## What changed
- Drop failed websocket connections immediately after a terminal stream
error instead of awaiting a graceful close handshake before forwarding
the error to the caller.
- Keep the success path and the closed-connection guard behavior
unchanged.
## Why this fixes the flake
- The failing integration test waits for the second websocket stream to
surface the model error before issuing a follow-up request.
- On slower runners, the old error path awaited
`ws_stream.close().await` before sending the error downstream. If that
close handshake stalled, the test kept waiting for an error that had
already happened server-side and nextest timed it out.
- Dropping the failed websocket immediately makes the terminal error
observable right away and marks the session closed so the next request
reconnects cleanly instead of depending on a best-effort close
handshake.
## Code or test?
- This is a production logic fix in `codex-api`. The existing websocket
integration test already exercises the regression path.
- include the requested sub-agent model and reasoning effort in the
spawn begin event\n- render that metadata next to the spawned agent name
and role in the TUI transcript
---------
Co-authored-by: Codex <noreply@openai.com>
Summary
- update the code-mode handler, runner, instructions, and error text to
refer to the `exec` tool name everywhere that used to say `code_mode`
- ensure generated documentation strings and tool specs describe `exec`
and rely on the shared `PUBLIC_TOOL_NAME`
- refresh the suite tests so they invoke `exec` instead of the old name
Testing
- Not run (not requested)
## Summary
- update the guardian prompting
- clarify the guardian rejection message so an action may still proceed
if the user explicitly approves it after being informed of the risk
## Testing
- cargo run on selected examples
## Summary
- add `skill_approval` to `RejectConfig` and the app-server v2
`AskForApproval::Reject` payload so skill-script prompts can be
configured independently from sandbox and rule-based prompts
- update Unix shell escalation to reject prompts based on the actual
decision source, keeping prefix rules tied to `rules`, unmatched command
fallbacks tied to `sandbox_approval`, and skill scripts tied to
`skill_approval`
- regenerate the affected protocol/config schemas and expand
unit/integration coverage for the new flag and skill approval behavior
Pass more params to /compact. This should give us parity with the
/responses endpoint to improve caching.
I'm torn about the MCP await. Blocking will give us parity but it seems
like we explicitly don't block on MCPs. Happy either way
Summary
- document how code-mode can import `output_text`/`output_image` and
ensure `add_content` stays compatible
- add a synthetic `@openai/code_mode` module that appends content items
and validates inputs
- cover the new behavior with integration tests for structured text and
image outputs
Testing
- Not run (not requested)
- clarify the `close_agent` tool description so it nudges models to
close agents they no longer need
- keep the change scoped to the tool spec text only
Co-authored-by: Codex <noreply@openai.com>
Summary
- document that `@openai/code_mode` exposes
`set_max_output_tokens_per_exec_call` and that `code_mode` truncates the
final Rust-side output when the budget is exceeded
- enforce the configured budget in the Rust tool runner, reusing
truncation helpers so text-only outputs follow the unified-exec wrapper
and mixed outputs still fit within the limit
- ensure the new behavior is covered by a code-mode integration test and
string spec update
Testing
- Not run (not requested)
Summary
- drop `McpToolOutput` in favor of `CallToolResult`, moving its helpers
to keep MCP tooling focused on the final result shape
- wire the new schema definitions through code mode, context, handlers,
and spec modules so MCP tools serialize the exact output shape expected
by the model
- extend code mode tests to cover multiple MCP call scenarios and ensure
the serialized data matches the new schema
- refresh JS runner helpers and protocol models alongside the schema
changes
Testing
- Not run (not requested)
- add `model` and `reasoning_effort` to the `spawn_agent` schema so the
values pass through
- validate requested models against `model.model` and only check that
the selected model supports the requested reasoning effort
---------
Co-authored-by: Codex <noreply@openai.com>
## Summary
- add ARC monitor support for MCP tool calls by serializing MCP approval
requests into the ARC action shape and sending the relevant
conversation/policy context to the `/api/codex/safety/arc` endpoint
- route ARC outcomes back into MCP approval flow so `ask-user` falls
back to a user prompt and `steer-model` blocks the tool call, with
guardian/ARC tests covering the new request shape
- update the TUI approval copy from “Approve Once” to “Allow” / “Allow
for this session” and refresh the related
snapshots
---------
Co-authored-by: Fouad Matin <fouad@openai.com>
Co-authored-by: Fouad Matin <169186268+fouad-openai@users.noreply.github.com>
Summary
- document output types for the various tool handlers and registry so
the API exposes richer descriptions
- update unified execution helpers and client tests to align with the
new output metadata
- clean up unused helpers across tool dispatch paths
Testing
- Not run (not requested)
There are some bug investigations that currently require us to ask users
for their user ID even though they've already uploaded logs and session
details via `/feedback`. This frustrates users and increases the time
for diagnosis.
This PR includes the ChatGPT user ID in the metadata uploaded for
`/feedback` (both the TUI and app-server).
Replace the Unix shell lookup path in `codex-rs/core/src/shell.rs` to
use
`libc::getpwuid_r()` instead of `libc::getpwuid()` when resolving the
current
user's shell.
Why:
- `getpwuid()` can return pointers into libc-managed shared storage
- on the musl static Linux build, concurrent callers can race on that
storage
- this matches the crash pattern reported in tmux/Linux sessions with
parallel
shell activity
Refs:
- Fixes#13842
Addresses #13586
This doesn't affect our CI scripts. It was user-reported.
Summary
- add `wiremock::ResponseTemplate` and `body_string_contains` imports
behind `#[cfg(not(debug_assertions))]` in
`codex-rs/core/tests/suite/view_image.rs` so release builds only pull
the helpers they actually use
## Summary
- update the unified exec test to use truncated_output() instead of the
removed output field
- fix the compile failure on latest main after ExecCommandToolOutput
changed shape
## What changed
- The retry test now uses the same streaming SSE test server used by
production-style tests instead of a wiremock sequence.
- The fixture is resolved via `find_resource!`, and the test asserts
that exactly two outbound requests were sent.
## Why this fixes the flake
- The old wiremock sequence approximated early-close behavior, but it
did not reproduce the same streaming semantics the real client sees.
- That meant the retry path depended on mock implementation details
instead of on the actual transport behavior we care about.
- Switching to the streaming SSE helper makes the test exercise the real
early-close/retry contract, and counting requests directly verifies that
we retried exactly once rather than merely hoping the sequence aligned.
## Scope
- Test-only change.
- collect input/output transcript deltas into active handoff transcript
state
- attach and clear that transcript on each handoff, and regenerate
schema/tests
### Purpose
While trying to build out CLI-Tools for the agent to use under skills we
have found that those tools sometimes need to invoke a user elicitation.
These elicitations are handled out of band of the codex app-server but
need to indicate to the exec manager that the command running is not
going to progress on the usual timeout horizon.
### Example
Model calls universal exec:
`$ download-credit-card-history --start-date 2026-01-19 --end-date
2026-02-19 > credit_history.jsonl`
download-cred-card-history might hit a hosted/preauthenticated service
to fetch data. That service might decide that the request requires an
end user approval the access to the personal data. It should be able to
signal to the running thread that the command in question is blocked on
user elicitation. In that case we want the exec to continue, but the
timeout to not expire on the tool call, essentially freezing time until
the user approves or rejects the command at which point the tool would
signal the app-server to decrement the outstanding elicitation count.
Now timeouts would proceed as normal.
### What's Added
- New v2 RPC methods:
- thread/increment_elicitation
- thread/decrement_elicitation
- Protocol updates in:
- codex-rs/app-server-protocol/src/protocol/common.rs
- codex-rs/app-server-protocol/src/protocol/v2.rs
- App-server handlers wired in:
- codex-rs/app-server/src/codex_message_processor.rs
### Behavior
- Counter starts at 0 per thread.
- increment atomically increases the counter.
- decrement atomically decreases the counter; decrement at 0 returns
invalid request.
- Transition rules:
- 0 -> 1: broadcast pause state, pausing all active stopwatches
immediately.
- \>0 -> >0: remain paused.
- 1 -> 0: broadcast unpause state, resuming stopwatches.
- Core thread/session logic:
- codex-rs/core/src/codex_thread.rs
- codex-rs/core/src/codex.rs
- codex-rs/core/src/mcp_connection_manager.rs
### Exec-server stopwatch integration
- Added centralized stopwatch tracking/controller:
- codex-rs/exec-server/src/posix/stopwatch_controller.rs
- Hooked pause/unpause broadcast handling + stopwatch registration:
- codex-rs/exec-server/src/posix/mcp.rs
- codex-rs/exec-server/src/posix/stopwatch.rs
- codex-rs/exec-server/src/posix.rs
Summary
- relocate truncation logic for exec command output into the new
`ExecCommandToolOutput` response helper instead of centralized handler
code
- update all affected tools and unified exec handling to use the new
response item structure and eliminate `Function(FunctionToolOutput)`
responses
- adjust context, registry, and handler interfaces to align with the new
response semantics and error fields
Testing
- Not run (not requested)
**Summary**
- allow `code_mode` to pass enabled tools metadata to the runner and
expose them via `tools.js`
- import tools inside JavaScript rather than relying only on globals or
proxies for nested tool calls
- update specs, docs, and tests to exercise the new bridge and explain
the tooling changes
**Testing**
- Not run (not requested)
(Experimental)
This PR adds a first MVP for hooks, with SessionStart and Stop
The core design is:
- hooks live in a dedicated engine under codex-rs/hooks
- each hook type has its own event-specific file
- hook execution is synchronous and blocks normal turn progression while
running
- matching hooks run in parallel, then their results are aggregated into
a normalized HookRunSummary
On the AppServer side, hooks are exposed as operational metadata rather
than transcript-native items:
- new live notifications: hook/started, hook/completed
- persisted/replayed hook results live on Turn.hookRuns
- we intentionally did not add hook-specific ThreadItem variants
Hooks messages are not persisted, they remain ephemeral. The context
changes they add are (they get appended to the user's prompt)
## Summary
- preserve unknown `:special_path` tokens, including nested entries, so
older Codex builds warn and ignore instead of failing config load
- fail closed with a startup warning when a permissions profile has
missing or empty filesystem entries instead of aborting profile
compilation
- normalize Windows verbatim paths like `\?\C:\...` before absolute-path
validation while keeping explicit errors for truly invalid paths
## Testing
- just fmt
- cargo test -p codex-core permissions_profiles_allow
- cargo test -p codex-core
normalize_absolute_path_for_platform_simplifies_windows_verbatim_paths
- cargo test -p codex-protocol
unknown_special_paths_are_ignored_by_legacy_bridge
- cargo clippy -p codex-core -p codex-protocol --all-targets -- -D
warnings
- cargo clean
## Summary
This is a fast follow to the initial `[permissions]` structure.
- keep the new split-policy carveout behavior for narrower non-write
entries under broader writable roots
- preserve legacy `WorkspaceWrite` semantics by using a cwd-aware bridge
that drops only redundant nested readable roots when projecting from
`SandboxPolicy`
- route the legacy macOS seatbelt adapter through that same legacy
bridge so redundant nested readable roots do not become read-only
carveouts on macOS
- derive the legacy bridge for `command_exec` using the sandbox root cwd
rather than the request cwd so policy derivation matches later sandbox
enforcement
- add regression coverage for the legacy macOS nested-readable-root case
## Examples
### Legacy `workspace-write` on macOS
A legacy `workspace-write` policy can redundantly list a nested readable
root under an already-writable workspace root.
For example, legacy config can effectively mean:
- workspace root (`.` / `cwd`) is writable
- `docs/` is also listed in `readable_roots`
The new shared split-policy helper intentionally treats a narrower
non-write entry under a broader writable root as a carveout for real
`[permissions]` configs. Without this fast follow, the unchanged macOS
seatbelt legacy adapter could project that legacy shape into a
`FileSystemSandboxPolicy` that treated `docs/` like a read-only carveout
under the writable workspace root. In practice, legacy callers on macOS
could unexpectedly lose write access inside `docs/`, even though that
path was writable before the `[permissions]` migration work.
This change fixes that by routing the legacy seatbelt path through the
cwd-aware legacy bridge, so:
- legacy `workspace-write` keeps `docs/` writable when `docs/` was only
a redundant readable root
- explicit `[permissions]` entries like `'.' = 'write'` and `'docs' =
'read'` still make `docs/` read-only, which is the new intended
split-policy behavior
### Legacy `command_exec` with a subdirectory cwd
`command_exec` can run a command from a request cwd that is narrower
than the sandbox root cwd.
For example:
- sandbox root cwd is `/repo`
- request cwd is `/repo/subdir`
- legacy policy is still `workspace-write` rooted at `/repo`
Before this fast follow, `command_exec` derived the legacy bridge using
the request cwd, but the sandbox was later built using the sandbox root
cwd. That mismatch could miss redundant legacy readable roots during
projection and accidentally reintroduce read-only carveouts for paths
that should still be writable under the legacy model.
This change fixes that by deriving the legacy bridge with the same
sandbox root cwd that sandbox enforcement later uses.
## Verification
- `just fmt`
- `cargo test -p codex-core
seatbelt_legacy_workspace_write_nested_readable_root_stays_writable`
- `cargo test -p codex-core test_sandbox_config_parsing`
- `cargo clippy -p codex-core -p codex-app-server --all-targets -- -D
warnings`
- `cargo clean`
## Summary
We need to support allowing request_permissions calls when using
`Reject` policy
<img width="1133" height="588" alt="Screenshot 2026-03-09 at 12 06
40 PM"
src="https://github.com/user-attachments/assets/a8df987f-c225-4866-b8ab-5590960daec5"
/>
Note that this is a backwards-incompatible change for Reject policy. I'm
not sure if we need to add a default based on our current use/setup
## Testing
- [x] Added tests
- [x] Tested locally
## Summary
The apply_patch tool should also respect AdditionalPermissions
## Testing
- [x] Added unit tests
---------
Co-authored-by: Codex <noreply@openai.com>
This changes the web_search tool spec in codex-core to use dedicated
Responses-API payload structs instead of shared config types and custom
serializers.
Previously, `ToolSpec::WebSearch` stored `WebSearchFilters` and
`WebSearchUserLocation` directly and relied on hand-written serializers
to shape the outgoing JSON. This worked, but it mixed config/schema
types with the OpenAI Responses payload contract and created an easy
place for drift if those shared types changed later.
### Why
This keeps the boundary clearer:
- app-server/config/schema types stay focused on config
- Responses tool payload types stay focused on the OpenAI wire format
It also makes the serialization behavior obvious from the structs
themselves, instead of hiding it in custom serializer functions.
## Summary
request_permissions flows should support persisting results for the
session.
Open Question: Still deciding if we need within-turn approvals - this
adds complexity but I could see it being useful
## Testing
- [x] Updated unit tests
---------
Co-authored-by: Codex <noreply@openai.com>
## What changed
- The RMCP streamable HTTP tests now wait for metadata and tool
readiness before issuing tool calls.
- OAuth state is isolated per test home.
- The helper server startup path now uses bounded bind retries so
transient `AddrInUse` collisions do not fail the test immediately.
## Why this fixes the flake
- The old tests could begin issuing tool requests before the helper
server had finished advertising its metadata and tools, so the first
request sometimes raced the server startup sequence.
- On top of that, shared OAuth state and occasional bind collisions on
CI runners introduced cross-test environmental noise unrelated to the
functionality under test.
- Readiness polling makes the client wait for an observable “server is
ready” signal, while isolated state and bounded bind retries remove
external contention that was causing intermittent failures.
## Scope
- Test-only change.
This cleans up a bunch of metric plumbing that had started to drift.
The main change is making `codex-otel` the canonical home for shared
metric definitions and metric tag helpers. I moved the `turn/thread`
metric names that were still duplicated into the OTEL metric registry,
added a shared `metrics::tags` module for common tag keys and session
tag construction, and updated `SessionTelemetry` to build its metadata
tags through that shared path.
On the codex-core side, TTFT/TTFM now use the shared metric-name
constants instead of local string definitions. I also switched the
obvious remaining turn/thread metric callsites over to the shared
constants, and added a small helper so TTFT/TTFM can attach an optional
sanitized client.name tag from TurnContext.
This should make follow-on telemetry work less ad hoc:
- one canonical place for metric names
- one canonical place for common metric tag keys/builders
- less duplication between `codex-core` and `codex-otel`
add `plugin/uninstall` app-server endpoint to fully rm plugin from
plugins cache dir and rm entry from user config file.
plugin-enablement is session-scoped, so uninstalls are only picked up in
new sessions (like installs).
added tests.
## Summary
Alternative to #14061 - we need to use a child process on windows to
correctly validate Powershell behavior.
## Testing
- [x] These are tests
## What changed
- add a bounded `resume_until_initial_messages` helper in
`core/tests/suite/resume.rs`
- retry the resume call until `initial_messages` contains the fully
persisted final turn shape before asserting
## Why this fixes flakiness
The old test resumed once immediately after `TurnComplete` and sometimes
read rollout state before the final turn had been persisted. That made
the assertion race persistence timing instead of checking the resumed
message shape. The new helper polls for up to two seconds in 10ms steps
and only asserts once the expected message sequence is actually present,
so the test waits for the real readiness condition instead of depending
on a lucky timing window.
## Scope
- test-only
- no production logic change
## Summary
- align the guardian permission test with the actual sandbox policy it
widens and use a slightly larger Windows-only timeout budget
- expose the additional-permissions normalization helper to the guardian
test module
- replace the guardian popup snapshot assertion with targeted string
assertions
## Why this fixes the flake
This group was carrying two separate sources of drift. The guardian core
test widened derived sandbox policies without updating the source
sandbox policy, and it used a Windows command/timeout combination that
was too tight on slower runners. Separately, the TUI test was
snapshotting the full popup even though unrelated feature text changes
were the only thing moving. The new assertions keep coverage on the
guardian entry itself while removing unrelated snapshot churn.
pakrym-oaiandMichael Bolin
Ahmed IbrahimandMichael Bolin
maja-openaiandMichael Bolin
Celia ChenandMichael Bolin
Rasmus RygaardandMichael Bolin
Leo ShimonakaandMichael Bolin
gabec-openaiandMichael Bolin
Won ParkandMichael Bolin
xl-openaiandMichael Bolin
Eric TrautandMichael Bolin
Channing CongerandGitHub
Dylan HurdandGitHub
Andrei EternalandGitHub
viyatb-oaiandGitHub
Owen LinandGitHub
sayan-oaiandGitHub