## What changed
- Move assessment event construction, metrics, and analytics tracking into `codex-guardian-reviewer` through `ReviewReport`.
- Store denial accounting in thread extension data through `ReviewDenials`, with core retaining turn interruption and lifecycle cleanup.
- Move failed-review record selection and bounded serialization into the extension, with core supplying captured review context.
## Testing
Add coverage that denial accounting clears on turn completion and interruption. Move the oversized-record test into the extension and exercise the new API.
GitOrigin-RevId: 6e5b3d9d4128b356f99b1d0e74a68111e40261ab
## What changed
Introduce `ExecutedToolCalls` to own optional shared recorder state and centralize feature checks, call recording, and accepted result source lookup. Update session, tool runtime, MCP, and Code Mode callers to use this interface.
Move prompt attachment, retry handling, and metadata budgeting into the `request_metadata` module, preserving existing recording and request behavior.
## Testing
Add coverage for all session and turn feature combinations, lazy result source lookup, and unchanged prompts when recording is disabled. Retain the existing metadata budget and cell completeness tests in the new module.
GitOrigin-RevId: bc19f34317232b6599c7ec71c15db4a80123ff6f
## Why
Git root discovery is optional metadata work, but filesystem probes can block. They should not exhaust Tokio's blocking pool, delay runtime shutdown, or prevent later turns from observing repository changes.
## What changed
- Add a shared `GitRootDiscovery` service that coalesces concurrent lookups for the same working directory and limits probes across directories.
- Run probes on detached threads, retain in-flight work across caller cancellation, and discard completed results instead of caching them.
- Use the service for turn and memory metadata enrichment, abort unused turn enrichment when its state is dropped, and limit memory metadata waits to one second.
## Testing
Add coverage for probe sharing, capacity limits, cancellation, fresh discovery, runtime shutdown, memory timeouts, and repositories restored after startup prewarming.
GitOrigin-RevId: bca46fc263e7a12a2f69146d8a0b3e7c7e0846cb
## What changed
- Import skill metadata and mention helpers directly from `codex-skills`.
- Import host skill loading types directly from `codex-skills-extension`.
- Make the remaining `codex-core` skills module private and remove its skill API re-exports, including the `HostSkillsService` re-export from `codex-core-api`.
GitOrigin-RevId: 26bf8fffd980df9c4d0c308d7e983ceab9a5df82
## Why
MCP tool handlers and their Code Mode definitions are immutable for the lifetime
of a stable MCP binding, so rebuilding them for every sampling step repeats schema
construction unnecessarily.
## What changed
- Cache MCP handlers per session and reuse them while the same binding remains
active, invalidating the cache when the binding changes.
- Lazily cache schema-free Code Mode definitions on MCP handlers while continuing
to rebuild definitions for step-local tools whose specifications may change.
- Reapply current app enablement, tool policy, exposure, and plugin schema budgets
whenever cached handlers are registered for a step.
## Testing
- Cover handler and Code Mode definition reuse, binding invalidation, changing app
policy, stable continuation schemas, and step-local custom tool updates.
GitOrigin-RevId: 69c5011fb97c08580ac476e7fb2286b1e92e0359
## What changed
- Move host skill discovery, snapshot caching, and configuration handling from
`codex-core-skills` to `codex-skills-extension`.
- Rename `SkillsService` and `SkillsLoadInput` to `HostSkillsService` and
`HostSkillsLoadInput` to distinguish the host implementation from other skill
providers.
- Keep shared loading primitives and skill outcome modeling in
`codex-core-skills`.
GitOrigin-RevId: d81a21791d2ef8c066e157e7e538b8cb7ee4c24b
## Why
App-server clients need to advertise structured MCP extension settings, including supported MCP App UI MIME types, rather than only opting into OpenAI form elicitation with a boolean.
## What changed
- Add an `extensions` map to initialize capabilities and preserve `mcpServerOpenaiFormElicitation` as a legacy alias for `openai/form`.
- Capture the declared extension profile when a thread is started, resumed, or forked, propagate it to subagents, and advertise it to downstream MCP servers during initialization.
- Keep the profile stable for the lifetime of the loaded session instead of changing it on later turns or direct tool calls.
## Testing
- Cover extension conversion, downstream MCP initialization, session isolation, legacy form support, and subagent inheritance.
GitOrigin-RevId: fbcedbb74ce788e574b0f884a4c45c4cedb9de54
## What changed
- Add the opt-in `executed_tool_call_metadata` feature to record model-attempted direct and code-mode tool calls.
- Attach recorded names and arguments to the matching tool output in the next Responses request, preserving the metadata across sampling retries.
- Bound pending calls and argument sizes, and emit truncation metadata when limits are exceeded.
## Testing
- Cover disabled-by-default behavior, namespaced and nested calls, blocked or failed attempts, retry handling, and recorder limits.
GitOrigin-RevId: 5ce917b73e797b0f6904d46477b9d2cf8ea71bf2
## What changed
- Store a request-logging-disabled `RouteAwareClientPool` in session services and use it for file creation, blob upload, and finalization requests.
- Preserve the existing system-root fallback for transport-default proxy routes, and emit a warning event when that fallback is used.
- Remove the direct `reqwest` dependency from `codex-api` now that uploads use the shared HTTP abstraction.
## Testing
- Add coverage for completing multiple uploads through a shared client pool.
GitOrigin-RevId: 01842415cad2d349a6bd3869abb3332e8110ebce
## What changed
- Make `McpRuntime` own the published MCP configuration, connections, elicitation routing, and selected capability roots for a thread.
- Capture immutable MCP bindings for model steps and tool calls so in-flight work keeps a consistent connection set and approval authority while refreshed state is published atomically.
- Mark MCP state dirty when relevant configuration, plugins, environments, authentication, or elicitation capabilities change, then rebuild it before the next sampling request or out-of-band MCP operation.
- Separate config reloads from server invalidation: `ReloadMcpConfig` applies resolved MCP inputs, while `RefreshMcpServers` requests reinitialization from the thread's latest state.
## Testing
- Cover refreshed state visibility for existing turns, stable step bindings, current approval authority, cancelled refresh retries, resource-client reconciliation, and Apps recovery between sampling requests.
GitOrigin-RevId: 59eabb1aa8dc083426bd18ef4d3630508f376401
## What changed
- Use `McpConnectionSet` as the primary name for the collection of active MCP
server connections throughout `codex-mcp` and its callers.
- Keep `McpConnectionManager` as a backward-compatible type alias.
GitOrigin-RevId: 9753838b4fd0a0b6429a9c4af0c077137616f388
## What changed
- Change `ReviewDecision::Denied` to carry a rejection string and update the generated protocol schemas.
- Preserve specific rejection reasons through command, patch, network, MCP, delegated, and automatic approval flows so tool results can return them to the model.
- Distinguish invalid approval responses from user declines and truncate oversized rejection messages before adding them to model context.
## Testing
- Add coverage for denial serialization, invalid app-server approval responses, dropped automatic reviews, network rejection propagation, and rejection-message truncation.
GitOrigin-RevId: d48a55643aa03c7ecfa08a8e4b669913eec5464a
## What changed
- Add a thread-owned `McpRuntime` that atomically publishes connection-set
replacements and owns shutdown.
- Route `McpResourceClient` snapshots and cache keys through that runtime so
extension clients observe refreshed MCP servers.
- Keep model-step configuration snapshots separate from ownership of live MCP
connections.
## Testing
- Add coverage that captures an extension's session resource client, refreshes
the MCP server configuration, and verifies that the existing client sees the
newly published server.
GitOrigin-RevId: 351b2e25594b8eb6b91e50ff8d06b7d49768ae1c
## What changed
- Let deferred environments provide selected capability roots with their ready signal.
- Validate that those roots have unique, non-empty IDs, belong to the registering environment, and stay within the root limit.
- Include roots from ready turn environments when resolving MCP contributions, and refresh the MCP runtime when the selected root set changes.
- Expose the exact ready root set to MCP contributors so executor plugins become available with their environment.
## Testing
- Cover ready-root propagation, validation failures, replacement isolation, reconnection, and MCP plugin availability refresh.
GitOrigin-RevId: ec3498aab1164824025094e96a9b1063b7b731ad
## Why
Code-mode tool results could return to the model while an MCP
elicitation was still waiting for user input. This differed from
parallel tool calling and could let the model continue before the user
resolved the request.
We need one session-level view of outstanding elicitations so tool
runtimes can consistently hold results until every pending elicitation
is resolved.
## What changed
- Added a counted, session-owned ElicitationService with RAII
registrations.
- Registered both core-originated and server-originated MCP elicitations
with the service.
- Migrated out-of-band elicitation tracking and unified exec timeout
pausing to the shared service.
- Made code-mode functions.exec and functions.wait capture their runtime
result normally, then hold it before returning while an elicitation is
outstanding.
- Kept terminate: true immediate; only its result is held.
- Preserved model-visible wall time across the elicitation hold.
- Kept the behavior session-scoped, with concurrent elicitations holding
the pause until all registrations are released.
## Why
Generic plugin guidance is currently emitted only with initial context
from host plugin state. An executor-selected plugin can become available
later in the same turn, making its skills and tools usable without ever
telling the model how plugin capabilities should be used.
## What
- project every ready selected plugin package, including skill-only
plugins
- carry plugin availability with the exact MCP runtime projection while
preserving MCP manager reuse when servers and connectors are unchanged
- move generic plugin guidance from the static initial-context path into
persisted World State
- recognize legacy and retained plugin fragments so resume and
compaction do not duplicate guidance
## Testing
- `just test -p codex-mcp-extension`
- `just test -p codex-core plugins_instructions`
- `just test -p codex-core
plugin_availability_change_reuses_the_mcp_manager`
- `just test -p codex-app-server --test all selected_capabilit`
## Why
Selected plugin metadata is stable, but MCP processes are live runtime
state. They need different lifetimes:
- the MCP extension caches manifest, MCP, and connector declarations for
each stable selected root;
- each model step projects that cached metadata through the roots that
resolved as ready for that exact step;
- the MCP manager is rebuilt only when that availability projection
changes.
This matches executor skills: both features consume the same resolved
step roots instead of inferring readiness from the turn's selected
environments.
## Behavior
```text
E1 not ready for this step
-> no E1 MCP servers or connectors
-> cached plugin metadata stays in ext/mcp
E1 becomes ready
-> reuse cached metadata
-> publish one MCP runtime containing E1 capabilities
same ready roots on the next step
-> reuse the exact runtime; no rediscovery and no MCP restart
resume
-> create new extension thread state and a new MCP runtime
```
All model-facing consumers use the same step snapshot:
```text
resolved selected roots
|
v
extension MCP/connector projection
|
v
{ MCP config, connector snapshot, MCP manager }
|
+-> advertise model tools
+-> build app/connector tools
+-> execute MCP calls
```
## Cache contract
The existing MCP extension owns a cache keyed by the full
`SelectedCapabilityRoot`:
```rust
let state = thread_store.get_or_init(SelectedExecutorPluginMcpState::default);
```
The cache lives with extension thread state. Environment availability
filters projection but does not invalidate metadata. Resume creates new
thread state. There is no file watcher or executor generation because
contents behind a stable environment/root are assumed stable.
## What changes
- Keeps executor plugin discovery and cached metadata in `ext/mcp`.
- Caches MCP and connector declarations together per selected root.
- Uses the step's already-resolved capability roots, including lazy
environments that are not turn environments.
- Reuses the current MCP runtime when the ready-root projection is
unchanged.
- Uses the same step MCP manager and connector snapshot for
model-visible tools and execution.
- Resolves direct thread-scoped MCP requests from the current
selected-root projection.
## Deliberately out of scope
- `app/list` remains based on the latest global host-plugin state; this
PR does not make its response or notifications thread-specific.
- `required = true` startup semantics do not apply to delayed executor
MCP activation.
- No filesystem/content invalidation.
- No transport-disconnect watcher.
- No executor generations or environment replacement semantics.
- No client sharing across complete manager replacements.
## Stack
1. Extension-owned World State sections.
2. Project executor skills through World State.
3. Pin one MCP runtime to each model step.
4. **This PR:** project selected MCP and connector state from
extension-owned metadata.
5. Integration coverage for selected capability availability and resume.
## Verification
-
`selected_plugin_servers_use_managed_requirements_for_the_selected_root_id`
- The stacked integration PR covers unavailable to ready activation,
unchanged-runtime reuse, skills, MCP tools, connector attribution, and
cold resume.
## Why
An MCP refresh can replace the session's current manager while a model
step is still running. The step must execute calls through the same
manager whose tools it advertised.
## Boundary
```text
current session MCP runtime
|
| capture once for this model step
v
StepContext.mcp
- exact MCP config
- exact connection manager
- exact runtime environment context
```
```rust
pub struct McpRuntimeSnapshot {
config: Arc<McpConfig>,
manager: Arc<McpConnectionManager>,
runtime_context: McpRuntimeContext,
}
```
## Example
```text
step A captures runtime A and advertises A's tools
refresh publishes runtime B
step A tool call -> runtime A
next step -> runtime B
```
Capturing the snapshot is only an `Arc` clone. It does not restart MCPs
or make an RPC.
## What changes
- Captures one MCP runtime in `StepContext`.
- Uses it for tool planning, tool calls, resources, approvals, connector
attribution, and elicitation.
- Publishes replacement runtimes atomically.
- Lets an old runtime live only while an in-flight step or request still
holds its `Arc`.
Most of this diff is mechanical routing from the session-global manager
to `step_context.mcp`; it does not introduce selected-plugin discovery
yet.
## What does not change
- No plugin or extension migration.
- No new MCP cache policy.
- No environment file watching.
- No client sharing between separate managers.
## Stack
1. Extension-owned World State sections.
2. Project executor skills through World State.
3. **This PR:** pin one MCP runtime to each model step.
4. Project selected MCP/app/connector metadata by environment
availability.
5. One end-to-end integration scenario.
## Why
`selectedCapabilityRoots` is durable thread intent: “use this capability
root from environment `worker`.”
The important product assumption is:
> One environment ID always names the same logical executor and stable
contents.
`worker` does not silently change from executor A to an unrelated
executor B. The process-local connection handle for `worker` can still
be replaced while Codex is running, though, for example when
`environment/add` registers a fresh handle for the same logical
environment.
The thread should persist only the stable selection. Each model step
should pair that selection with the exact ready handle captured for that
step.
## The boundary
```text
persisted thread intent
plugin@1 -> environment "worker"
|
| capture the current step
v
model-step view
unavailable, or
plugin@1 + worker's exact captured ready handle
```
The environment ID is the stable identity and cache key. The
`Arc<Environment>` is only a process-local handle retained so consumers
of one model step use the same captured environment. It is never
persisted and it does not imply different environment contents.
## What changes
### Persist the stable selection
Selected roots are written into `SessionMeta` and restored with the
thread. Forked subagents inherit the same selections, including
bounded-history forks.
Only stable data is persisted: root ID, environment ID, and root path.
### Capture readiness together with the exact handle
The environment snapshot records:
```rust
environment_id -> Some(Arc<Environment>) // ready in this step
environment_id -> None // still starting in this step
```
This prevents readiness and execution from coming from different
registry snapshots.
For example:
```text
step snapshot: worker -> handle A, ready
environment/add: worker -> fresh handle B for the same logical environment
current step: plugin@1 still uses captured handle A
```
Without carrying handle A in the snapshot, the resolver could combine “A
was ready” with handle B and treat B as ready before it had finished
starting.
This does not change cache invalidation. Stable capability metadata
remains identified by environment ID and capability root. Replacing a
process-local handle under the same stable environment ID does not
invalidate or rediscover that metadata.
### Resolve availability per model step
- A ready captured environment produces resolved roots using its
captured handle.
- A starting, missing, or failed environment is omitted from that step.
- A selected lazy environment that is outside the turn's captured
environment set is asked to start, and a later step can observe it as
ready.
- No capability files are scanned here.
Transient transport disconnects remain the remote client's reconnect
concern. This PR models initial attachment/readiness; it does not add
live socket-connectivity state.
## Example
```text
thread selection: plugin@1 -> environment "worker"
step 1: worker is starting -> plugin@1 unavailable
step 2: worker is ready -> plugin@1 resolves through worker's captured handle
step 3: fresh local handle -> current step remains pinned; a later step captures its own view
```
Temporary unavailability does not discard the durable selection. Later
PRs can retain stable metadata caches while projecting only currently
available capabilities into model-visible World State.
## Compatibility
The app-server request shape does not change. Older rollouts without
`selected_capability_roots` deserialize to an empty list.
## Stack
1. **This PR:** persist stable selected roots and resolve them through
an exact model-step handle.
2. #29960: cache stable skill metadata and project available skills into
World State.
3. #29946: cache stable plugin declarations and manage the separate live
MCP runtime.
## Why
With deferred executors, a turn can begin before a remote environment
attaches. AGENTS.md discovery previously ran only during session setup,
so instructions from a later environment never reached the model or the
session instruction sources.
WorldState persistence has now landed, so this uses the durable
model-visible baseline directly instead of carrying a temporary
resume/fork compatibility path.
## What
- Add an `AgentsMdManager` in `SessionServices` to own host
instructions, loaded state, and refresh caching.
- When `DeferredExecutor` is enabled, refresh AGENTS.md when attached
environment selections change and freeze the result in the corresponding
`StepContext`.
- Represent AGENTS.md as a persisted WorldState section for every
session, with bounded initial, replacement, and removal updates.
- Remove duplicate AGENTS.md state and rendering from
`SessionConfiguration` and `TurnContext`.
- Build initial context, per-request updates, and compaction context
from the same step-scoped value.
- On resume and fork, compare current instructions with the restored
WorldState baseline and inject a replacement exactly once when they
differ.
Builds on #29833, #29835, and #29837.
## Tests
- Covers a remote environment becoming ready mid-turn, with AGENTS.md
appearing on the next request exactly once and updating canonical
instruction sources.
- Covers full, unchanged, replaced, and removed AGENTS.md WorldState
rendering.
- Covers changed instructions across cold resume and fork without
duplicate reinjection.
- Covers remote-v2 compaction retaining creation-time instructions in
the live session and cold resume appending one replacement when the
source changed.
- Ran focused `codex-core` AGENTS.md, WorldState, and context-update
test suites.
Stacked on #28822.
## Summary
- add a host-injectable current-time provider with a built-in system
implementation
- record UTC developer reminders in history immediately before due model
requests
- keep cadence state per session and force a refresh after compaction
This does NOT include the app server client <-> server clock logic. This
PR is only for the reminder message & system clock that will be used in
prod.
## Testing
- `just test -p codex-core varlatency_`
- `just clippy -p codex-core -p codex-app-server -p codex-mcp-server -p
codex-thread-manager-sample`
- `just fmt`
## Why
Host skill discovery was still exposed as a manager even though it is a
process-owned service shared by sessions, the app-server catalog, and
file-watcher invalidation. The skills extension also consumed an ad hoc
loaded-skills wrapper instead of a named immutable snapshot.
## What changed
- replace `SkillsManager` with concrete `SkillsService`
- make the service cache and return immutable `HostSkillsSnapshot`
values
- migrate the skills extension host provider to the snapshot boundary
- migrate app-server catalog, watcher, and invalidation paths to the
service
This keeps the service limited to host discovery, caching, roots, and
invalidation. Catalog rendering and invocation remain extension
responsibilities for the next stacked change.
## Why
Shell snapshots are currently session-scoped even though shell and cwd
are properties of a selected turn environment. That makes snapshot
refresh depend on separate session-cwd plumbing, prevents retained
environments from retaining their snapshot work, and can make snapshot
construction use a different shell than command execution.
This follows #27955 by making the retained thread-environment service
own environment snapshot lifecycles. Session configuration remains the
requested selection state, while `ThreadEnvironments` remains the source
of successfully resolved environments.
## What changed
- Configure the shell-snapshot builder before initial environment
resolution.
- Start each local environment snapshot task when its `TurnEnvironment`
is built and retain that shared task while environment ID and cwd still
match.
- Inherit retained environment snapshots into spawned child threads.
- Carry the selected `TurnEnvironment` through shell runtimes so
snapshot construction and command execution use the same
environment-specific shell and cwd.
- Load project instructions and warm plugins/skills after initial
environment resolution.
- Continue decoding invalid UTF-8 instruction files lossily without
emitting a startup warning.
- Keep requested selections in `SessionConfiguration`; failed or
duplicate resolutions only affect the resolved environment snapshot.
## Validation
- `cargo check -p codex-core --tests`
- `just test -p codex-home instructions` (6 passed)
- Focused environment, instruction, shell-snapshot, and user-shell tests
(84 passed)
- Focused shell-snapshot, user-shell, and unified-exec tests (126
passed; two event-timing tests passed on retry)
## Why
Selected execution environments are thread-scoped resources, but startup
and turn construction repeatedly resolved their IDs and working
directories. That discarded existing environment handles and shell
metadata even when a selection had not changed.
Session configuration updates also need to affect future turns without
changing the resolved environment set already captured by a running
turn.
## What changed
- Create a `ThreadEnvironments` service inside `Codex` from the spawned
`EnvironmentManager` and raw environment selections, then store it on
`SessionServices`.
- Split service construction from `update_selections`, allowing session
configuration updates to mutate the resolved set in place.
- Retain an existing `TurnEnvironment` when its environment ID and
working directory match; resolve only added or changed selections and
remove selections that are no longer present.
- Normalize duplicate IDs by keeping the first selection and skip
individual selections that fail to resolve instead of rejecting the
entire update.
- Give each `TurnContext` a cloned `TurnEnvironmentSnapshot`, so later
session configuration updates affect future turns without rewriting an
active turn.
- Reuse the service-owned environment manager and resolved snapshot for
startup work, MCP initialization, and child-thread spawning instead of
flowing resolved environments through spawn arguments.
## Test plan
- `cargo check -p codex-core --tests`
- `just test -p codex-core environment_selection`
- `just test -p codex-core turn_environments`
- `just test -p codex-core
session_update_settings_does_not_rewrite_sticky_environment_cwds`
- `just test -p codex-core
default_turn_does_not_overlay_legacy_fallback_cwd_onto_stored_thread_environments`
## Why
Tool router construction rebuilds the deferred-tool BM25 index during
session initialization and before each sampling continuation, even when
the searchable tool metadata is unchanged. Local profiling measured
`append_tool_search_executor` at roughly 113 ms per continuation, making
repeated index construction the largest measured router-building cost.
## What changed
- Add a session-scoped `ToolSearchHandlerCache` so continuations and
user turns can reuse the existing handler.
- Key reuse on the complete ordered `Vec<ToolSearchInfo>`, rebuilding
when searchable text, loadable tool specs, source metadata, or ordering
changes.
- Build handlers outside the cache lock and recheck before publishing
them, avoiding holding the mutex during index construction.
## Verification
- `cache_reuses_identical_search_infos_and_rebuilds_changed_inputs`
covers exact cache reuse and invalidation when the ordered search
metadata changes.
- Local rollout profiling showed the initial router build populating the
cache and unchanged later continuations reusing it:
- uncached: 118 ms median across 14 spans from 3 rollouts
- cached: 4 ms median across 12 spans from 3 rollouts
## Why
Shell snapshot lifecycle state was split between `Shell` and
`SessionServices`: `Shell` carried the receiver while session code
exposed and forwarded the raw sender. That coupled shell identity to
mutable snapshot state and made refresh, inheritance, and file lifetime
harder to reason about.
## What changed
- make each `Arc<ShellSnapshot>` represent one cwd-specific snapshot
generation
- store the active generation in `SessionServices` with `ArcSwapOption`
- have construction start the background build and expose only a
cwd-validated snapshot path
- use `ShellSnapshotFile` ownership to delete snapshot files
automatically
- pass snapshot paths explicitly to shell runtimes instead of storing
snapshot state on `Shell`
- preserve inherited and in-flight generations by pinning their `Arc`
while they are in use
## Test plan
- `cargo check -p codex-core --lib`
- `just test -p codex-core 'shell_snapshot::tests'`
- `just test -p codex-core
shell_command_snapshot_still_intercepts_apply_patch`
- `just test -p codex-core
shell_snapshot_deleted_after_shutdown_with_skills`
## Why
`selectedCapabilityRoots` belongs to one thread, but MCP contributors
previously received only the global Codex config. That left no clean way
for a selected executor capability to contribute MCP servers to its own
thread.
## What this PR does
- Gives MCP contributors a small context containing the config and, for
a running thread, its frozen host-seeded inputs.
- Uses the same thread inputs during startup, status queries, refreshes,
and skill dependency checks.
- Keeps threadless MCP operations and the existing hosted Apps behavior
unchanged.
- Adds coverage showing that two threads resolve independent
registrations and that later lifecycle mutations do not change the
frozen MCP inputs.
This PR does not discover plugin manifests, add MCP servers, or launch
anything new. It only establishes the thread-scoped registration
boundary.
## Follow-ups
- Resolve selected executor plugin roots through their owning
environment filesystem.
- Convert their stdio MCP declarations into environment-bound
registrations and add an executor MCP end-to-end test.
## Verification
- `just fmt`
- `cargo check --tests -p codex-protocol -p codex-extension-api -p
codex-mcp-extension -p codex-core -p codex-app-server`
Tests and Clippy were not run.
## Why
#27198 made the extension-owned `codex_apps` MCP connection the hosted
plugin runtime, but its `mcp/skill` resources still bypassed the skills
extension. App-server could list and read those resources through
generic MCP APIs, but a thread with no selected environment did not
expose them in the model's skills catalog or load their `SKILL.md`
through `$skill`.
Hosted skills should stay remote while using the same typed catalog,
source authority, deduplication, bounded contextual catalog, and
selected-skill prompt injection as host and executor skills. They should
not be downloaded or exposed as ambient filesystem paths.
## What changed
- Add a session-scoped `McpResourceClient` over the replaceable MCP
connection manager so resource list/read calls follow startup and
refresh replacements.
- Add a `BackendSkillProvider` that pages `codex_apps` resources,
accepts bounded and validated `mcp/skill` entries, and reads a selected
skill's `SKILL.md` through the same MCP connection.
- Register the remote provider in app-server and include it in the
skills catalog even when a thread has no selected capability roots or
executor.
- Contribute hosted skill metadata through the bounded
`AvailableSkillsInstructions` developer-context path, exclude remote
entries from per-turn catalog injection, and classify `<skills>`
messages as contextual developer content so rollback can trim and
rebuild them correctly.
## Testing
- Extend the app-server MCP resource integration test with
`environments: []` to exercise two-page discovery, filter a
non-`mcp/skill` resource, verify the escaped developer catalog entry and
user-role `<skill>` fragment containing the fetched `SKILL.md`, and
preserve generic MCP resource reads.
- Add core event-mapping coverage that classifies `<skills>` developer
messages as contextual history.
## Summary
We originally addressed startup prewarming holding the read side of
`RwLock<McpConnectionManager>` by snapshotting tool-list state. Review
feedback identified the broader ownership problem: the outer
synchronization should only publish or retrieve the current manager,
while MCP operations rely on the manager's internal synchronization. A
follow-up preserved operation retirement with a separate gate, but
further review questioned whether that synchronization was actually
required and whether we could support latest-wins replacement instead.
This PR now stores the current MCP manager in `ArcSwap`. Each operation
uses `load_full()` to obtain an owned `Arc<McpConnectionManager>`, then
performs MCP I/O without retaining the publication mechanism. Refresh
cancels obsolete startup work, constructs a replacement, and atomically
publishes it. New operations see the latest manager, while operations
that already loaded the previous manager retain a valid handle. Refresh
happens at a turn boundary, so there should be no active user tool calls
to drain.
Git history supports dropping the outer `RwLock`. It was introduced in
`03ffe4d595` on November 17, 2025 for non-blocking MCP startup: the
session published an empty manager, startup initialized that same object
while holding the write lock, and readers waited for initialization.
`7cd2e84026` on February 19, 2026 removed that two-phase initialization
in favor of constructing a fresh manager and swapping it in, explicitly
noting that `Option` or `OnceCell` could replace the placeholder design.
Hot reload later reused the existing lock to publish a replacement, but
I found no indication that the lock was introduced to guarantee
in-flight tool calls finish before refresh or shutdown.
Terminal shutdown remains separate from refresh: it aborts startup
prewarming and active tasks before shutting down the current manager, so
tool calls may be interrupted and no model WebSocket work continues
after shutdown. Focused regression coverage exercises pending tool-list
cancellation, deferred refresh, and startup-prewarm shutdown.
## Why
Required MCP server startup was enforced in `Session::new` after
`McpConnectionManager` had already created the clients. That split let
other manager construction paths bypass the same requirement and exposed
manager internals solely so the session could validate them. Keeping
required-server readiness in the constructor gives every caller one
consistent startup contract.
## What changed
- make `McpConnectionManager::new` return `anyhow::Result<Self>` and
fail when an enabled, required server cannot initialize
- pass the startup cancellation token into the constructor so
required-server waits remain cancellable
- propagate constructor failures through resource reads, connector
discovery, and MCP status collection
- preserve the active manager and cancellation token when a refreshed
replacement fails
- keep required-startup failure collection private and cover the
constructor error contract directly
## Validation
- updated the focused connection-manager test to assert the complete
required-server startup error
- local tests not run; relying on CI
## Why
Once a named permission profile is selected, runtime state has to keep
that profile identity intact instead of collapsing back to anonymous
effective permissions. The session refresh path also needs to rebuild
profile-derived network proxy state so active profile switches take
effect consistently.
## What changed
- Preserve the active permission profile through session updates.
- Rebuild profile-derived runtime/network configuration when the active
profile changes.
- Keep the runtime path aligned with the current session configuration
APIs.
- Tighten the affected tests, including the Windows delete-pending
memory-file case that was intermittently tripping CI.
## Stack
1. **This PR**: runtime/session/network propagation for active
permission profiles.
2. [#23708](https://github.com/openai/codex/pull/23708): TUI selection
plumbing and guardrail flow.
3. [#21559](https://github.com/openai/codex/pull/21559): profile-aware
`/permissions` menu and custom profile display.
<img width="1296" height="906" alt="image"
src="https://github.com/user-attachments/assets/077fa3a7-80cb-4925-80b1-d2395018d90a"
/>
## Why
[#21736](https://github.com/openai/codex/pull/21736) introduces the
typed extension API, but the runtime does not yet carry a registry
through thread/session startup or give contributors host-owned stores to
read from. This PR wires that host-side path so later feature migrations
can move product-specific behavior behind typed contributions without
adding another bespoke seam directly to `codex-core`.
## What changed
- Thread `ExtensionRegistry<Config>` through `ThreadManager`,
`CodexSpawnArgs`, `Session`, and sub-agent spawn paths.
- Wire `ThreadStartContributor` and `ContextContributor`
- Expose the small supporting surface needed by non-core callers that
construct threads directly, including `empty_extension_registry()`
through `codex-core-api`.
This PR lands the host plumbing only: the app-server registry is still
empty, and concrete feature migrations are intended to follow
separately.
## Why
PR #21460 reverted the earlier move of skills change watching from
`codex-core` into app-server. This reapplies that boundary change so
app-server owns client-facing `skills/changed` notifications and core no
longer carries the watcher.
## What
- Restore the app-server `SkillsWatcher` and register it from thread
listener setup.
- Remove the core-owned skills watcher and its core live-reload
integration surface.
- Restore app-server coverage for `skills/changed` notifications after a
watched skill file changes.
## Validation
- `cargo test -p codex-app-server --test all
suite::v2::skills_list::skills_changed_notification_is_emitted_after_skill_change
-- --exact --nocapture`
- `cargo test -p codex-core --lib --no-run`
## Summary
TL;DR: teaches `codex-rs` / app-server to request a desktop-provided
attestation token and attach it as `x-oai-attestation` on the scoped
ChatGPT Codex request paths.
## Details
This PR teaches the Codex app-server runtime how to request and attach
an attestation token. It does not generate DeviceCheck tokens directly;
instead, it relies on the connected desktop app to advertise that it can
generate attestation and then asks that app for a fresh header value
when needed.
The flow is:
1. The Codex desktop app connects to app-server.
2. During `initialize`, the app can advertise that it supports
`requestAttestation`.
3. Before app-server calls selected ChatGPT Codex endpoints, it sends
the internal server request `attestation/generate` to the app.
4. app-server receives a pre-encoded header value back.
5. app-server forwards that value as `x-oai-attestation` on the scoped
outbound requests.
The code in this repo is mostly protocol and runtime plumbing: it adds
the app-server request/response shape, introduces an attestation
provider in core, wires that provider into Responses / compaction /
realtime setup paths, and covers the intended scoping with tests. The
signed macOS DeviceCheck generation remains owned by the desktop app PR.
## Related PR
- Codex desktop app implementation:
https://github.com/openai/openai/pull/878649
## Validation
<details>
<summary>Tests run</summary>
```sh
cargo test -p codex-app-server-protocol
cargo test -p codex-core attestation --lib
cargo test -p codex-app-server --lib attestation
```
Also ran:
```sh
just fix -p codex-core
just fix -p codex-app-server
just fix -p codex-app-server-protocol
just fmt
just write-app-server-schema
```
</details>
<details>
<summary>E2E DeviceCheck validation</summary>
First validated the signed desktop app boundary directly: launched a
packaged signed `Codex.app`, sent `attestation/generate`, decoded the
returned `v1.` attestation header, and validated the extracted
DeviceCheck token with `personal/jm/verify_devicecheck_token.py` using
bundle ID `com.openai.codex`. Apple returned `status_code: 200` and
`is_ok: true`.
Then ran the fuller app + app-server flow. The packaged `Codex.app`
launched a current-branch app-server via `CODEX_CLI_PATH`, and a local
MITM proxy intercepted outbound `chatgpt.com` traffic. The app-server
requested `attestation/generate` from the real Electron app process, and
the intercepted `/backend-api/codex/responses` traffic included
`x-oai-attestation` on both routes:
```text
GET /backend-api/codex/responses Upgrade: websocket x-oai-attestation: present
POST /backend-api/codex/responses Upgrade: none x-oai-attestation: present
```
The captured header decoded to a DeviceCheck token that also validated
with Apple for `com.openai.codex` (`status_code: 200`, `is_ok: true`,
team `2DC432GLL2`).
</details>
---------
Co-authored-by: Codex <noreply@openai.com>
## Why
Skills update notifications are app-server API behavior, but the watcher
lived in `codex-core` and surfaced through
`EventMsg::SkillsUpdateAvailable`. Moving the watcher out keeps core
focused on thread execution and lets app-server own both cache
invalidation and the `skills/changed` notification.
## What changed
- Added an app-server-owned skills watcher that watches local skill
roots, clears the shared skills cache, and emits `skills/changed`
directly.
- Registers skill watches from the common app-server thread listener
attach path, including direct starts, resumes, and app-server-observed
child or forked threads.
- Stores the `WatchRegistration` on `ThreadState`, so listener
replacement, thread teardown, idle unload, and app-server shutdown
deregister by dropping the RAII guard.
- Removed `EventMsg::SkillsUpdateAvailable`, the core watcher, and the
old core live-reload test.
- Extended the app-server skills change test to verify a cached skills
list is refreshed after a filesystem change without forcing reload.
## Validation
- `cargo check -p codex-core -p codex-app-server -p codex-mcp-server -p
codex-rollout -p codex-rollout-trace`
- `cargo test -p codex-app-server
skills_changed_notification_is_emitted_after_skill_change`
## Why
After `hooks/list` exposes the hook inventory, clients need a way to
persist user hook preferences, make those changes effective in
already-open sessions, and distinguish user-controllable hooks from
managed requirements without adding another bespoke app-server write
API.
## What
- Extends `hooks/list` entries with effective `enabled` state.
- Persists user-level hook state under `hooks.state.<hook-id>` so the
model can grow beyond a single boolean over time.
- Uses the existing `config/batchWrite` path for hook state updates
instead of introducing a dedicated hook write RPC.
- Refreshes live session hook engines after config writes so
already-open threads observe updated enablement without a restart.
## Stack
1. openai/codex#19705
2. openai/codex#19778
3. This PR - openai/codex#19840
4. openai/codex#19882
## Reviewer Notes
The generated schema files account for much of the raw diff. The core
behavior is in:
- `hooks/src/config_rules.rs`, which resolves per-hook user state from
the config layer stack.
- `hooks/src/engine/discovery.rs`, which projects effective enablement
into `hooks/list` from source-derived managedness.
- `config/src/hook_config.rs`, which defines the new `hooks.state`
representation.
- `core/src/session/mod.rs`, which rebuilds live hook state after user
config reloads.
---------
Co-authored-by: Codex <noreply@openai.com>
## Why
`codex-models-manager` had grown to own provider-specific concerns:
constructing OpenAI-compatible `/models` requests, resolving provider
auth, emitting request telemetry, and deciding how provider catalogs
should be sourced. That made the manager harder to reuse for providers
whose model catalog is not fetched from the OpenAI `/models` endpoint,
such as Amazon Bedrock.
This change moves provider-specific model discovery behind
provider-owned implementations, so the models manager can focus on
refresh policy, cache behavior, picker ordering, and model metadata
merging.
## What Changed
- Introduced a `ModelsManager` trait with separate `OpenAiModelsManager`
and `StaticModelsManager` implementations.
- Added `ModelsEndpointClient` so OpenAI-compatible HTTP fetching lives
outside `codex-models-manager`.
- Moved `/models` request construction, provider auth resolution,
timeout handling, and request telemetry into `codex-model-provider` via
`OpenAiModelsEndpoint`.
- Added provider-owned `models_manager(...)` construction so configured
OpenAI-compatible providers use `OpenAiModelsManager`, while
static/catalog-backed providers can return `StaticModelsManager`.
- Added an Amazon Bedrock static model catalog for the GPT OSS Bedrock
model IDs.
- Updated core/session/thread manager code and tests to depend on
`Arc<dyn ModelsManager>`.
- Moved offline model test helpers into
`codex_models_manager::test_support`.
## Metadata References
The Bedrock catalog metadata is based on the official Amazon Bedrock
OpenAI model documentation:
- [Amazon Bedrock OpenAI
models](https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-openai.html)
lists the Bedrock model IDs, text input/output modalities, and `128,000`
token context window for `gpt-oss-20b` and `gpt-oss-120b`.
- [Amazon Bedrock `gpt-oss-120b` model
card](https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-openai-gpt-oss-120b.html)
lists the `bedrock-runtime` model ID `openai.gpt-oss-120b-1:0`, the
`bedrock-mantle` model ID `openai.gpt-oss-120b`, text-only modalities,
and `128K` context window.
- [OpenAI `gpt-oss-120b` model
docs](https://developers.openai.com/api/docs/models/gpt-oss-120b)
document configurable reasoning effort with `low`, `medium`, and `high`,
plus text input/output modality.
The display names, default reasoning effort, and priority ordering are
Codex-local catalog choices.
## Test Plan
- Manually verified app-server model listing with an AWS profile:
```shell
CODEX_HOME="$(mktemp -d)" cargo run -p codex-app-server-test-client -- \
--codex-bin ./target/debug/codex \
-c 'model_provider="amazon-bedrock"' \
-c 'model_providers.amazon-bedrock.aws.profile="codex-bedrock"' \
-c 'model_providers.amazon-bedrock.aws.region="us-west-2"' \
model-list
```
The response returned the Bedrock catalog with `openai.gpt-oss-120b-1:0`
as the default model and `openai.gpt-oss-20b-1:0` as the second listed
model, both text-only and supporting low/medium/high reasoning effort.
## Summary
Adds the debug CLI entry point for reducing recorded rollout traces.
This gives developers a direct way to inspect whether the emitted trace
stream reduces into the expected conversation/runtime model.
## Stack
This is PR 5/5 in the rollout trace stack.
- [#18876](https://github.com/openai/codex/pull/18876): Add rollout
trace crate
- [#18877](https://github.com/openai/codex/pull/18877): Record core
session rollout traces
- [#18878](https://github.com/openai/codex/pull/18878): Trace tool and
code-mode boundaries
- [#18879](https://github.com/openai/codex/pull/18879): Trace sessions
and multi-agent edges
- [#18880](https://github.com/openai/codex/pull/18880): Add debug trace
reduction command
## Review Notes
This PR is intentionally last: it depends on the trace crate, core
recorder, runtime/tool events, and session/agent edge data all existing.
The command should remain a debug/developer tool and avoid adding new
runtime behavior.
The useful review question is whether the CLI exposes the reducer in the
smallest practical way for local inspection without turning the debug
command into a supported user-facing workflow.
Begin migrating the thread write codepaths to ThreadStore.
This starts using ThreadStore inside of core session code, not only in
the app server code.
Rework the interfaces around thread recording/persistence. We're left
with the following:
* `ThreadManager`: owns the process-level registry of loaded threads and
handles cross-thread orchestration: start, resume, fork, lookup, remove,
and route ops to running CodexThreads.
* `CodexThread`: represents one loaded/running thread from the outside.
It is the handle app-server and callers use to submit ops, inspect
session metadata, and shut the thread down.
* `LiveThread`: session-owned persistence lifecycle handle for one
active thread. Core session code uses it to append rollout items,
materialize lazy persistence, flush, shutdown, discard init-failed
writers, and load that thread’s persisted history.
* `ThreadStore`: storage backend abstraction. It answers “how are
threads persisted, read, listed, updated, archived?” Local and remote
implementations live behind this trait.
* `LocalThreadStore`: local ThreadStore implementation. It owns the
file/sqlite-specific details and keeps RolloutRecorder as a local
implementation detail.
This is a few too many Thread abstractions for my liking, but they do
all represent different concepts / needs / layers.
Migration note: in places where the core code explicitly requires a
path, rather than a thread ID, throw an error if we're running with a
remote store.
Cover the new local live-writer lifecycle with focused tests and
preserve app-server thread-start behavior, including ephemeral pathless
sessions.
## Summary
Short circuit the convo if auto-review hits too many denials
## Testing
- [x] Added unit tests
---------
Co-authored-by: Codex <noreply@openai.com>
## Summary
Wires rollout trace recording into `codex-core` session and turn
execution. This records the core model request/response, compaction, and
session lifecycle boundaries needed for replay without yet tracing every
nested runtime/tool boundary.
## Stack
This is PR 2/5 in the rollout trace stack.
- [#18876](https://github.com/openai/codex/pull/18876): Add rollout
trace crate
- [#18877](https://github.com/openai/codex/pull/18877): Record core
session rollout traces
- [#18878](https://github.com/openai/codex/pull/18878): Trace tool and
code-mode boundaries
- [#18879](https://github.com/openai/codex/pull/18879): Trace sessions
and multi-agent edges
- [#18880](https://github.com/openai/codex/pull/18880): Add debug trace
reduction command
## Review Notes
This layer is the first live integration point. The important review
question is whether trace recording is isolated from normal session
behavior: trace failures should not become user-visible execution
failures, and recording should preserve the existing turn/session
lifecycle semantics.
The PR depends on the reducer/data model from the first stack entry and
only introduces the core recorder surface that later PRs use for richer
runtime and relationship events.
## Summary
This PR fully reverts the previously merged Agent Identity runtime
integration from the old stack:
https://github.com/openai/codex/pull/17387/changes
It removes the Codex-side task lifecycle wiring, rollout/session
persistence, feature flag plumbing, lazy `auth.json` mutation,
background task auth paths, and request callsite changes introduced by
that stack.
This leaves the repo in a clean pre-AgentIdentity integration state so
the follow-up PRs can reintroduce the pieces in smaller reviewable
layers.
## Stack
1. This PR: full revert
2. https://github.com/openai/codex/pull/18871: move Agent Identity
business logic into a crate
3. https://github.com/openai/codex/pull/18785: add explicit
AgentIdentity auth mode and startup task allocation
4. https://github.com/openai/codex/pull/18811: migrate auth callsites
through AuthProvider
## Testing
Tests: targeted Rust checks, cargo-shear, Bazel lock check, and CI.
Builds on top of #17659
Move the filesystem + sqlite thread listing-related operations inside of
a local ThreadStore implementation and call ThreadStore from the places
that used to perform these filesystem/sqlite operations.
This is the first of a series of PRs that will implement the rest of the
local ThreadStore.
Testing:
- added unit tests for the thread store implementation
- adjusted some unit tests in the realtime + personality packages whose
callsites changed. Specifically I'm trying to hide ThreadMetadata inside
of the local implementation and make ThreadMetadata a sqlite
implementation detail concern rather than a public interface, preferring
the more generate StoredThread interface instead
- added a corner case test for the personality migration package that
wasn't covered by the existing test suite
- adjust the behavior of searched thread listing to run the existing
local rollout repair/backfill pass _before_ querying SQLite results, so
callers using ThreadStore::list_threads do not miss matches after a
partial metadata warm-up
## Summary
Stack PR 2 of 4 for feature-gated agent identity support.
This PR adds agent identity registration behind
`features.use_agent_identity`. It keeps the app-server protocol
unchanged and starts registration after ChatGPT auth exists rather than
requiring a client restart.
## Stack
- PR1: https://github.com/openai/codex/pull/17385 - add
`features.use_agent_identity`
- PR2: https://github.com/openai/codex/pull/17386 - this PR
- PR3: https://github.com/openai/codex/pull/17387 - register agent tasks
when enabled
- PR4: https://github.com/openai/codex/pull/17388 - use `AgentAssertion`
downstream when enabled
## Validation
Covered as part of the local stack validation pass:
- `just fmt`
- `cargo test -p codex-core --lib agent_identity`
- `cargo test -p codex-core --lib agent_assertion`
- `cargo test -p codex-core --lib websocket_agent_task`
- `cargo test -p codex-api api_bridge`
- `cargo build -p codex-cli --bin codex`
## Notes
The full local app-server E2E path is still being debugged after PR
creation. The current branch stack is directionally ready for review
while that follow-up continues.
## Description
This PR introduces `review_id` as the stable identifier for guardian
reviews and exposes it in app-server `item/autoApprovalReview/started`
and `item/autoApprovalReview/completed` events.
Internally, guardian rejection state is now keyed by `review_id` instead
of the reviewed tool item ID. `target_item_id` is still included when a
review maps to a concrete thread item, but it is no longer overloaded as
the review lifecycle identifier.
## Motivation
We'd like to give users the ability to preempt a guardian review while
it's running (approve or decline).
However, we can't implement the API that allows the user to override a
running guardian review because we didn't have a unique `review_id` per
guardian review. Using `target_item_id` is not correct since:
- with execve reviews, there can be multiple execve calls (and therefore
guardian reviews) per shell command
- with network policy reviews, there is no target item ID
The PR that actually implements user overrides will use `review_id` as
the stable identifier.
jif
