Vec::new();
let mut start = 0;
let mut it = token.char_indices().peekable();
while let Some((i, ch)) = it.next() {
if ch == '&' || ch == '|' {
if i > start {
parts.push(token[start..i].to_string());
}
// Detect doubled operator: && or ||
let op_len = match it.peek() {
Some(&(j, next)) if next == ch => {
it.next(); // consume second char
(j + next.len_utf8()) - i
}
_ => ch.len_utf8(),
};
parts.push(token[i..i + op_len].to_string());
start = i + op_len;
}
}
if start < token.len() {
parts.push(token[start..].to_string());
}
parts.retain(|s| !s.trim().is_empty());
parts
}
fn has_force_delete_cmdlet(tokens: &[String]) -> bool {
const DELETE_CMDLETS: &[&str] = &["remove-item", "ri", "rm", "del", "erase", "rd", "rmdir"];
// Hard separators that end a command segment (so -Force must be in same segment)
const SEG_SEPS: &[char] = &[';', '|', '&', '\n', '\r', '\t'];
// Soft separators: punctuation that can stick to tokens (blocks, parens, brackets, commas, etc.)
const SOFT_SEPS: &[char] = &['{', '}', '(', ')', '[', ']', ',', ';'];
// Build rough command segments first
let mut segments: Vec<Vec<String>> = vec![Vec::new()];
for tok in tokens {
// If token itself contains segment separators, split it (best-effort)
let mut cur = String::new();
for ch in tok.chars() {
if SEG_SEPS.contains(&ch) {
let s = cur.trim();
if let Some(msg) = segments.last_mut()
&& !s.is_empty()
{
msg.push(s.to_string());
}
cur.clear();
if let Some(last) = segments.last()
&& !last.is_empty()
{
segments.push(Vec::new());
}
} else {
cur.push(ch);
}
}
let s = cur.trim();
if let Some(segment) = segments.last_mut()
&& !s.is_empty()
{
segment.push(s.to_string());
}
}
// Now, inside each segment, normalize tokens by splitting on soft punctuation
segments.into_iter().any(|seg| {
let atoms = seg
.iter()
.flat_map(|t| t.split(|c| SOFT_SEPS.contains(&c)))
.map(str::trim)
.filter(|s| !s.is_empty());
let mut has_delete = false;
let mut has_force = false;
for a in atoms {
if DELETE_CMDLETS.iter().any(|cmd| a.eq_ignore_ascii_case(cmd)) {
has_delete = true;
}
if a.eq_ignore_ascii_case("-force")
|| a.get(..7)
.is_some_and(|p| p.eq_ignore_ascii_case("-force:"))
{
has_force = true;
}
}
has_delete && has_force
})
}
/// Check for /f or /F flag in CMD del/erase arguments.
fn has_force_flag_cmd(args: &[String]) -> bool {
args.iter().any(|a| a.eq_ignore_ascii_case("/f"))
}
/// Check for /s or /S flag in CMD rd/rmdir arguments.
fn has_recursive_flag_cmd(args: &[String]) -> bool {
args.iter().any(|a| a.eq_ignore_ascii_case("/s"))
}
/// Check for /q or /Q flag in CMD rd/rmdir arguments.
fn has_quiet_flag_cmd(args: &[String]) -> bool {
args.iter().any(|a| a.eq_ignore_ascii_case("/q"))
}
fn args_have_url(args: &[String]) -> bool {
args.iter().any(|arg| looks_like_url(arg))
}
@@ -313,4 +469,287 @@ mod tests {
"."
])));
}
// Force delete tests for PowerShell
#[test]
fn powershell_remove_item_force_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"Remove-Item test -Force"
])));
}
#[test]
fn powershell_remove_item_recurse_force_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"Remove-Item test -Recurse -Force"
])));
}
#[test]
fn powershell_ri_alias_force_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"pwsh",
"-Command",
"ri test -Force"
])));
}
#[test]
fn powershell_remove_item_without_force_is_not_flagged() {
assert!(!is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"Remove-Item test"
])));
}
// Force delete tests for CMD
#[test]
fn cmd_del_force_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd", "/c", "del", "/f", "test.txt"
])));
}
#[test]
fn cmd_erase_force_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd", "/c", "erase", "/f", "test.txt"
])));
}
#[test]
fn cmd_del_without_force_is_not_flagged() {
assert!(!is_dangerous_command_windows(&vec_str(&[
"cmd", "/c", "del", "test.txt"
])));
}
#[test]
fn cmd_rd_recursive_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd", "/c", "rd", "/s", "/q", "test"
])));
}
#[test]
fn cmd_rd_without_quiet_is_not_flagged() {
assert!(!is_dangerous_command_windows(&vec_str(&[
"cmd", "/c", "rd", "/s", "test"
])));
}
#[test]
fn cmd_rmdir_recursive_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd", "/c", "rmdir", "/s", "/q", "test"
])));
}
// Test exact scenario from issue #8567
#[test]
fn powershell_remove_item_path_recurse_force_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"Remove-Item -Path 'test' -Recurse -Force"
])));
}
#[test]
fn powershell_remove_item_force_with_semicolon_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"Remove-Item test -Force; Write-Host done"
])));
}
#[test]
fn powershell_remove_item_force_inside_block_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"if ($true) { Remove-Item test -Force}"
])));
}
#[test]
fn powershell_remove_item_force_inside_brackets_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"[void]( Remove-Item test -Force)]"
])));
}
#[test]
fn cmd_del_path_containing_f_is_not_flagged() {
assert!(!is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
"del",
"C:/foo/bar.txt"
])));
}
#[test]
fn cmd_rd_path_containing_s_is_not_flagged() {
assert!(!is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
"rd",
"C:/source"
])));
}
#[test]
fn cmd_bypass_chained_del_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd", "/c", "echo", "hello", "&", "del", "/f", "file.txt"
])));
}
#[test]
fn powershell_chained_no_space_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"Write-Host hi;Remove-Item -Force C:\\tmp"
])));
}
#[test]
fn powershell_comma_separated_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"del,-Force,C:\\foo"
])));
}
#[test]
fn cmd_echo_del_is_not_dangerous() {
assert!(!is_dangerous_command_windows(&vec_str(&[
"cmd", "/c", "echo", "del", "/f"
])));
}
#[test]
fn cmd_del_single_string_argument_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
"del /f file.txt"
])));
}
#[test]
fn cmd_del_chained_single_string_argument_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
"echo hello & del /f file.txt"
])));
}
#[test]
fn cmd_chained_no_space_del_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
"echo hi&del /f file.txt"
])));
}
#[test]
fn cmd_chained_andand_no_space_del_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
"echo hi&&del /f file.txt"
])));
}
#[test]
fn cmd_chained_oror_no_space_del_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
"echo hi||del /f file.txt"
])));
}
#[test]
fn cmd_start_url_single_string_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
"start https://example.com"
])));
}
#[test]
fn cmd_chained_no_space_rmdir_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
"echo hi&rmdir /s /q testdir"
])));
}
#[test]
fn cmd_del_force_uppercase_flag_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd", "/c", "DEL", "/F", "file.txt"
])));
}
#[test]
fn cmdexe_r_del_force_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd.exe", "/r", "del", "/f", "file.txt"
])));
}
#[test]
fn cmd_start_quoted_url_single_string_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
r#"start "https://example.com""#
])));
}
#[test]
fn cmd_start_title_then_url_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"cmd",
"/c",
r#"start "" https://example.com"#
])));
}
#[test]
fn powershell_rm_alias_force_is_dangerous() {
assert!(is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"rm test -Force"
])));
}
#[test]
fn powershell_benign_force_separate_command_is_not_dangerous() {
assert!(!is_dangerous_command_windows(&vec_str(&[
"powershell",
"-Command",
"Get-ChildItem -Force; Remove-Item test"
])));
}
}

View File

@@ -15,6 +15,7 @@ use crate::protocol::EventMsg;
use crate::protocol::TurnContextItem;
use crate::protocol::TurnStartedEvent;
use crate::protocol::WarningEvent;
use crate::session_prefix::TURN_ABORTED_OPEN_TAG;
use crate::truncate::TruncationPolicy;
use crate::truncate::approx_token_count;
use crate::truncate::truncate_text;
@@ -44,7 +45,11 @@ pub(crate) async fn run_inline_auto_compact_task(
turn_context: Arc<TurnContext>,
) {
let prompt = turn_context.compact_prompt().to_string();
let input = vec![UserInput::Text { text: prompt }];
let input = vec![UserInput::Text {
text: prompt,
// Compaction prompt is synthesized; no UI element ranges to preserve.
text_elements: Vec::new(),
}];
run_compact_task_inner(sess, turn_context, input).await;
}
@@ -86,7 +91,6 @@ async fn run_compact_task_inner(
model: turn_context.client.get_model(),
effort: turn_context.client.get_reasoning_effort(),
summary: turn_context.client.get_reasoning_summary(),
base_instructions: turn_context.base_instructions.clone(),
user_instructions: turn_context.user_instructions.clone(),
developer_instructions: turn_context.developer_instructions.clone(),
final_output_json_schema: turn_context.final_output_json_schema.clone(),
@@ -100,6 +104,7 @@ async fn run_compact_task_inner(
let turn_input_len = turn_input.len();
let prompt = Prompt {
input: turn_input,
base_instructions: sess.get_base_instructions().await,
..Default::default()
};
let attempt_result = drain_to_completed(&sess, turn_context.as_ref(), &prompt).await;
@@ -163,7 +168,7 @@ async fn run_compact_task_inner(
let summary_text = format!("{SUMMARY_PREFIX}\n{summary_suffix}");
let user_messages = collect_user_messages(history_items);
let initial_context = sess.build_initial_context(turn_context.as_ref());
let initial_context = sess.build_initial_context(turn_context.as_ref()).await;
let mut new_history = build_compacted_history(initial_context, &user_messages, &summary_text);
let ghost_snapshots: Vec<ResponseItem> = history_items
.iter()
@@ -219,11 +224,31 @@ pub(crate) fn collect_user_messages(items: &[ResponseItem]) -> Vec<String> {
Some(user.message())
}
}
_ => None,
_ => collect_turn_aborted_marker(item),
})
.collect()
}
fn collect_turn_aborted_marker(item: &ResponseItem) -> Option<String> {
let ResponseItem::Message { role, content, .. } = item else {
return None;
};
if role != "user" {
return None;
}
let text = content_items_to_text(content)?;
if text
.trim_start()
.to_ascii_lowercase()
.starts_with(TURN_ABORTED_OPEN_TAG)
{
Some(text)
} else {
None
}
}
pub(crate) fn is_summary_message(message: &str) -> bool {
message.starts_with(format!("{SUMMARY_PREFIX}\n").as_str())
}
@@ -312,6 +337,9 @@ async fn drain_to_completed(
sess.record_into_history(std::slice::from_ref(&item), turn_context)
.await;
}
Ok(ResponseEvent::ServerReasoningIncluded(included)) => {
sess.set_server_reasoning_included(included).await;
}
Ok(ResponseEvent::RateLimits(snapshot)) => {
sess.update_rate_limits(turn_context, snapshot).await;
}
@@ -330,6 +358,7 @@ async fn drain_to_completed(
mod tests {
use super::*;
use crate::session_prefix::TURN_ABORTED_OPEN_TAG;
use pretty_assertions::assert_eq;
#[test]
@@ -482,4 +511,41 @@ mod tests {
};
assert_eq!(summary, summary_text);
}
#[test]
fn build_compacted_history_preserves_turn_aborted_markers() {
let marker = format!(
"{TURN_ABORTED_OPEN_TAG}\n <turn_id>turn-1</turn_id>\n <reason>interrupted</reason>\n</turn_aborted>"
);
let items = vec![
ResponseItem::Message {
id: None,
role: "user".to_string(),
content: vec![ContentItem::InputText {
text: marker.clone(),
}],
},
ResponseItem::Message {
id: None,
role: "user".to_string(),
content: vec![ContentItem::InputText {
text: "real user message".to_string(),
}],
},
];
let user_messages = collect_user_messages(&items);
let history = build_compacted_history(Vec::new(), &user_messages, "SUMMARY");
let found_marker = history.iter().any(|item| match item {
ResponseItem::Message { role, content, .. } if role == "user" => {
content_items_to_text(content).is_some_and(|text| text == marker)
}
_ => false,
});
assert!(
found_marker,
"expected compacted history to retain <turn_aborted> marker"
);
}
}

View File

@@ -54,7 +54,8 @@ async fn run_remote_compact_task_inner_impl(
input: history.for_prompt(),
tools: vec![],
parallel_tool_calls: false,
base_instructions_override: turn_context.base_instructions.clone(),
base_instructions: sess.get_base_instructions().await,
personality: None,
output_schema: None,
};

View File

@@ -1,14 +1,16 @@
use crate::config::CONFIG_TOML_FILE;
use crate::config::types::McpServerConfig;
use crate::config::types::Notice;
use crate::path_utils::resolve_symlink_write_paths;
use crate::path_utils::write_atomically;
use anyhow::Context;
use codex_protocol::config_types::TrustLevel;
use codex_protocol::openai_models::ReasoningEffort;
use std::collections::BTreeMap;
use std::path::Path;
use std::path::PathBuf;
use tempfile::NamedTempFile;
use tokio::task;
use toml_edit::ArrayOfTables;
use toml_edit::DocumentMut;
use toml_edit::Item as TomlItem;
use toml_edit::Table as TomlTable;
@@ -36,6 +38,8 @@ pub enum ConfigEdit {
RecordModelMigrationSeen { from: String, to: String },
/// Replace the entire `[mcp_servers]` table.
ReplaceMcpServers(BTreeMap<String, McpServerConfig>),
/// Set or clear a skill config entry under `[[skills.config]]`.
SetSkillConfig { path: PathBuf, enabled: bool },
/// Set trust_level under `[projects."<path>"]`,
/// migrating inline tables to explicit tables.
SetProjectTrustLevel { path: PathBuf, level: TrustLevel },
@@ -298,6 +302,9 @@ impl ConfigDocument {
value(*acknowledged),
)),
ConfigEdit::ReplaceMcpServers(servers) => Ok(self.replace_mcp_servers(servers)),
ConfigEdit::SetSkillConfig { path, enabled } => {
Ok(self.set_skill_config(path.as_path(), *enabled))
}
ConfigEdit::SetPath { segments, value } => Ok(self.insert(segments, value.clone())),
ConfigEdit::ClearPath { segments } => Ok(self.clear_owned(segments)),
ConfigEdit::SetProjectTrustLevel { path, level } => {
@@ -387,6 +394,113 @@ impl ConfigDocument {
true
}
fn set_skill_config(&mut self, path: &Path, enabled: bool) -> bool {
let normalized_path = normalize_skill_config_path(path);
let mut remove_skills_table = false;
let mut mutated = false;
{
let root = self.doc.as_table_mut();
let skills_item = match root.get_mut("skills") {
Some(item) => item,
None => {
if enabled {
return false;
}
root.insert(
"skills",
TomlItem::Table(document_helpers::new_implicit_table()),
);
let Some(item) = root.get_mut("skills") else {
return false;
};
item
}
};
if document_helpers::ensure_table_for_write(skills_item).is_none() {
if enabled {
return false;
}
*skills_item = TomlItem::Table(document_helpers::new_implicit_table());
}
let Some(skills_table) = skills_item.as_table_mut() else {
return false;
};
let config_item = match skills_table.get_mut("config") {
Some(item) => item,
None => {
if enabled {
return false;
}
skills_table.insert("config", TomlItem::ArrayOfTables(ArrayOfTables::new()));
let Some(item) = skills_table.get_mut("config") else {
return false;
};
item
}
};
if !matches!(config_item, TomlItem::ArrayOfTables(_)) {
if enabled {
return false;
}
*config_item = TomlItem::ArrayOfTables(ArrayOfTables::new());
}
let TomlItem::ArrayOfTables(overrides) = config_item else {
return false;
};
let existing_index = overrides.iter().enumerate().find_map(|(idx, table)| {
table
.get("path")
.and_then(|item| item.as_str())
.map(Path::new)
.map(normalize_skill_config_path)
.filter(|value| *value == normalized_path)
.map(|_| idx)
});
if enabled {
if let Some(index) = existing_index {
overrides.remove(index);
mutated = true;
if overrides.is_empty() {
skills_table.remove("config");
if skills_table.is_empty() {
remove_skills_table = true;
}
}
}
} else if let Some(index) = existing_index {
for (idx, table) in overrides.iter_mut().enumerate() {
if idx == index {
table["path"] = value(normalized_path);
table["enabled"] = value(false);
mutated = true;
break;
}
}
} else {
let mut entry = TomlTable::new();
entry.set_implicit(false);
entry["path"] = value(normalized_path);
entry["enabled"] = value(false);
overrides.push(entry);
mutated = true;
}
}
if remove_skills_table {
let root = self.doc.as_table_mut();
root.remove("skills");
}
mutated
}
fn scoped_segments(&self, scope: Scope, segments: &[&str]) -> Vec<String> {
let resolved: Vec<String> = segments
.iter()
@@ -494,6 +608,13 @@ impl ConfigDocument {
}
}
fn normalize_skill_config_path(path: &Path) -> String {
dunce::canonicalize(path)
.unwrap_or_else(|_| path.to_path_buf())
.to_string_lossy()
.to_string()
}
/// Persist edits using a blocking strategy.
pub fn apply_blocking(
codex_home: &Path,
@@ -505,10 +626,14 @@ pub fn apply_blocking(
}
let config_path = codex_home.join(CONFIG_TOML_FILE);
let serialized = match std::fs::read_to_string(&config_path) {
Ok(contents) => contents,
Err(err) if err.kind() == std::io::ErrorKind::NotFound => String::new(),
Err(err) => return Err(err.into()),
let write_paths = resolve_symlink_write_paths(&config_path)?;
let serialized = match write_paths.read_path {
Some(path) => match std::fs::read_to_string(&path) {
Ok(contents) => contents,
Err(err) if err.kind() == std::io::ErrorKind::NotFound => String::new(),
Err(err) => return Err(err.into()),
},
None => String::new(),
};
let doc = if serialized.is_empty() {
@@ -534,22 +659,13 @@ pub fn apply_blocking(
return Ok(());
}
std::fs::create_dir_all(codex_home).with_context(|| {
write_atomically(&write_paths.write_path, &document.doc.to_string()).with_context(|| {
format!(
"failed to create Codex home directory at {}",
codex_home.display()
"failed to persist config.toml at {}",
write_paths.write_path.display()
)
})?;
let tmp = NamedTempFile::new_in(codex_home)?;
std::fs::write(tmp.path(), document.doc.to_string()).with_context(|| {
format!(
"failed to write temporary config file at {}",
tmp.path().display()
)
})?;
tmp.persist(config_path)?;
Ok(())
}
@@ -693,6 +809,8 @@ mod tests {
use crate::config::types::McpServerTransportConfig;
use codex_protocol::openai_models::ReasoningEffort;
use pretty_assertions::assert_eq;
#[cfg(unix)]
use std::os::unix::fs::symlink;
use tempfile::tempdir;
use toml::Value as TomlValue;
@@ -737,6 +855,54 @@ model_reasoning_effort = "high"
assert_eq!(contents, "enabled = true\n");
}
#[test]
fn set_skill_config_writes_disabled_entry() {
let tmp = tempdir().expect("tmpdir");
let codex_home = tmp.path();
ConfigEditsBuilder::new(codex_home)
.with_edits([ConfigEdit::SetSkillConfig {
path: PathBuf::from("/tmp/skills/demo/SKILL.md"),
enabled: false,
}])
.apply_blocking()
.expect("persist");
let contents =
std::fs::read_to_string(codex_home.join(CONFIG_TOML_FILE)).expect("read config");
let expected = r#"[[skills.config]]
path = "/tmp/skills/demo/SKILL.md"
enabled = false
"#;
assert_eq!(contents, expected);
}
#[test]
fn set_skill_config_removes_entry_when_enabled() {
let tmp = tempdir().expect("tmpdir");
let codex_home = tmp.path();
std::fs::write(
codex_home.join(CONFIG_TOML_FILE),
r#"[[skills.config]]
path = "/tmp/skills/demo/SKILL.md"
enabled = false
"#,
)
.expect("seed config");
ConfigEditsBuilder::new(codex_home)
.with_edits([ConfigEdit::SetSkillConfig {
path: PathBuf::from("/tmp/skills/demo/SKILL.md"),
enabled: true,
}])
.apply_blocking()
.expect("persist");
let contents =
std::fs::read_to_string(codex_home.join(CONFIG_TOML_FILE)).expect("read config");
assert_eq!(contents, "");
}
#[test]
fn blocking_set_model_preserves_inline_table_contents() {
let tmp = tempdir().expect("tmpdir");
@@ -784,6 +950,71 @@ profiles = { fast = { model = "gpt-4o", sandbox_mode = "strict" } }
);
}
#[cfg(unix)]
#[test]
fn blocking_set_model_writes_through_symlink_chain() {
let tmp = tempdir().expect("tmpdir");
let codex_home = tmp.path();
let target_dir = tempdir().expect("target dir");
let target_path = target_dir.path().join(CONFIG_TOML_FILE);
let link_path = codex_home.join("config-link.toml");
let config_path = codex_home.join(CONFIG_TOML_FILE);
symlink(&target_path, &link_path).expect("symlink link");
symlink("config-link.toml", &config_path).expect("symlink config");
apply_blocking(
codex_home,
None,
&[ConfigEdit::SetModel {
model: Some("gpt-5.1-codex".to_string()),
effort: Some(ReasoningEffort::High),
}],
)
.expect("persist");
let meta = std::fs::symlink_metadata(&config_path).expect("config metadata");
assert!(meta.file_type().is_symlink());
let contents = std::fs::read_to_string(&target_path).expect("read target");
let expected = r#"model = "gpt-5.1-codex"
model_reasoning_effort = "high"
"#;
assert_eq!(contents, expected);
}
#[cfg(unix)]
#[test]
fn blocking_set_model_replaces_symlink_on_cycle() {
let tmp = tempdir().expect("tmpdir");
let codex_home = tmp.path();
let link_a = codex_home.join("a.toml");
let link_b = codex_home.join("b.toml");
let config_path = codex_home.join(CONFIG_TOML_FILE);
symlink("b.toml", &link_a).expect("symlink a");
symlink("a.toml", &link_b).expect("symlink b");
symlink("a.toml", &config_path).expect("symlink config");
apply_blocking(
codex_home,
None,
&[ConfigEdit::SetModel {
model: Some("gpt-5.1-codex".to_string()),
effort: None,
}],
)
.expect("persist");
let meta = std::fs::symlink_metadata(&config_path).expect("config metadata");
assert!(!meta.file_type().is_symlink());
let contents = std::fs::read_to_string(&config_path).expect("read config");
let expected = r#"model = "gpt-5.1-codex"
"#;
assert_eq!(contents, expected);
}
#[test]
fn batch_write_table_upsert_preserves_inline_comments() {
let tmp = tempdir().expect("tmpdir");
@@ -1007,6 +1238,7 @@ hide_rate_limit_model_nudge = true
"#;
assert_eq!(contents, expected);
}
#[test]
fn blocking_set_hide_gpt5_1_migration_prompt_preserves_table() {
let tmp = tempdir().expect("tmpdir");
@@ -1123,6 +1355,7 @@ gpt-5 = "gpt-5.1"
cwd: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: Some(vec!["one".to_string(), "two".to_string()]),
@@ -1144,6 +1377,7 @@ gpt-5 = "gpt-5.1"
env_http_headers: None,
},
enabled: false,
disabled_reason: None,
startup_timeout_sec: Some(std::time::Duration::from_secs(5)),
tool_timeout_sec: None,
enabled_tools: None,
@@ -1208,6 +1442,7 @@ foo = { command = "cmd" }
cwd: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -1251,6 +1486,7 @@ foo = { command = "cmd" } # keep me
cwd: None,
},
enabled: false,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -1293,6 +1529,7 @@ foo = { command = "cmd", args = ["--flag"] } # keep me
cwd: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -1336,6 +1573,7 @@ foo = { command = "cmd" }
cwd: None,
},
enabled: false,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,

View File

@@ -1,17 +1,22 @@
use crate::auth::AuthCredentialsStoreMode;
use crate::config::edit::ConfigEdit;
use crate::config::edit::ConfigEditsBuilder;
use crate::config::types::DEFAULT_OTEL_ENVIRONMENT;
use crate::config::types::History;
use crate::config::types::McpServerConfig;
use crate::config::types::McpServerDisabledReason;
use crate::config::types::McpServerTransportConfig;
use crate::config::types::Notice;
use crate::config::types::Notifications;
use crate::config::types::OtelConfig;
use crate::config::types::OtelConfigToml;
use crate::config::types::OtelExporterKind;
use crate::config::types::Personality;
use crate::config::types::SandboxWorkspaceWrite;
use crate::config::types::ScrollInputMode;
use crate::config::types::ShellEnvironmentPolicy;
use crate::config::types::ShellEnvironmentPolicyToml;
use crate::config::types::SkillsConfig;
use crate::config::types::Tui;
use crate::config::types::UriBasedFileOpener;
use crate::config_loader::ConfigLayerStack;
@@ -19,6 +24,7 @@ use crate::config_loader::ConfigRequirements;
use crate::config_loader::LoaderOverrides;
use crate::config_loader::McpServerIdentity;
use crate::config_loader::McpServerRequirement;
use crate::config_loader::Sourced;
use crate::config_loader::load_config_layers_state;
use crate::features::Feature;
use crate::features::FeatureOverrides;
@@ -42,6 +48,7 @@ use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::config_types::SandboxMode;
use codex_protocol::config_types::TrustLevel;
use codex_protocol::config_types::Verbosity;
use codex_protocol::config_types::WebSearchMode;
use codex_protocol::openai_models::ReasoningEffort;
use codex_rmcp_client::OAuthCredentialsStoreMode;
use codex_utils_absolute_path::AbsolutePathBuf;
@@ -76,8 +83,6 @@ pub use constraint::ConstraintResult;
pub use service::ConfigService;
pub use service::ConfigServiceError;
const OPENAI_DEFAULT_REVIEW_MODEL: &str = "gpt-5.1-codex-max";
pub use codex_git::GhostSnapshotConfig;
/// Maximum number of bytes of the documentation that will be embedded. Larger
@@ -108,8 +113,8 @@ pub struct Config {
/// Optional override of model selection.
pub model: Option<String>,
/// Model used specifically for review sessions. Defaults to "gpt-5.1-codex-max".
pub review_model: String,
/// Model used specifically for review sessions.
pub review_model: Option<String>,
/// Size of the context window for the model, in tokens.
pub model_context_window: Option<i64>,
@@ -123,6 +128,9 @@ pub struct Config {
/// Info needed to make an API request to the model.
pub model_provider: ModelProviderInfo,
/// Optionally specify the personality of the model
pub model_personality: Option<Personality>,
/// Approval policy for executing commands.
pub approval_policy: Constrained<AskForApproval>,
@@ -338,7 +346,8 @@ pub struct Config {
/// model info's default preference.
pub include_apply_patch_tool: bool,
pub tools_web_search_request: bool,
/// Explicit or feature-derived web search mode.
pub web_search_mode: Option<WebSearchMode>,
/// If set to `true`, used only the experimental unified exec tool.
pub use_experimental_unified_exec_tool: bool,
@@ -390,6 +399,7 @@ pub struct ConfigBuilder {
cli_overrides: Option<Vec<(String, TomlValue)>>,
harness_overrides: Option<ConfigOverrides>,
loader_overrides: Option<LoaderOverrides>,
fallback_cwd: Option<PathBuf>,
}
impl ConfigBuilder {
@@ -413,21 +423,29 @@ impl ConfigBuilder {
self
}
pub fn fallback_cwd(mut self, fallback_cwd: Option<PathBuf>) -> Self {
self.fallback_cwd = fallback_cwd;
self
}
pub async fn build(self) -> std::io::Result<Config> {
let Self {
codex_home,
cli_overrides,
harness_overrides,
loader_overrides,
fallback_cwd,
} = self;
let codex_home = codex_home.map_or_else(find_codex_home, std::io::Result::Ok)?;
let cli_overrides = cli_overrides.unwrap_or_default();
let harness_overrides = harness_overrides.unwrap_or_default();
let mut harness_overrides = harness_overrides.unwrap_or_default();
let loader_overrides = loader_overrides.unwrap_or_default();
let cwd = match harness_overrides.cwd.as_deref() {
let cwd_override = harness_overrides.cwd.as_deref().or(fallback_cwd.as_deref());
let cwd = match cwd_override {
Some(path) => AbsolutePathBuf::try_from(path)?,
None => AbsolutePathBuf::current_dir()?,
};
harness_overrides.cwd = Some(cwd.to_path_buf());
let config_layer_stack =
load_config_layers_state(&codex_home, Some(cwd), &cli_overrides, loader_overrides)
.await?;
@@ -460,6 +478,28 @@ impl Config {
.await
}
/// Load a default configuration when user config files are invalid.
pub fn load_default_with_cli_overrides(
cli_overrides: Vec<(String, TomlValue)>,
) -> std::io::Result<Self> {
let codex_home = find_codex_home()?;
let mut merged = toml::Value::try_from(ConfigToml::default()).map_err(|e| {
std::io::Error::new(
std::io::ErrorKind::InvalidData,
format!("failed to serialize default config: {e}"),
)
})?;
let cli_layer = crate::config_loader::build_cli_overrides_layer(&cli_overrides);
crate::config_loader::merge_toml_values(&mut merged, &cli_layer);
let config_toml = deserialize_config_toml_with_base(merged, &codex_home)?;
Self::load_config_with_layer_stack(
config_toml,
ConfigOverrides::default(),
codex_home,
ConfigLayerStack::default(),
)
}
/// This is a secondary way of creating [Config], which is appropriate when
/// the harness is meant to be used with a specific configuration that
/// ignores user settings. For example, the `codex exec` subcommand is
@@ -504,7 +544,7 @@ pub async fn load_config_as_toml_with_cli_overrides(
Ok(cfg)
}
fn deserialize_config_toml_with_base(
pub(crate) fn deserialize_config_toml_with_base(
root_value: TomlValue,
config_base_dir: &Path,
) -> std::io::Result<ConfigToml> {
@@ -518,25 +558,32 @@ fn deserialize_config_toml_with_base(
fn filter_mcp_servers_by_requirements(
mcp_servers: &mut HashMap<String, McpServerConfig>,
mcp_requirements: Option<&BTreeMap<String, McpServerRequirement>>,
mcp_requirements: Option<&Sourced<BTreeMap<String, McpServerRequirement>>>,
) {
let Some(allowlist) = mcp_requirements else {
return;
};
let source = allowlist.source.clone();
for (name, server) in mcp_servers.iter_mut() {
let allowed = allowlist
.value
.get(name)
.is_some_and(|requirement| mcp_server_matches_requirement(requirement, server));
if !allowed {
if allowed {
server.disabled_reason = None;
} else {
server.enabled = false;
server.disabled_reason = Some(McpServerDisabledReason::Requirements {
source: source.clone(),
});
}
}
}
fn constrain_mcp_servers(
mcp_servers: HashMap<String, McpServerConfig>,
mcp_requirements: Option<&BTreeMap<String, McpServerRequirement>>,
mcp_requirements: Option<&Sourced<BTreeMap<String, McpServerRequirement>>>,
) -> ConstraintResult<Constrained<HashMap<String, McpServerConfig>>> {
if mcp_requirements.is_none() {
return Ok(Constrained::allow_any(mcp_servers));
@@ -719,30 +766,17 @@ pub fn set_default_oss_provider(codex_home: &Path, provider: &str) -> std::io::R
));
}
}
let config_path = codex_home.join(CONFIG_TOML_FILE);
// Read existing config or create empty string if file doesn't exist
let content = match std::fs::read_to_string(&config_path) {
Ok(content) => content,
Err(e) if e.kind() == std::io::ErrorKind::NotFound => String::new(),
Err(e) => return Err(e),
};
// Parse as DocumentMut for editing while preserving structure
let mut doc = content.parse::<DocumentMut>().map_err(|e| {
std::io::Error::new(
std::io::ErrorKind::InvalidData,
format!("failed to parse config.toml: {e}"),
)
})?;
// Set the default_oss_provider at root level
use toml_edit::value;
doc["oss_provider"] = value(provider);
// Write the modified document back
std::fs::write(&config_path, doc.to_string())?;
Ok(())
let edits = [ConfigEdit::SetPath {
segments: vec!["oss_provider".to_string()],
value: value(provider),
}];
ConfigEditsBuilder::new(codex_home)
.with_edits(edits)
.apply_blocking()
.map_err(|err| std::io::Error::other(format!("failed to persist config.toml: {err}")))
}
/// Base config deserialized from ~/.codex/config.toml.
@@ -786,6 +820,12 @@ pub struct ConfigToml {
#[serde(default)]
pub developer_instructions: Option<String>,
/// Optional path to a file containing model instructions that will override
/// the built-in instructions for the selected model. Users are STRONGLY
/// DISCOURAGED from using this field, as deviating from the instructions
/// sanctioned by Codex will likely degrade model performance.
pub model_instructions_file: Option<AbsolutePathBuf>,
/// Compact prompt used for history compaction.
pub compact_prompt: Option<String>,
@@ -869,14 +909,24 @@ pub struct ConfigToml {
/// Override to force-enable reasoning summaries for the configured model.
pub model_supports_reasoning_summaries: Option<bool>,
/// EXPERIMENTAL
/// Optionally specify a personality for the model
pub model_personality: Option<Personality>,
/// Base URL for requests to ChatGPT (as opposed to the OpenAI API).
pub chatgpt_base_url: Option<String>,
pub projects: Option<HashMap<String, ProjectConfig>>,
/// Controls the web search tool mode: disabled, cached, or live.
pub web_search: Option<WebSearchMode>,
/// Nested tools section for feature toggles
pub tools: Option<ToolsToml>,
/// User-level skill config entries keyed by SKILL.md path.
pub skills: Option<SkillsConfig>,
/// Centralized feature flags (new). Prefer this over individual toggles.
#[serde(default)]
// Injects known feature keys into the schema and forbids unknown keys.
@@ -921,6 +971,8 @@ pub struct ConfigToml {
pub notice: Option<Notice>,
/// Legacy, now use features
/// Deprecated: ignored. Use `model_instructions_file`.
#[schemars(skip)]
pub experimental_instructions_file: Option<AbsolutePathBuf>,
pub experimental_compact_prompt_file: Option<AbsolutePathBuf>,
pub experimental_use_unified_exec_tool: Option<bool>,
@@ -1158,6 +1210,24 @@ pub fn resolve_oss_provider(
}
}
/// Resolve the web search mode from explicit config and feature flags.
fn resolve_web_search_mode(
config_toml: &ConfigToml,
config_profile: &ConfigProfile,
features: &Features,
) -> Option<WebSearchMode> {
if let Some(mode) = config_profile.web_search.or(config_toml.web_search) {
return Some(mode);
}
if features.enabled(Feature::WebSearchCached) {
return Some(WebSearchMode::Cached);
}
if features.enabled(Feature::WebSearchRequest) {
return Some(WebSearchMode::Live);
}
None
}
impl Config {
#[cfg(test)]
fn load_from_base_config_with_overrides(
@@ -1222,6 +1292,7 @@ impl Config {
};
let features = Features::from_config(&cfg, &config_profile, feature_overrides);
let web_search_mode = resolve_web_search_mode(&cfg, &config_profile, &features);
#[cfg(target_os = "windows")]
{
// Base flag controls sandbox on/off; elevated only applies when base is enabled.
@@ -1292,6 +1363,12 @@ impl Config {
|| cfg.sandbox_mode.is_some();
let mut model_providers = built_in_model_providers();
if features.enabled(Feature::ResponsesWebsockets)
&& let Some(provider) = model_providers.get_mut("openai")
&& provider.is_openai()
{
provider.wire_api = crate::model_provider_info::WireApi::ResponsesWebsocket;
}
// Merge user-defined providers into the built-in list.
for (key, provider) in cfg.model_providers.into_iter() {
model_providers.entry(key).or_insert(provider);
@@ -1341,7 +1418,6 @@ impl Config {
};
let include_apply_patch_tool_flag = features.enabled(Feature::ApplyPatchFreeform);
let tools_web_search_request = features.enabled(Feature::WebSearchRequest);
let use_experimental_unified_exec_tool = features.enabled(Feature::UnifiedExec);
let forced_chatgpt_workspace_id =
@@ -1370,14 +1446,12 @@ impl Config {
// Load base instructions override from a file if specified. If the
// path is relative, resolve it against the effective cwd so the
// behaviour matches other path-like config values.
let experimental_instructions_path = config_profile
.experimental_instructions_file
let model_instructions_path = config_profile
.model_instructions_file
.as_ref()
.or(cfg.experimental_instructions_file.as_ref());
let file_base_instructions = Self::try_read_non_empty_file(
experimental_instructions_path,
"experimental instructions file",
)?;
.or(cfg.model_instructions_file.as_ref());
let file_base_instructions =
Self::try_read_non_empty_file(model_instructions_path, "model instructions file")?;
let base_instructions = base_instructions.or(file_base_instructions);
let developer_instructions = developer_instructions.or(cfg.developer_instructions);
@@ -1391,10 +1465,7 @@ impl Config {
)?;
let compact_prompt = compact_prompt.or(file_compact_prompt);
// Default review model when not set in config; allow CLI override to take precedence.
let review_model = override_review_model
.or(cfg.review_model)
.unwrap_or_else(default_review_model);
let review_model = override_review_model.or(cfg.review_model);
let check_for_update_on_startup = cfg.check_for_update_on_startup.unwrap_or(true);
@@ -1403,7 +1474,7 @@ impl Config {
let ConfigRequirements {
approval_policy: mut constrained_approval_policy,
sandbox_policy: mut constrained_sandbox_policy,
mcp_server_requirements,
mcp_servers,
} = requirements;
constrained_approval_policy
@@ -1413,11 +1484,8 @@ impl Config {
.set(sandbox_policy)
.map_err(|e| std::io::Error::new(std::io::ErrorKind::InvalidInput, format!("{e}")))?;
let mcp_servers =
constrain_mcp_servers(cfg.mcp_servers.clone(), mcp_server_requirements.as_ref())
.map_err(|e| {
std::io::Error::new(std::io::ErrorKind::InvalidInput, format!("{e}"))
})?;
let mcp_servers = constrain_mcp_servers(cfg.mcp_servers.clone(), mcp_servers.as_ref())
.map_err(|e| std::io::Error::new(std::io::ErrorKind::InvalidInput, format!("{e}")))?;
let config = Self {
model,
@@ -1435,6 +1503,7 @@ impl Config {
notify: cfg.notify,
user_instructions,
base_instructions,
model_personality: config_profile.model_personality.or(cfg.model_personality),
developer_instructions,
compact_prompt,
// The config.toml omits "_mode" because it's a config file. However, "_mode"
@@ -1488,7 +1557,7 @@ impl Config {
forced_chatgpt_workspace_id,
forced_login_method,
include_apply_patch_tool: include_apply_patch_tool_flag,
tools_web_search_request,
web_search_mode,
use_experimental_unified_exec_tool,
ghost_snapshot,
features,
@@ -1625,8 +1694,28 @@ impl Config {
}
}
fn default_review_model() -> String {
OPENAI_DEFAULT_REVIEW_MODEL.to_string()
pub(crate) fn uses_deprecated_instructions_file(config_layer_stack: &ConfigLayerStack) -> bool {
config_layer_stack
.layers_high_to_low()
.into_iter()
.any(|layer| toml_uses_deprecated_instructions_file(&layer.config))
}
fn toml_uses_deprecated_instructions_file(value: &TomlValue) -> bool {
let Some(table) = value.as_table() else {
return false;
};
if table.contains_key("experimental_instructions_file") {
return true;
}
let Some(profiles) = table.get("profiles").and_then(TomlValue::as_table) else {
return false;
};
profiles.values().any(|profile| {
profile.as_table().is_some_and(|profile_table| {
profile_table.contains_key("experimental_instructions_file")
})
})
}
/// Returns the path to the Codex configuration directory, which can be
@@ -1673,6 +1762,7 @@ mod tests {
use crate::config::types::HistoryPersistence;
use crate::config::types::McpServerTransportConfig;
use crate::config::types::Notifications;
use crate::config_loader::RequirementSource;
use crate::features::Feature;
use super::*;
@@ -1694,6 +1784,7 @@ mod tests {
cwd: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -1710,6 +1801,7 @@ mod tests {
env_http_headers: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -1942,9 +2034,9 @@ trust_level = "trusted"
(MATCHED_URL_SERVER.to_string(), http_mcp(GOOD_URL)),
(DIFFERENT_NAME_SERVER.to_string(), stdio_mcp("same-cmd")),
]);
filter_mcp_servers_by_requirements(
&mut servers,
Some(&BTreeMap::from([
let source = RequirementSource::LegacyManagedConfigTomlFromMdm;
let requirements = Sourced::new(
BTreeMap::from([
(
MISMATCHED_URL_SERVER.to_string(),
McpServerRequirement {
@@ -1977,20 +2069,29 @@ trust_level = "trusted"
},
},
),
])),
]),
source.clone(),
);
filter_mcp_servers_by_requirements(&mut servers, Some(&requirements));
let reason = Some(McpServerDisabledReason::Requirements { source });
assert_eq!(
servers
.iter()
.map(|(name, server)| (name.clone(), server.enabled))
.collect::<HashMap<String, bool>>(),
.map(|(name, server)| (
name.clone(),
(server.enabled, server.disabled_reason.clone())
))
.collect::<HashMap<String, (bool, Option<McpServerDisabledReason>)>>(),
HashMap::from([
(MISMATCHED_URL_SERVER.to_string(), false),
(MISMATCHED_COMMAND_SERVER.to_string(), false),
(MATCHED_URL_SERVER.to_string(), true),
(MATCHED_COMMAND_SERVER.to_string(), true),
(DIFFERENT_NAME_SERVER.to_string(), false),
(MISMATCHED_URL_SERVER.to_string(), (false, reason.clone())),
(
MISMATCHED_COMMAND_SERVER.to_string(),
(false, reason.clone()),
),
(MATCHED_URL_SERVER.to_string(), (true, None)),
(MATCHED_COMMAND_SERVER.to_string(), (true, None)),
(DIFFERENT_NAME_SERVER.to_string(), (false, reason)),
])
);
}
@@ -2007,11 +2108,14 @@ trust_level = "trusted"
assert_eq!(
servers
.iter()
.map(|(name, server)| (name.clone(), server.enabled))
.collect::<HashMap<String, bool>>(),
.map(|(name, server)| (
name.clone(),
(server.enabled, server.disabled_reason.clone())
))
.collect::<HashMap<String, (bool, Option<McpServerDisabledReason>)>>(),
HashMap::from([
("server-a".to_string(), true),
("server-b".to_string(), true),
("server-a".to_string(), (true, None)),
("server-b".to_string(), (true, None)),
])
);
}
@@ -2023,16 +2127,22 @@ trust_level = "trusted"
("server-b".to_string(), http_mcp("https://example.com/b")),
]);
filter_mcp_servers_by_requirements(&mut servers, Some(&BTreeMap::new()));
let source = RequirementSource::LegacyManagedConfigTomlFromMdm;
let requirements = Sourced::new(BTreeMap::new(), source.clone());
filter_mcp_servers_by_requirements(&mut servers, Some(&requirements));
let reason = Some(McpServerDisabledReason::Requirements { source });
assert_eq!(
servers
.iter()
.map(|(name, server)| (name.clone(), server.enabled))
.collect::<HashMap<String, bool>>(),
.map(|(name, server)| (
name.clone(),
(server.enabled, server.disabled_reason.clone())
))
.collect::<HashMap<String, (bool, Option<McpServerDisabledReason>)>>(),
HashMap::from([
("server-a".to_string(), false),
("server-b".to_string(), false),
("server-a".to_string(), (false, reason.clone())),
("server-b".to_string(), (false, reason)),
])
);
}
@@ -2167,6 +2277,47 @@ trust_level = "trusted"
Ok(())
}
#[test]
fn web_search_mode_uses_none_if_unset() {
let cfg = ConfigToml::default();
let profile = ConfigProfile::default();
let features = Features::with_defaults();
assert_eq!(resolve_web_search_mode(&cfg, &profile, &features), None);
}
#[test]
fn web_search_mode_prefers_profile_over_legacy_flags() {
let cfg = ConfigToml::default();
let profile = ConfigProfile {
web_search: Some(WebSearchMode::Live),
..Default::default()
};
let mut features = Features::with_defaults();
features.enable(Feature::WebSearchCached);
assert_eq!(
resolve_web_search_mode(&cfg, &profile, &features),
Some(WebSearchMode::Live)
);
}
#[test]
fn web_search_mode_disabled_overrides_legacy_request() {
let cfg = ConfigToml {
web_search: Some(WebSearchMode::Disabled),
..Default::default()
};
let profile = ConfigProfile::default();
let mut features = Features::with_defaults();
features.enable(Feature::WebSearchRequest);
assert_eq!(
resolve_web_search_mode(&cfg, &profile, &features),
Some(WebSearchMode::Disabled)
);
}
#[test]
fn profile_legacy_toggles_override_base() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
@@ -2195,6 +2346,52 @@ trust_level = "trusted"
Ok(())
}
#[tokio::test]
async fn project_profile_overrides_user_profile() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
let workspace = TempDir::new()?;
let workspace_key = workspace.path().to_string_lossy().replace('\\', "\\\\");
std::fs::write(
codex_home.path().join(CONFIG_TOML_FILE),
format!(
r#"
profile = "global"
[profiles.global]
model = "gpt-global"
[profiles.project]
model = "gpt-project"
[projects."{workspace_key}"]
trust_level = "trusted"
"#,
),
)?;
let project_config_dir = workspace.path().join(".codex");
std::fs::create_dir_all(&project_config_dir)?;
std::fs::write(
project_config_dir.join(CONFIG_TOML_FILE),
r#"
profile = "project"
"#,
)?;
let config = ConfigBuilder::default()
.codex_home(codex_home.path().to_path_buf())
.harness_overrides(ConfigOverrides {
cwd: Some(workspace.path().to_path_buf()),
..Default::default()
})
.build()
.await?;
assert_eq!(config.active_profile.as_deref(), Some("project"));
assert_eq!(config.model.as_deref(), Some("gpt-project"));
Ok(())
}
#[test]
fn profile_sandbox_mode_overrides_base() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
@@ -2320,6 +2517,30 @@ trust_level = "trusted"
Ok(())
}
#[test]
fn responses_websockets_feature_updates_openai_provider() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
let mut entries = BTreeMap::new();
entries.insert("responses_websockets".to_string(), true);
let cfg = ConfigToml {
features: Some(crate::features::FeaturesToml { entries }),
..Default::default()
};
let config = Config::load_from_base_config_with_overrides(
cfg,
ConfigOverrides::default(),
codex_home.path().to_path_buf(),
)?;
assert_eq!(
config.model_provider.wire_api,
crate::model_provider_info::WireApi::ResponsesWebsocket
);
Ok(())
}
#[test]
fn config_honors_explicit_file_oauth_store_mode() -> std::io::Result<()> {
let codex_home = TempDir::new()?;
@@ -2413,6 +2634,7 @@ trust_level = "trusted"
cwd: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: Some(Duration::from_secs(3)),
tool_timeout_sec: Some(Duration::from_secs(5)),
enabled_tools: None,
@@ -2566,6 +2788,7 @@ bearer_token = "secret"
cwd: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -2634,6 +2857,7 @@ ZIG_VAR = "3"
cwd: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -2682,6 +2906,7 @@ ZIG_VAR = "3"
cwd: Some(cwd_path.clone()),
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -2728,6 +2953,7 @@ ZIG_VAR = "3"
env_http_headers: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: Some(Duration::from_secs(2)),
tool_timeout_sec: None,
enabled_tools: None,
@@ -2790,6 +3016,7 @@ startup_timeout_sec = 2.0
)])),
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: Some(Duration::from_secs(2)),
tool_timeout_sec: None,
enabled_tools: None,
@@ -2864,6 +3091,7 @@ X-Auth = "DOCS_AUTH"
)])),
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: Some(Duration::from_secs(2)),
tool_timeout_sec: None,
enabled_tools: None,
@@ -2891,6 +3119,7 @@ X-Auth = "DOCS_AUTH"
env_http_headers: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -2956,6 +3185,7 @@ url = "https://example.com/mcp"
)])),
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: Some(Duration::from_secs(2)),
tool_timeout_sec: None,
enabled_tools: None,
@@ -2973,6 +3203,7 @@ url = "https://example.com/mcp"
cwd: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -3053,6 +3284,7 @@ url = "https://example.com/mcp"
cwd: None,
},
enabled: false,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: None,
@@ -3095,6 +3327,7 @@ url = "https://example.com/mcp"
cwd: None,
},
enabled: true,
disabled_reason: None,
startup_timeout_sec: None,
tool_timeout_sec: None,
enabled_tools: Some(vec!["allowed".to_string()]),
@@ -3464,7 +3697,7 @@ model_verbosity = "high"
assert_eq!(
Config {
model: Some("o3".to_string()),
review_model: OPENAI_DEFAULT_REVIEW_MODEL.to_string(),
review_model: None,
model_context_window: None,
model_auto_compact_token_limit: None,
model_provider_id: "openai".to_string(),
@@ -3496,6 +3729,7 @@ model_verbosity = "high"
model_reasoning_summary: ReasoningSummary::Detailed,
model_supports_reasoning_summaries: None,
model_verbosity: None,
model_personality: None,
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
base_instructions: None,
developer_instructions: None,
@@ -3503,7 +3737,7 @@ model_verbosity = "high"
forced_chatgpt_workspace_id: None,
forced_login_method: None,
include_apply_patch_tool: false,
tools_web_search_request: false,
web_search_mode: None,
use_experimental_unified_exec_tool: false,
ghost_snapshot: GhostSnapshotConfig::default(),
features: Features::with_defaults(),
@@ -3551,7 +3785,7 @@ model_verbosity = "high"
)?;
let expected_gpt3_profile_config = Config {
model: Some("gpt-3.5-turbo".to_string()),
review_model: OPENAI_DEFAULT_REVIEW_MODEL.to_string(),
review_model: None,
model_context_window: None,
model_auto_compact_token_limit: None,
model_provider_id: "openai-chat-completions".to_string(),
@@ -3583,6 +3817,7 @@ model_verbosity = "high"
model_reasoning_summary: ReasoningSummary::default(),
model_supports_reasoning_summaries: None,
model_verbosity: None,
model_personality: None,
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
base_instructions: None,
developer_instructions: None,
@@ -3590,7 +3825,7 @@ model_verbosity = "high"
forced_chatgpt_workspace_id: None,
forced_login_method: None,
include_apply_patch_tool: false,
tools_web_search_request: false,
web_search_mode: None,
use_experimental_unified_exec_tool: false,
ghost_snapshot: GhostSnapshotConfig::default(),
features: Features::with_defaults(),
@@ -3653,7 +3888,7 @@ model_verbosity = "high"
)?;
let expected_zdr_profile_config = Config {
model: Some("o3".to_string()),
review_model: OPENAI_DEFAULT_REVIEW_MODEL.to_string(),
review_model: None,
model_context_window: None,
model_auto_compact_token_limit: None,
model_provider_id: "openai".to_string(),
@@ -3685,6 +3920,7 @@ model_verbosity = "high"
model_reasoning_summary: ReasoningSummary::default(),
model_supports_reasoning_summaries: None,
model_verbosity: None,
model_personality: None,
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
base_instructions: None,
developer_instructions: None,
@@ -3692,7 +3928,7 @@ model_verbosity = "high"
forced_chatgpt_workspace_id: None,
forced_login_method: None,
include_apply_patch_tool: false,
tools_web_search_request: false,
web_search_mode: None,
use_experimental_unified_exec_tool: false,
ghost_snapshot: GhostSnapshotConfig::default(),
features: Features::with_defaults(),
@@ -3741,7 +3977,7 @@ model_verbosity = "high"
)?;
let expected_gpt5_profile_config = Config {
model: Some("gpt-5.1".to_string()),
review_model: OPENAI_DEFAULT_REVIEW_MODEL.to_string(),
review_model: None,
model_context_window: None,
model_auto_compact_token_limit: None,
model_provider_id: "openai".to_string(),
@@ -3773,6 +4009,7 @@ model_verbosity = "high"
model_reasoning_summary: ReasoningSummary::Detailed,
model_supports_reasoning_summaries: None,
model_verbosity: Some(Verbosity::High),
model_personality: None,
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
base_instructions: None,
developer_instructions: None,
@@ -3780,7 +4017,7 @@ model_verbosity = "high"
forced_chatgpt_workspace_id: None,
forced_login_method: None,
include_apply_patch_tool: false,
tools_web_search_request: false,
web_search_mode: None,
use_experimental_unified_exec_tool: false,
ghost_snapshot: GhostSnapshotConfig::default(),
features: Features::with_defaults(),

View File

@@ -3,10 +3,12 @@ use schemars::JsonSchema;
use serde::Deserialize;
use serde::Serialize;
use crate::config::types::Personality;
use crate::protocol::AskForApproval;
use codex_protocol::config_types::ReasoningSummary;
use codex_protocol::config_types::SandboxMode;
use codex_protocol::config_types::Verbosity;
use codex_protocol::config_types::WebSearchMode;
use codex_protocol::openai_models::ReasoningEffort;
/// Collection of common configuration options that a user can define as a unit
@@ -23,7 +25,12 @@ pub struct ConfigProfile {
pub model_reasoning_effort: Option<ReasoningEffort>,
pub model_reasoning_summary: Option<ReasoningSummary>,
pub model_verbosity: Option<Verbosity>,
pub model_personality: Option<Personality>,
pub chatgpt_base_url: Option<String>,
/// Optional path to a file containing model instructions.
pub model_instructions_file: Option<AbsolutePathBuf>,
/// Deprecated: ignored. Use `model_instructions_file`.
#[schemars(skip)]
pub experimental_instructions_file: Option<AbsolutePathBuf>,
pub experimental_compact_prompt_file: Option<AbsolutePathBuf>,
pub include_apply_patch_tool: Option<bool>,
@@ -31,6 +38,7 @@ pub struct ConfigProfile {
pub experimental_use_freeform_apply_patch: Option<bool>,
pub tools_web_search: Option<bool>,
pub tools_view_image: Option<bool>,
pub web_search: Option<WebSearchMode>,
pub analytics: Option<crate::config::types::AnalyticsConfigToml>,
/// Optional feature toggles scoped to this profile.
#[serde(default)]

View File

@@ -9,6 +9,9 @@ use crate::config_loader::LoaderOverrides;
use crate::config_loader::load_config_layers_state;
use crate::config_loader::merge_toml_values;
use crate::path_utils;
use crate::path_utils::SymlinkWritePaths;
use crate::path_utils::resolve_symlink_write_paths;
use crate::path_utils::write_atomically;
use codex_app_server_protocol::Config as ApiConfig;
use codex_app_server_protocol::ConfigBatchWriteParams;
use codex_app_server_protocol::ConfigLayerMetadata;
@@ -27,6 +30,7 @@ use std::borrow::Cow;
use std::path::Path;
use std::path::PathBuf;
use thiserror::Error;
use tokio::task;
use toml::Value as TomlValue;
use toml_edit::Item as TomlItem;
@@ -362,19 +366,30 @@ impl ConfigService {
async fn create_empty_user_layer(
config_toml: &AbsolutePathBuf,
) -> Result<ConfigLayerEntry, ConfigServiceError> {
let toml_value = match tokio::fs::read_to_string(config_toml).await {
Ok(contents) => toml::from_str(&contents).map_err(|e| {
ConfigServiceError::toml("failed to parse existing user config.toml", e)
})?,
Err(e) => {
if e.kind() == std::io::ErrorKind::NotFound {
tokio::fs::write(config_toml, "").await.map_err(|e| {
ConfigServiceError::io("failed to create empty user config.toml", e)
})?;
let SymlinkWritePaths {
read_path,
write_path,
} = resolve_symlink_write_paths(config_toml.as_path())
.map_err(|err| ConfigServiceError::io("failed to resolve user config path", err))?;
let toml_value = match read_path {
Some(path) => match tokio::fs::read_to_string(&path).await {
Ok(contents) => toml::from_str(&contents).map_err(|e| {
ConfigServiceError::toml("failed to parse existing user config.toml", e)
})?,
Err(err) if err.kind() == std::io::ErrorKind::NotFound => {
write_empty_user_config(write_path.clone()).await?;
TomlValue::Table(toml::map::Map::new())
} else {
return Err(ConfigServiceError::io("failed to read user config.toml", e));
}
Err(err) => {
return Err(ConfigServiceError::io(
"failed to read user config.toml",
err,
));
}
},
None => {
write_empty_user_config(write_path).await?;
TomlValue::Table(toml::map::Map::new())
}
};
Ok(ConfigLayerEntry::new(
@@ -385,6 +400,13 @@ async fn create_empty_user_layer(
))
}
async fn write_empty_user_config(write_path: PathBuf) -> Result<(), ConfigServiceError> {
task::spawn_blocking(move || write_atomically(&write_path, ""))
.await
.map_err(|err| ConfigServiceError::anyhow("config persistence task panicked", err.into()))?
.map_err(|err| ConfigServiceError::io("failed to create empty user config.toml", err))
}
fn parse_value(value: JsonValue) -> Result<Option<TomlValue>, String> {
if value.is_null() {
return Ok(None);

View File

@@ -3,10 +3,13 @@
// Note this file should generally be restricted to simple struct/enum
// definitions that do not contain business logic.
use crate::config_loader::RequirementSource;
pub use codex_protocol::config_types::AltScreenMode;
pub use codex_protocol::config_types::WebSearchMode;
use codex_utils_absolute_path::AbsolutePathBuf;
use std::collections::BTreeMap;
use std::collections::HashMap;
use std::fmt;
use std::path::PathBuf;
use std::time::Duration;
use wildmatch::WildMatchPattern;
@@ -19,6 +22,23 @@ use serde::de::Error as SerdeError;
pub const DEFAULT_OTEL_ENVIRONMENT: &str = "dev";
#[derive(Debug, Clone, PartialEq, Eq)]
pub enum McpServerDisabledReason {
Unknown,
Requirements { source: RequirementSource },
}
impl fmt::Display for McpServerDisabledReason {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
match self {
McpServerDisabledReason::Unknown => write!(f, "unknown"),
McpServerDisabledReason::Requirements { source } => {
write!(f, "requirements ({source})")
}
}
}
}
#[derive(Serialize, Debug, Clone, PartialEq)]
pub struct McpServerConfig {
#[serde(flatten)]
@@ -28,6 +48,10 @@ pub struct McpServerConfig {
#[serde(default = "default_enabled")]
pub enabled: bool,
/// Reason this server was disabled after applying requirements.
#[serde(skip)]
pub disabled_reason: Option<McpServerDisabledReason>,
/// Startup timeout in seconds for initializing MCP server & initially listing tools.
#[serde(
default,
@@ -159,6 +183,7 @@ impl<'de> Deserialize<'de> for McpServerConfig {
startup_timeout_sec,
tool_timeout_sec,
enabled,
disabled_reason: None,
enabled_tools,
disabled_tools,
})
@@ -398,10 +423,11 @@ impl Default for Notifications {
/// Terminals generally encode both mouse wheels and trackpads as the same "scroll up/down" mouse
/// button events, without a magnitude. This setting controls whether Codex uses a heuristic to
/// infer wheel vs trackpad per stream, or forces a specific behavior.
#[derive(Serialize, Deserialize, Debug, Clone, Copy, PartialEq, Eq, JsonSchema)]
#[derive(Serialize, Deserialize, Debug, Clone, Copy, PartialEq, Eq, JsonSchema, Default)]
#[serde(rename_all = "snake_case")]
pub enum ScrollInputMode {
/// Infer wheel vs trackpad behavior per scroll stream.
#[default]
Auto,
/// Always treat scroll events as mouse-wheel input (fixed lines per tick).
Wheel,
@@ -409,12 +435,6 @@ pub enum ScrollInputMode {
Trackpad,
}
impl Default for ScrollInputMode {
fn default() -> Self {
Self::Auto
}
}
/// Collection of settings that are specific to the TUI.
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Default, JsonSchema)]
#[schemars(deny_unknown_fields)]
@@ -580,6 +600,20 @@ impl Notice {
pub(crate) const TABLE_KEY: &'static str = "notice";
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Eq, JsonSchema)]
#[schemars(deny_unknown_fields)]
pub struct SkillConfig {
pub path: AbsolutePathBuf,
pub enabled: bool,
}
#[derive(Serialize, Deserialize, Debug, Clone, Default, PartialEq, Eq, JsonSchema)]
#[schemars(deny_unknown_fields)]
pub struct SkillsConfig {
#[serde(default, skip_serializing_if = "Vec::is_empty")]
pub config: Vec<SkillConfig>,
}
#[derive(Serialize, Deserialize, Debug, Clone, PartialEq, Default, JsonSchema)]
#[schemars(deny_unknown_fields)]
pub struct SandboxWorkspaceWrite {
@@ -714,6 +748,14 @@ impl Default for ShellEnvironmentPolicy {
}
}
#[derive(Default, Debug, Clone, Eq, PartialEq, Serialize, Deserialize, JsonSchema)]
#[serde(rename_all = "kebab-case")]
pub enum Personality {
Friendly,
#[default]
Pragmatic,
}
#[cfg(test)]
mod tests {
use super::*;

View File

@@ -44,7 +44,7 @@ impl fmt::Display for RequirementSource {
pub struct ConfigRequirements {
pub approval_policy: Constrained<AskForApproval>,
pub sandbox_policy: Constrained<SandboxPolicy>,
pub mcp_server_requirements: Option<BTreeMap<String, McpServerRequirement>>,
pub mcp_servers: Option<Sourced<BTreeMap<String, McpServerRequirement>>>,
}
impl Default for ConfigRequirements {
@@ -52,7 +52,7 @@ impl Default for ConfigRequirements {
Self {
approval_policy: Constrained::allow_any_from_default(),
sandbox_policy: Constrained::allow_any(SandboxPolicy::ReadOnly),
mcp_server_requirements: None,
mcp_servers: None,
}
}
}
@@ -74,7 +74,7 @@ pub struct McpServerRequirement {
pub struct ConfigRequirementsToml {
pub allowed_approval_policies: Option<Vec<AskForApproval>>,
pub allowed_sandbox_modes: Option<Vec<SandboxModeRequirement>>,
pub mcp_server_requirements: Option<BTreeMap<String, McpServerRequirement>>,
pub mcp_servers: Option<BTreeMap<String, McpServerRequirement>>,
}
/// Value paired with the requirement source it came from, for better error
@@ -103,7 +103,7 @@ impl<T> std::ops::Deref for Sourced<T> {
pub struct ConfigRequirementsWithSources {
pub allowed_approval_policies: Option<Sourced<Vec<AskForApproval>>>,
pub allowed_sandbox_modes: Option<Sourced<Vec<SandboxModeRequirement>>>,
pub mcp_server_requirements: Option<Sourced<BTreeMap<String, McpServerRequirement>>>,
pub mcp_servers: Option<Sourced<BTreeMap<String, McpServerRequirement>>>,
}
impl ConfigRequirementsWithSources {
@@ -134,7 +134,7 @@ impl ConfigRequirementsWithSources {
{
allowed_approval_policies,
allowed_sandbox_modes,
mcp_server_requirements,
mcp_servers,
}
);
}
@@ -143,12 +143,12 @@ impl ConfigRequirementsWithSources {
let ConfigRequirementsWithSources {
allowed_approval_policies,
allowed_sandbox_modes,
mcp_server_requirements,
mcp_servers,
} = self;
ConfigRequirementsToml {
allowed_approval_policies: allowed_approval_policies.map(|sourced| sourced.value),
allowed_sandbox_modes: allowed_sandbox_modes.map(|sourced| sourced.value),
mcp_server_requirements: mcp_server_requirements.map(|sourced| sourced.value),
mcp_servers: mcp_servers.map(|sourced| sourced.value),
}
}
}
@@ -184,7 +184,7 @@ impl ConfigRequirementsToml {
pub fn is_empty(&self) -> bool {
self.allowed_approval_policies.is_none()
&& self.allowed_sandbox_modes.is_none()
&& self.mcp_server_requirements.is_none()
&& self.mcp_servers.is_none()
}
}
@@ -195,7 +195,7 @@ impl TryFrom<ConfigRequirementsWithSources> for ConfigRequirements {
let ConfigRequirementsWithSources {
allowed_approval_policies,
allowed_sandbox_modes,
mcp_server_requirements,
mcp_servers,
} = toml;
let approval_policy: Constrained<AskForApproval> = match allowed_approval_policies {
@@ -273,7 +273,7 @@ impl TryFrom<ConfigRequirementsWithSources> for ConfigRequirements {
Ok(ConfigRequirements {
approval_policy,
sandbox_policy,
mcp_server_requirements: mcp_server_requirements.map(|sourced| sourced.value),
mcp_servers,
})
}
}
@@ -291,15 +291,14 @@ mod tests {
let ConfigRequirementsToml {
allowed_approval_policies,
allowed_sandbox_modes,
mcp_server_requirements,
mcp_servers,
} = toml;
ConfigRequirementsWithSources {
allowed_approval_policies: allowed_approval_policies
.map(|value| Sourced::new(value, RequirementSource::Unknown)),
allowed_sandbox_modes: allowed_sandbox_modes
.map(|value| Sourced::new(value, RequirementSource::Unknown)),
mcp_server_requirements: mcp_server_requirements
.map(|value| Sourced::new(value, RequirementSource::Unknown)),
mcp_servers: mcp_servers.map(|value| Sourced::new(value, RequirementSource::Unknown)),
}
}
@@ -319,7 +318,7 @@ mod tests {
let other = ConfigRequirementsToml {
allowed_approval_policies: Some(allowed_approval_policies.clone()),
allowed_sandbox_modes: Some(allowed_sandbox_modes.clone()),
mcp_server_requirements: None,
mcp_servers: None,
};
target.merge_unset_fields(source.clone(), other);
@@ -332,7 +331,7 @@ mod tests {
source.clone()
)),
allowed_sandbox_modes: Some(Sourced::new(allowed_sandbox_modes, source)),
mcp_server_requirements: None,
mcp_servers: None,
}
);
}
@@ -360,7 +359,7 @@ mod tests {
source_location,
)),
allowed_sandbox_modes: None,
mcp_server_requirements: None,
mcp_servers: None,
}
);
Ok(())
@@ -396,7 +395,7 @@ mod tests {
existing_source,
)),
allowed_sandbox_modes: None,
mcp_server_requirements: None,
mcp_servers: None,
}
);
Ok(())
@@ -561,35 +560,38 @@ mod tests {
#[test]
fn deserialize_mcp_server_requirements() -> Result<()> {
let toml_str = r#"
[mcp_server_requirements.docs.identity]
[mcp_servers.docs.identity]
command = "codex-mcp"
[mcp_server_requirements.remote.identity]
[mcp_servers.remote.identity]
url = "https://example.com/mcp"
"#;
let requirements: ConfigRequirements =
with_unknown_source(from_str(toml_str)?).try_into()?;
assert_eq!(
requirements.mcp_server_requirements,
Some(BTreeMap::from([
(
"docs".to_string(),
McpServerRequirement {
identity: McpServerIdentity::Command {
command: "codex-mcp".to_string(),
requirements.mcp_servers,
Some(Sourced::new(
BTreeMap::from([
(
"docs".to_string(),
McpServerRequirement {
identity: McpServerIdentity::Command {
command: "codex-mcp".to_string(),
},
},
},
),
(
"remote".to_string(),
McpServerRequirement {
identity: McpServerIdentity::Url {
url: "https://example.com/mcp".to_string(),
),
(
"remote".to_string(),
McpServerRequirement {
identity: McpServerIdentity::Url {
url: "https://example.com/mcp".to_string(),
},
},
},
),
]))
),
]),
RequirementSource::Unknown,
))
);
Ok(())
}

View File

@@ -12,10 +12,13 @@ mod tests;
use crate::config::CONFIG_TOML_FILE;
use crate::config::ConfigToml;
use crate::config::deserialize_config_toml_with_base;
use crate::config_loader::config_requirements::ConfigRequirementsWithSources;
use crate::config_loader::layer_io::LoadedConfigLayers;
use crate::git_info::resolve_root_git_project_for_trust;
use codex_app_server_protocol::ConfigLayerSource;
use codex_protocol::config_types::SandboxMode;
use codex_protocol::config_types::TrustLevel;
use codex_protocol::protocol::AskForApproval;
use codex_utils_absolute_path::AbsolutePathBuf;
use codex_utils_absolute_path::AbsolutePathBufGuard;
@@ -30,7 +33,9 @@ pub use config_requirements::McpServerIdentity;
pub use config_requirements::McpServerRequirement;
pub use config_requirements::RequirementSource;
pub use config_requirements::SandboxModeRequirement;
pub use config_requirements::Sourced;
pub use merge::merge_toml_values;
pub(crate) use overrides::build_cli_overrides_layer;
pub use state::ConfigLayerEntry;
pub use state::ConfigLayerStack;
pub use state::ConfigLayerStackOrdering;
@@ -62,9 +67,9 @@ const DEFAULT_PROJECT_ROOT_MARKERS: &[&str] = &[".git"];
/// - admin: managed preferences (*)
/// - system `/etc/codex/config.toml`
/// - user `${CODEX_HOME}/config.toml`
/// - cwd `${PWD}/config.toml`
/// - tree parent directories up to root looking for `./.codex/config.toml`
/// - repo `$(git rev-parse --show-toplevel)/.codex/config.toml`
/// - cwd `${PWD}/config.toml` (only when the directory is trusted)
/// - tree parent directories up to root looking for `./.codex/config.toml` (trusted only)
/// - repo `$(git rev-parse --show-toplevel)/.codex/config.toml` (trusted only)
/// - runtime e.g., --config flags, model selector in UI
///
/// (*) Only available on macOS via managed device profiles.
@@ -112,6 +117,12 @@ pub async fn load_config_layers_state(
let mut layers = Vec::<ConfigLayerEntry>::new();
let cli_overrides_layer = if cli_overrides.is_empty() {
None
} else {
Some(overrides::build_cli_overrides_layer(cli_overrides))
};
// Include an entry for the "system" config folder, loading its config.toml,
// if it exists.
let system_config_toml_file = if cfg!(unix) {
@@ -156,17 +167,22 @@ pub async fn load_config_layers_state(
for layer in &layers {
merge_toml_values(&mut merged_so_far, &layer.config);
}
if let Some(cli_overrides_layer) = cli_overrides_layer.as_ref() {
merge_toml_values(&mut merged_so_far, cli_overrides_layer);
}
let project_root_markers = project_root_markers_from_config(&merged_so_far)?
.unwrap_or_else(default_project_root_markers);
let project_root = find_project_root(&cwd, &project_root_markers).await?;
let project_layers = load_project_layers(&cwd, &project_root).await?;
layers.extend(project_layers);
if let Some(project_root) =
trusted_project_root(&merged_so_far, &cwd, &project_root_markers, codex_home).await?
{
let project_layers = load_project_layers(&cwd, &project_root).await?;
layers.extend(project_layers);
}
}
// Add a layer for runtime overrides from the CLI or UI, if any exist.
if !cli_overrides.is_empty() {
let cli_overrides_layer = overrides::build_cli_overrides_layer(cli_overrides);
if let Some(cli_overrides_layer) = cli_overrides_layer {
layers.push(ConfigLayerEntry::new(
ConfigLayerSource::SessionFlags,
cli_overrides_layer,
@@ -386,6 +402,44 @@ fn default_project_root_markers() -> Vec<String> {
.collect()
}
async fn trusted_project_root(
merged_config: &TomlValue,
cwd: &AbsolutePathBuf,
project_root_markers: &[String],
config_base_dir: &Path,
) -> io::Result<Option<AbsolutePathBuf>> {
let config_toml = deserialize_config_toml_with_base(merged_config.clone(), config_base_dir)?;
let project_root = find_project_root(cwd, project_root_markers).await?;
let projects = config_toml.projects.unwrap_or_default();
let cwd_key = cwd.as_path().to_string_lossy().to_string();
let project_root_key = project_root.as_path().to_string_lossy().to_string();
let repo_root_key = resolve_root_git_project_for_trust(cwd.as_path())
.map(|root| root.to_string_lossy().to_string());
let trust_level = projects
.get(&cwd_key)
.and_then(|project| project.trust_level)
.or_else(|| {
projects
.get(&project_root_key)
.and_then(|project| project.trust_level)
})
.or_else(|| {
repo_root_key
.as_ref()
.and_then(|root| projects.get(root))
.and_then(|project| project.trust_level)
});
if matches!(trust_level, Some(TrustLevel::Trusted)) {
Ok(Some(project_root))
} else {
Ok(None)
}
}
/// Takes a `toml::Value` parsed from a config.toml file and walks through it,
/// resolving any `AbsolutePathBuf` fields against `base_dir`, returning a new
/// `toml::Value` with the same shape but with paths resolved.
@@ -602,7 +656,7 @@ mod unit_tests {
let contents = r#"
# This is a field recognized by config.toml that is an AbsolutePathBuf in
# the ConfigToml struct.
experimental_instructions_file = "./some_file.md"
model_instructions_file = "./some_file.md"
# This is a field recognized by config.toml.
model = "gpt-1000"
@@ -615,7 +669,7 @@ foo = "xyzzy"
let normalized_toml_value = resolve_relative_paths_in_config_toml(user_config, base_dir)?;
let mut expected_toml_value = toml::map::Map::new();
expected_toml_value.insert(
"experimental_instructions_file".to_string(),
"model_instructions_file".to_string(),
TomlValue::String(
AbsolutePathBuf::resolve_path_against_base("./some_file.md", base_dir)?
.as_path()

View File

@@ -1,10 +1,10 @@
use toml::Value as TomlValue;
pub(super) fn default_empty_table() -> TomlValue {
pub(crate) fn default_empty_table() -> TomlValue {
TomlValue::Table(Default::default())
}
pub(super) fn build_cli_overrides_layer(cli_overrides: &[(String, TomlValue)]) -> TomlValue {
pub(crate) fn build_cli_overrides_layer(cli_overrides: &[(String, TomlValue)]) -> TomlValue {
let mut root = default_empty_table();
for (path, value) in cli_overrides {
apply_toml_override(&mut root, path, value.clone());

View File

@@ -3,19 +3,47 @@ use super::load_config_layers_state;
use crate::config::CONFIG_TOML_FILE;
use crate::config::ConfigBuilder;
use crate::config::ConfigOverrides;
use crate::config::ConfigToml;
use crate::config::ProjectConfig;
use crate::config_loader::ConfigLayerEntry;
use crate::config_loader::ConfigRequirements;
use crate::config_loader::config_requirements::ConfigRequirementsWithSources;
use crate::config_loader::fingerprint::version_for_toml;
use crate::config_loader::load_requirements_toml;
use codex_protocol::config_types::TrustLevel;
use codex_protocol::protocol::AskForApproval;
#[cfg(target_os = "macos")]
use codex_protocol::protocol::SandboxPolicy;
use codex_utils_absolute_path::AbsolutePathBuf;
use pretty_assertions::assert_eq;
use std::collections::HashMap;
use std::path::Path;
use tempfile::tempdir;
use toml::Value as TomlValue;
async fn make_config_for_test(
codex_home: &Path,
project_path: &Path,
trust_level: TrustLevel,
project_root_markers: Option<Vec<String>>,
) -> std::io::Result<()> {
tokio::fs::write(
codex_home.join(CONFIG_TOML_FILE),
toml::to_string(&ConfigToml {
projects: Some(HashMap::from([(
project_path.to_string_lossy().to_string(),
ProjectConfig {
trust_level: Some(trust_level),
},
)])),
project_root_markers,
..Default::default()
})
.expect("serialize config"),
)
.await
}
#[tokio::test]
async fn merges_managed_config_layer_on_top() {
let tmp = tempdir().expect("tempdir");
@@ -365,6 +393,7 @@ async fn project_layers_prefer_closest_cwd() -> std::io::Result<()> {
let codex_home = tmp.path().join("home");
tokio::fs::create_dir_all(&codex_home).await?;
make_config_for_test(&codex_home, &project_root, TrustLevel::Trusted, None).await?;
let cwd = AbsolutePathBuf::from_absolute_path(&nested)?;
let layers = load_config_layers_state(
&codex_home,
@@ -409,10 +438,10 @@ async fn project_paths_resolve_relative_to_dot_codex_and_override_in_order() ->
tokio::fs::write(project_root.join(".git"), "gitdir: here").await?;
let root_cfg = r#"
experimental_instructions_file = "root.txt"
model_instructions_file = "root.txt"
"#;
let nested_cfg = r#"
experimental_instructions_file = "child.txt"
model_instructions_file = "child.txt"
"#;
tokio::fs::write(project_root.join(".codex").join(CONFIG_TOML_FILE), root_cfg).await?;
tokio::fs::write(nested.join(".codex").join(CONFIG_TOML_FILE), nested_cfg).await?;
@@ -429,6 +458,7 @@ experimental_instructions_file = "child.txt"
let codex_home = tmp.path().join("home");
tokio::fs::create_dir_all(&codex_home).await?;
make_config_for_test(&codex_home, &project_root, TrustLevel::Trusted, None).await?;
let config = ConfigBuilder::default()
.codex_home(codex_home)
@@ -458,6 +488,7 @@ async fn project_layer_is_added_when_dot_codex_exists_without_config_toml() -> s
let codex_home = tmp.path().join("home");
tokio::fs::create_dir_all(&codex_home).await?;
make_config_for_test(&codex_home, &project_root, TrustLevel::Trusted, None).await?;
let cwd = AbsolutePathBuf::from_absolute_path(&nested)?;
let layers = load_config_layers_state(
&codex_home,
@@ -486,6 +517,95 @@ async fn project_layer_is_added_when_dot_codex_exists_without_config_toml() -> s
Ok(())
}
#[tokio::test]
async fn project_layers_skipped_when_untrusted_or_unknown() -> std::io::Result<()> {
let tmp = tempdir()?;
let project_root = tmp.path().join("project");
let nested = project_root.join("child");
tokio::fs::create_dir_all(nested.join(".codex")).await?;
tokio::fs::write(
nested.join(".codex").join(CONFIG_TOML_FILE),
"foo = \"child\"\n",
)
.await?;
let cwd = AbsolutePathBuf::from_absolute_path(&nested)?;
let codex_home_untrusted = tmp.path().join("home_untrusted");
tokio::fs::create_dir_all(&codex_home_untrusted).await?;
make_config_for_test(
&codex_home_untrusted,
&project_root,
TrustLevel::Untrusted,
None,
)
.await?;
let layers_untrusted = load_config_layers_state(
&codex_home_untrusted,
Some(cwd.clone()),
&[] as &[(String, TomlValue)],
LoaderOverrides::default(),
)
.await?;
let project_layers_untrusted = layers_untrusted
.layers_high_to_low()
.into_iter()
.filter(|layer| matches!(layer.name, super::ConfigLayerSource::Project { .. }))
.count();
assert_eq!(project_layers_untrusted, 0);
assert_eq!(layers_untrusted.effective_config().get("foo"), None);
let codex_home_unknown = tmp.path().join("home_unknown");
tokio::fs::create_dir_all(&codex_home_unknown).await?;
let layers_unknown = load_config_layers_state(
&codex_home_unknown,
Some(cwd),
&[] as &[(String, TomlValue)],
LoaderOverrides::default(),
)
.await?;
let project_layers_unknown = layers_unknown
.layers_high_to_low()
.into_iter()
.filter(|layer| matches!(layer.name, super::ConfigLayerSource::Project { .. }))
.count();
assert_eq!(project_layers_unknown, 0);
assert_eq!(layers_unknown.effective_config().get("foo"), None);
Ok(())
}
#[tokio::test]
async fn cli_overrides_with_relative_paths_do_not_break_trust_check() -> std::io::Result<()> {
let tmp = tempdir()?;
let project_root = tmp.path().join("project");
let nested = project_root.join("child");
tokio::fs::create_dir_all(&nested).await?;
tokio::fs::write(project_root.join(".git"), "gitdir: here").await?;
let codex_home = tmp.path().join("home");
tokio::fs::create_dir_all(&codex_home).await?;
make_config_for_test(&codex_home, &project_root, TrustLevel::Trusted, None).await?;
let cwd = AbsolutePathBuf::from_absolute_path(&nested)?;
let cli_overrides = vec![(
"model_instructions_file".to_string(),
TomlValue::String("relative.md".to_string()),
)];
load_config_layers_state(
&codex_home,
Some(cwd),
&cli_overrides,
LoaderOverrides::default(),
)
.await?;
Ok(())
}
#[tokio::test]
async fn project_root_markers_supports_alternate_markers() -> std::io::Result<()> {
let tmp = tempdir()?;
@@ -507,11 +627,11 @@ async fn project_root_markers_supports_alternate_markers() -> std::io::Result<()
let codex_home = tmp.path().join("home");
tokio::fs::create_dir_all(&codex_home).await?;
tokio::fs::write(
codex_home.join(CONFIG_TOML_FILE),
r#"
project_root_markers = [".hg"]
"#,
make_config_for_test(
&codex_home,
&project_root,
TrustLevel::Trusted,
Some(vec![".hg".to_string()]),
)
.await?;

View File

@@ -1,12 +1,13 @@
use crate::codex::TurnContext;
use crate::context_manager::normalize;
use crate::instructions::SkillInstructions;
use crate::instructions::UserInstructions;
use crate::session_prefix::is_session_prefix;
use crate::truncate::TruncationPolicy;
use crate::truncate::approx_token_count;
use crate::truncate::approx_tokens_from_byte_count;
use crate::truncate::truncate_function_output_items_with_policy;
use crate::truncate::truncate_text;
use crate::user_instructions::SkillInstructions;
use crate::user_instructions::UserInstructions;
use crate::user_shell_command::is_user_shell_command_text;
use codex_protocol::models::ContentItem;
use codex_protocol::models::FunctionCallOutputContentItem;
@@ -97,7 +98,11 @@ impl ContextManager {
}
| ResponseItem::Compaction {
encrypted_content: content,
} => estimate_reasoning_length(content.len()) as i64,
} => {
let reasoning_bytes = estimate_reasoning_length(content.len());
i64::try_from(approx_tokens_from_byte_count(reasoning_bytes))
.unwrap_or(i64::MAX)
}
item => {
let serialized = serde_json::to_string(item).unwrap_or_default();
i64::try_from(approx_token_count(&serialized)).unwrap_or(i64::MAX)
@@ -124,34 +129,35 @@ impl ContextManager {
self.items = items;
}
pub(crate) fn replace_last_turn_images(&mut self, placeholder: &str) {
let Some(last_item) = self.items.last_mut() else {
return;
/// Replace image content in the last turn if it originated from a tool output.
/// Returns true when a tool image was replaced, false otherwise.
pub(crate) fn replace_last_turn_images(&mut self, placeholder: &str) -> bool {
let Some(index) = self.items.iter().rposition(|item| {
matches!(item, ResponseItem::FunctionCallOutput { .. })
|| matches!(item, ResponseItem::Message { role, .. } if role == "user")
}) else {
return false;
};
match last_item {
ResponseItem::Message { role, content, .. } if role == "user" => {
for item in content.iter_mut() {
if matches!(item, ContentItem::InputImage { .. }) {
*item = ContentItem::InputText {
text: placeholder.to_string(),
};
}
}
}
match &mut self.items[index] {
ResponseItem::FunctionCallOutput { output, .. } => {
let Some(content_items) = output.content_items.as_mut() else {
return;
return false;
};
let mut replaced = false;
let placeholder = placeholder.to_string();
for item in content_items.iter_mut() {
if matches!(item, FunctionCallOutputContentItem::InputImage { .. }) {
*item = FunctionCallOutputContentItem::InputText {
text: placeholder.to_string(),
text: placeholder.clone(),
};
replaced = true;
}
}
replaced
}
_ => {}
ResponseItem::Message { role, .. } if role == "user" => false,
_ => false,
}
}
@@ -230,12 +236,19 @@ impl ContextManager {
token_estimate as usize
}
pub(crate) fn get_total_token_usage(&self) -> i64 {
self.token_info
/// When true, the server already accounted for past reasoning tokens and
/// the client should not re-estimate them.
pub(crate) fn get_total_token_usage(&self, server_reasoning_included: bool) -> i64 {
let last_tokens = self
.token_info
.as_ref()
.map(|info| info.last_token_usage.total_tokens)
.unwrap_or(0)
.saturating_add(self.get_non_last_reasoning_items_tokens() as i64)
.unwrap_or(0);
if server_reasoning_included {
last_tokens
} else {
last_tokens.saturating_add(self.get_non_last_reasoning_items_tokens() as i64)
}
}
/// This function enforces a couple of invariants on the in-memory history:
@@ -316,12 +329,6 @@ fn estimate_reasoning_length(encoded_len: usize) -> usize {
.saturating_sub(650)
}
fn is_session_prefix(text: &str) -> bool {
let trimmed = text.trim_start();
let lowered = trimmed.to_ascii_lowercase();
lowered.starts_with("<environment_context>")
}
pub(crate) fn is_user_turn_boundary(item: &ResponseItem) -> bool {
let ResponseItem::Message { role, content, .. } = item else {
return false;

View File

@@ -3,6 +3,7 @@ use crate::truncate;
use crate::truncate::TruncationPolicy;
use codex_git::GhostCommit;
use codex_protocol::models::ContentItem;
use codex_protocol::models::FunctionCallOutputContentItem;
use codex_protocol::models::FunctionCallOutputPayload;
use codex_protocol::models::LocalShellAction;
use codex_protocol::models::LocalShellExecAction;
@@ -209,6 +210,58 @@ fn remove_first_item_removes_matching_call_for_output() {
assert_eq!(h.raw_items(), vec![]);
}
#[test]
fn replace_last_turn_images_replaces_tool_output_images() {
let items = vec![
user_input_text_msg("hi"),
ResponseItem::FunctionCallOutput {
call_id: "call-1".to_string(),
output: FunctionCallOutputPayload {
content: "ok".to_string(),
content_items: Some(vec![FunctionCallOutputContentItem::InputImage {
image_url: "data:image/png;base64,AAA".to_string(),
}]),
success: Some(true),
},
},
];
let mut history = create_history_with_items(items);
assert!(history.replace_last_turn_images("Invalid image"));
assert_eq!(
history.raw_items(),
vec![
user_input_text_msg("hi"),
ResponseItem::FunctionCallOutput {
call_id: "call-1".to_string(),
output: FunctionCallOutputPayload {
content: "ok".to_string(),
content_items: Some(vec![FunctionCallOutputContentItem::InputText {
text: "Invalid image".to_string(),
}]),
success: Some(true),
},
},
]
);
}
#[test]
fn replace_last_turn_images_does_not_touch_user_images() {
let items = vec![ResponseItem::Message {
id: None,
role: "user".to_string(),
content: vec![ContentItem::InputImage {
image_url: "data:image/png;base64,AAA".to_string(),
}],
}];
let mut history = create_history_with_items(items.clone());
assert!(!history.replace_last_turn_images("Invalid image"));
assert_eq!(history.raw_items(), items);
}
#[test]
fn remove_first_item_handles_local_shell_pair() {
let items = vec![

View File

@@ -169,7 +169,8 @@ pub fn build_reqwest_client() -> reqwest::Client {
let mut builder = reqwest::Client::builder()
// Set UA via dedicated helper to avoid header validation pitfalls
.user_agent(ua)
.default_headers(headers);
.default_headers(headers)
.use_rustls_tls();
if is_sandboxed() {
builder = builder.no_proxy();
}

View File

@@ -17,16 +17,11 @@ use codex_protocol::user_input::UserInput;
use tracing::warn;
use uuid::Uuid;
use crate::user_instructions::SkillInstructions;
use crate::user_instructions::UserInstructions;
use crate::instructions::SkillInstructions;
use crate::instructions::UserInstructions;
use crate::session_prefix::is_session_prefix;
use crate::user_shell_command::is_user_shell_command_text;
fn is_session_prefix(text: &str) -> bool {
let trimmed = text.trim_start();
let lowered = trimmed.to_ascii_lowercase();
lowered.starts_with("<environment_context>")
}
fn parse_user_message(message: &[ContentItem]) -> Option<UserMessageItem> {
if UserInstructions::is_user_instructions(message)
|| SkillInstructions::is_skill_instructions(message)
@@ -50,7 +45,11 @@ fn parse_user_message(message: &[ContentItem]) -> Option<UserMessageItem> {
if is_session_prefix(text) || is_user_shell_command_text(text) {
return None;
}
content.push(UserInput::Text { text: text.clone() });
content.push(UserInput::Text {
text: text.clone(),
// Model input content does not carry UI element ranges.
text_elements: Vec::new(),
});
}
ContentItem::InputImage { image_url } => {
content.push(UserInput::Image {
@@ -179,6 +178,7 @@ mod tests {
let expected_content = vec![
UserInput::Text {
text: "Hello world".to_string(),
text_elements: Vec::new(),
},
UserInput::Image { image_url: img1 },
UserInput::Image { image_url: img2 },
@@ -218,7 +218,10 @@ mod tests {
TurnItem::UserMessage(user) => {
let expected_content = vec![
UserInput::Image { image_url },
UserInput::Text { text: user_text },
UserInput::Text {
text: user_text,
text_elements: Vec::new(),
},
];
assert_eq!(user.content, expected_content);
}
@@ -255,7 +258,10 @@ mod tests {
TurnItem::UserMessage(user) => {
let expected_content = vec![
UserInput::Image { image_url },
UserInput::Text { text: user_text },
UserInput::Text {
text: user_text,
text_elements: Vec::new(),
},
];
assert_eq!(user.content, expected_content);
}

View File

@@ -32,6 +32,7 @@ use crate::sandboxing::SandboxPermissions;
use crate::spawn::StdioPolicy;
use crate::spawn::spawn_child_async;
use crate::text_encoding::bytes_to_string_smart;
use codex_utils_pty::process_group::kill_child_process_group;
pub const DEFAULT_EXEC_COMMAND_TIMEOUT_MS: u64 = 10_000;
@@ -46,6 +47,12 @@ const EXEC_TIMEOUT_EXIT_CODE: i32 = 124; // conventional timeout exit code
const READ_CHUNK_SIZE: usize = 8192; // bytes per read
const AGGREGATE_BUFFER_INITIAL_CAPACITY: usize = 8 * 1024; // 8 KiB
/// Hard cap on bytes retained from exec stdout/stderr/aggregated output.
///
/// This mirrors unified exec's output cap so a single runaway command cannot
/// OOM the process by dumping huge amounts of data to stdout/stderr.
const EXEC_OUTPUT_MAX_BYTES: usize = 1024 * 1024; // 1 MiB
/// Limit the number of ExecCommandOutputDelta events emitted per exec call.
/// Aggregation still collects full output; only the live event stream is capped.
pub(crate) const MAX_EXEC_OUTPUT_DELTAS_PER_CALL: usize = 10_000;
@@ -289,18 +296,32 @@ async fn exec_windows_sandbox(
};
let exit_status = synthetic_exit_status(capture.exit_code);
let mut stdout_text = capture.stdout;
if stdout_text.len() > EXEC_OUTPUT_MAX_BYTES {
stdout_text.truncate(EXEC_OUTPUT_MAX_BYTES);
}
let mut stderr_text = capture.stderr;
if stderr_text.len() > EXEC_OUTPUT_MAX_BYTES {
stderr_text.truncate(EXEC_OUTPUT_MAX_BYTES);
}
let stdout = StreamOutput {
text: capture.stdout,
text: stdout_text,
truncated_after_lines: None,
};
let stderr = StreamOutput {
text: capture.stderr,
text: stderr_text,
truncated_after_lines: None,
};
// Best-effort aggregate: stdout then stderr
let mut aggregated = Vec::with_capacity(stdout.text.len() + stderr.text.len());
append_all(&mut aggregated, &stdout.text);
append_all(&mut aggregated, &stderr.text);
// Best-effort aggregate: stdout then stderr (capped).
let mut aggregated = Vec::with_capacity(
stdout
.text
.len()
.saturating_add(stderr.text.len())
.min(EXEC_OUTPUT_MAX_BYTES),
);
append_capped(&mut aggregated, &stdout.text, EXEC_OUTPUT_MAX_BYTES);
append_capped(&mut aggregated, &stderr.text, EXEC_OUTPUT_MAX_BYTES);
let aggregated_output = StreamOutput {
text: aggregated,
truncated_after_lines: None,
@@ -489,8 +510,13 @@ impl StreamOutput<Vec<u8>> {
}
#[inline]
fn append_all(dst: &mut Vec<u8>, src: &[u8]) {
dst.extend_from_slice(src);
fn append_capped(dst: &mut Vec<u8>, src: &[u8], max_bytes: usize) {
if dst.len() >= max_bytes {
return;
}
let remaining = max_bytes.saturating_sub(dst.len());
let take = remaining.min(src.len());
dst.extend_from_slice(&src[..take]);
}
#[derive(Clone, Debug)]
@@ -583,19 +609,15 @@ async fn consume_truncated_output(
))
})?;
let (agg_tx, agg_rx) = async_channel::unbounded::<Vec<u8>>();
let stdout_handle = tokio::spawn(read_capped(
BufReader::new(stdout_reader),
stdout_stream.clone(),
false,
Some(agg_tx.clone()),
));
let stderr_handle = tokio::spawn(read_capped(
BufReader::new(stderr_reader),
stdout_stream.clone(),
true,
Some(agg_tx.clone()),
));
let (exit_status, timed_out) = tokio::select! {
@@ -661,15 +683,18 @@ async fn consume_truncated_output(
Duration::from_millis(IO_DRAIN_TIMEOUT_MS),
)
.await?;
drop(agg_tx);
let mut combined_buf = Vec::with_capacity(AGGREGATE_BUFFER_INITIAL_CAPACITY);
while let Ok(chunk) = agg_rx.recv().await {
append_all(&mut combined_buf, &chunk);
}
// Best-effort aggregate: stdout then stderr (capped).
let mut aggregated = Vec::with_capacity(
stdout
.text
.len()
.saturating_add(stderr.text.len())
.min(EXEC_OUTPUT_MAX_BYTES),
);
append_capped(&mut aggregated, &stdout.text, EXEC_OUTPUT_MAX_BYTES);
append_capped(&mut aggregated, &stderr.text, EXEC_OUTPUT_MAX_BYTES * 2);
let aggregated_output = StreamOutput {
text: combined_buf,
text: aggregated,
truncated_after_lines: None,
};
@@ -686,14 +711,11 @@ async fn read_capped<R: AsyncRead + Unpin + Send + 'static>(
mut reader: R,
stream: Option<StdoutStream>,
is_stderr: bool,
aggregate_tx: Option<Sender<Vec<u8>>>,
) -> io::Result<StreamOutput<Vec<u8>>> {
let mut buf = Vec::with_capacity(AGGREGATE_BUFFER_INITIAL_CAPACITY);
let mut buf = Vec::with_capacity(AGGREGATE_BUFFER_INITIAL_CAPACITY.min(EXEC_OUTPUT_MAX_BYTES));
let mut tmp = [0u8; READ_CHUNK_SIZE];
let mut emitted_deltas: usize = 0;
// No caps: append all bytes
loop {
let n = reader.read(&mut tmp).await?;
if n == 0 {
@@ -722,11 +744,7 @@ async fn read_capped<R: AsyncRead + Unpin + Send + 'static>(
emitted_deltas += 1;
}
if let Some(tx) = &aggregate_tx {
let _ = tx.send(tmp[..n].to_vec()).await;
}
append_all(&mut buf, &tmp[..n]);
append_capped(&mut buf, &tmp[..n], EXEC_OUTPUT_MAX_BYTES);
// Continue reading to EOF to avoid back-pressure
}
@@ -750,42 +768,11 @@ fn synthetic_exit_status(code: i32) -> ExitStatus {
std::process::ExitStatus::from_raw(code as u32)
}
#[cfg(unix)]
fn kill_child_process_group(child: &mut Child) -> io::Result<()> {
use std::io::ErrorKind;
if let Some(pid) = child.id() {
let pid = pid as libc::pid_t;
let pgid = unsafe { libc::getpgid(pid) };
if pgid == -1 {
let err = std::io::Error::last_os_error();
if err.kind() != ErrorKind::NotFound {
return Err(err);
}
return Ok(());
}
let result = unsafe { libc::killpg(pgid, libc::SIGKILL) };
if result == -1 {
let err = std::io::Error::last_os_error();
if err.kind() != ErrorKind::NotFound {
return Err(err);
}
}
}
Ok(())
}
#[cfg(not(unix))]
fn kill_child_process_group(_: &mut Child) -> io::Result<()> {
Ok(())
}
#[cfg(test)]
mod tests {
use super::*;
use std::time::Duration;
use tokio::io::AsyncWriteExt;
fn make_exec_output(
exit_code: i32,
@@ -847,6 +834,18 @@ mod tests {
));
}
#[tokio::test]
async fn read_capped_limits_retained_bytes() {
let (mut writer, reader) = tokio::io::duplex(1024);
let bytes = vec![b'a'; EXEC_OUTPUT_MAX_BYTES.saturating_add(128 * 1024)];
tokio::spawn(async move {
writer.write_all(&bytes).await.expect("write");
});
let out = read_capped(reader, None, false).await.expect("read");
assert_eq!(out.text.len(), EXEC_OUTPUT_MAX_BYTES);
}
#[cfg(unix)]
#[test]
fn sandbox_detection_flags_sigsys_exit_code() {

Some files were not shown because too many files have changed in this diff Show More