## Why
Local administrators need to require managed configuration from a specific customer-hosted server before a session can start.
## What changed
- Add a local managed `config_server` requirement with an HTTPS base URL and optional OAuth settings. Validate endpoints and derive the bundle route as `codex/config/bundle` beneath the base URL.
- Replace the entire `config_server` block when a higher-priority requirements layer supplies one, keeping the endpoint and OAuth settings together. Reject this setting in fetched policy.
- Track the applied server through configuration snapshots, role overrides, and refreshes. Reject session startup and managed refreshes when the applied server does not match the requirement; user refreshes retain the session's managed configuration.
## Testing
Add coverage for missing, mismatched, and matching server configuration at session startup, whole-block replacement, rejection of server settings in fetched policy, and preservation across role overrides and refreshes.
GitOrigin-RevId: 51e8a1738dabdb117ae858120f6fb09ae2ff1261
## Why
Remote credential masking previously fell back to unmasked execution whenever custom credential providers were configured.
## What changed
- Transport configured provider rules to the executor-local network proxy, where credential values and destination context are resolved.
- Advertise `credentialMaskingCustomProviders` and negotiate masking support. Keep the warning and unmasked fallback for executors that support only built-in providers.
- Discard provider rules when masking is disabled and omit their contents from debug output.
## Testing
Extend coverage for capability compatibility, provider configuration round trips, and fallback on older executors. Exercise configured-provider masking in remote request history and verify that masked placeholders are restored to real credentials in upstream request headers.
GitOrigin-RevId: b036f56900b343ddccc17b5293c82bc1f3426237
## Why
Executor hooks depended on V1 discovery snapshots, so plugins discovered through V2 could not run their cleanup hooks.
## What changed
Prepare executor hook sources during step setup from either V1 or V2 discovery and capture them in `StepContext` for stop, interrupt, and memory consolidation hooks. Restrict V2 sources to selected plugin installations while preserving plugin enablement, hook admission, MCP routing, and app tool policy checks.
## Testing
Add unit coverage for V2 hook admission, environment routing, disabled plugins, and failed discovery. Add integration coverage showing that `Stop` and `Interrupt` cleanup hooks run with V2 discovery alone and remain silent when the plugin is disabled.
GitOrigin-RevId: a4c0e9c56bdebcdf22cea63f02f8fb4c5eed358e
## What changed
- Add `CloudClient` and integrate it with `AppServerClient` and `AppServerRequestHandle` for requests, notifications, server-request replies, events, and shutdown.
- Keep retained request and reply handles bound to their cloud backend, with event delivery independent of request completion and a timeout for shutdown.
- Return no platform or Codex home metadata for cloud clients, and reject TUI filesystem operations for this backend.
## Testing
Add tests for backend ownership of retained handles, independent event delivery, errors after the command receiver closes, and shutdown timeouts when the worker does not respond.
GitOrigin-RevId: 4cf308dc0ea6fa6e7b14dd4542d93e8030b0162d
## What changed
- Add `OAuthSessionRequest` to pair access tokens with their authorization identity and retain rejected-token recovery across transient failures and retries. New sign-ins receive an identity that survives token refresh.
- Expose explicit login, logout, and authorization identity lookup, plus a per-manager option to disable automatic browser login while allowing credential reuse and refresh.
- Move serializable `OAuthSessionConfig` into `codex-model-provider-info`, add an optional `device_authorization_url` field, and share endpoint URL validation through `codex-http-client`.
- Distinguish permission failures, transient token endpoint failures, and invalid responses using `io::ErrorKind`.
## Testing
Add coverage for sign-in identity preservation through refresh retries, removal of saved and cached credentials on logout, and suppression of automatic browser login. Extend token endpoint error classification coverage.
GitOrigin-RevId: ae9f72e787171e179fe139a3b840a9f7e1e46506
## What changed
Promote `bedrock_setup_wizard` to stable and enable it by default for eligible onboarding. Users can still disable it with `features.bedrock_setup_wizard = false`.
## Testing
Update onboarding coverage to check the enabled default and explicit opt-out. Remove the explicit opt-in from daemon startup tests so they exercise the default configuration.
GitOrigin-RevId: 5bd911088681a31d34aa444398e6d3f7d1d87955
## Why
Rust v0 symbols can mention `codex_code_mode_host` as the instantiating crate of a dependency's generic function, so matching the crate name alone can select the wrong function.
## What changed
Demangle `nm` output and match `codex_code_mode_host::main` exactly, preserving spaces within demangled names. Include local symbols in the Linux lookup by removing `--extern-only`.
GitOrigin-RevId: b4ea4face3fe861966d3643d13db168143bcefa3
## What changed
Add optional `responsesapiClientMetadata` to `ThreadSampleParams` and forward it to the sample turn's Responses API metadata. Update the experimental protocol schema.
## Testing
Extend the thread sample integration test to verify that client metadata appears in both the request header and body, preserves the explicit `turn_trigger` when client metadata supplies a conflicting value, and does not carry over to the next parent turn.
GitOrigin-RevId: eb1f80824d476fc6c584a2b35d97e59d51fd4273
## Why
QPACK can expand compact response headers into large decoded field sections. Bound their size before `h3` constructs its capacity-limited `HeaderMap`.
## What changed
Set the HTTP/3 client's maximum field section size to 64 KiB so oversized responses are rejected with `HeaderTooBig`.
## Testing
Add a regression test that sends QPACK static references around the size limit and a much larger expansion. Verify that oversized responses are rejected and a subsequent valid response succeeds on the same connection.
GitOrigin-RevId: d7c83a7a755dfafae05919ec52542ffce40a1ef4
## What changed
When AWS profiles or environment credentials are detected, move manual Bedrock API key entry under “Other AWS sign-in methods” and update its description to include API keys. Keep API key entry directly available when no credentials are detected.
## Testing
Update onboarding method assertions and rendering snapshots to cover the grouped options and API key entry in the other sign-in methods menu.
GitOrigin-RevId: 992ecf7a2e3ac9f2f0eef2ef6534cf9beb88ea8e
## Why
When `-c` or `--config` appears at multiple command levels, Clap retains only the deepest scope's values, dropping earlier overrides.
## What changed
Collect config overrides from each command scope in order while preserving typed child arguments and their existing precedence. Arguments forwarded after `--` remain unchanged.
## Testing
Add regression coverage for overrides split across nested `mcp add` commands, forwarded `-c` arguments, and validation of an invalid override before `app-server` when another override follows it.
GitOrigin-RevId: 40c64eda866497e2f59491a523c6777a8bfe5d5f
## What changed
Keep the leading `/` in the textarea and render slash commands after the normal composer prompt, so `/model` appears as `› /model`.
Track the full `/command` token as a text element and remove stale slash-command elements when the draft changes. Update composer tests and snapshots for the new representation.
GitOrigin-RevId: 60119f18b69cec16f4e21839f45132231feb84b4
## Why
`thread/sample` previously required the latest source turn to have completed successfully, rejecting failed or interrupted turns even when the source was idle.
## What changed
Accept the latest completed, failed, or interrupted turn in both legacy and paginated history modes. Update the protocol documentation and validation error to reflect the accepted statuses, while continuing to reject stale or active sources.
## Testing
Expand integration coverage across all three terminal statuses and both history modes, checking that sampling preserves context and leaves the source unchanged. Also verify that sampling an active source is rejected before interrupting it.
GitOrigin-RevId: a5dfdb2800868ee178eae16b12ee7f6f9515898c
Use `ConfigEditsBuilder` to persist a concurrent MCP server edit during OAuth discovery without launching another CLI process inside the HTTP timeout. Preserve coverage that the concurrent server and issuer-bound callback are saved, and assert that OAuth fails with the expected invalid authorization endpoint error.
GitOrigin-RevId: 41d782f9dbbc96d528e4990171bffd87e95f6e4f
## Why
Runtime process handles are recyclable and thread-local. Using them as terminal session identities can merge separate commands or associate `write_stdin` operations with the wrong exec.
## What changed
- Give each exec a distinct terminal session and track the latest session for each thread-local process handle.
- Record the exec selected by `write_stdin` at runtime and use it to correct session membership during replay, including when dispatch is delayed until a handle is reused.
- Fall back to the latest session for the thread and handle when replaying older traces without runtime target events.
## Testing
Extend reducer coverage for process ID reuse across threads and successive execs, stdin session membership, and delayed runtime target resolution.
GitOrigin-RevId: 330c6fd26a4eebdef03ae600a030dc76920e0dd9
## What changed
Add the `codex-direct-tool-timing` crate to record tool execution, automated reviews, human waits, and dependency paths. Human waits take precedence over overlapping reviews, and dependency waits with unavailable history are recorded as unknown time.
Invocation guards complete each call once, explicitly or on drop. Separate invocation identities prevent late review or wait guards from affecting calls that reuse an ID; duplicate active IDs disable ambiguous path attribution. Batch timing exposes the last-finishing invocation's path once all calls complete.
## Testing
Add unit tests for parallel batch completion, duplicate and reused call IDs, cancellation, missing dependency history, overlapping reviews and human waits, and nested wait guards.
GitOrigin-RevId: 8b251f51a0c52342657f2d244524b562a371bd8e
Pause the Tokio clock before spawning the review in
`run_review_on_reused_session_waits_for_submitted_turn` so its one-second
deadline uses simulated time while the test verifies that stale turn completions
are ignored.
GitOrigin-RevId: c6712219f894835b576f499eb080ec15a3f09a78
## Why
A total-duration assertion does not prove that two shell commands overlap and depends on execution speed.
## What changed
Update `shell_tools_run_in_parallel` to wait for both commands to create start markers before releasing them through a shared file. Fail if both commands do not start within five seconds, and verify that each command exits successfully and returns its expected output.
GitOrigin-RevId: ba322157a60951ad5dd2f1b018ddc6007de1de89
## Why
Credential substitution needs to authorize the destination and the original request before exposing host secrets. Project configuration must not redirect trusted credential policies, and older clients must reject scoped policies rather than silently apply unconditional injection.
## What changed
- Add hook `schemes` and `ports` constraints and an exact whole-header `placeholder` for secret substitution. Require `network.mitm.version = 2` for these options in permission profiles.
- Restrict placeholder substitution to HTTPS on port 443 and reject missing, mismatched, or duplicate placeholder headers.
- Reject project changes to trusted MITM definitions and their inheritance chains. Allow self-contained project header-stripping policies only when no trusted MITM policy exists.
- Reject ambiguous query values, value-constrained headers, and semicolon paths during authorization, and preserve case-sensitive extension methods.
- Mark resolved secret headers sensitive, reject secret injection into `User-Agent`, restrict substitution headers, and limit secret files to 64 KiB.
## Testing
Add regression coverage for scoped TOML resolution, legacy-format rejection, unknown fields, project policy restrictions, origin and placeholder denials, ambiguous request matching, secret redaction, and hook application after credential brokerage.
GitOrigin-RevId: 78dc115ab6c88c9a734377626429bc4840731f89
## Why
Enterprise-managed MCP authentication previously required public OAuth clients at both the identity provider and the MCP authorization server, preventing use of registrations that require client secrets.
## What changed
- Accept separate `client_secret` and `token_endpoint_auth_method` settings for the enterprise IdP and MCP server OAuth registration. Support `client_secret_basic` and `client_secret_post`, defaulting to Basic when a secret is configured.
- Validate the registered method against server metadata and use it for enterprise login and both ID-JAG exchanges.
- Require a configured `oauth.authorization_server_issuer` for confidential MCP registrations and reject empty secrets or authentication methods without secrets.
- Require the account OAuth client to match the configured IdP client before reusing an account grant with a provisioned secret; otherwise use dedicated enterprise login.
## Testing
Extend configuration, login, exchange, and app-server lifecycle tests to cover confidential registrations, separate IdP and resource credentials, authentication method validation, and exclusion of client secrets from authorization URLs, stored login grants, and logs.
GitOrigin-RevId: b3be378c2b1bb67e6a3652c648e72fd2e21e956d
## Why
HTTPS requests and WebSocket connections on Windows need to authenticate when a proxy rejects `CONNECT` with status `407`.
## What changed
- Retry proxy challenges through the existing Windows authentication helper, using the signed-in user's Negotiate or NTLM credentials. Treat an origin `407` as a normal HTTP response.
- Bridge authenticated tunnels to the HTTP client through destination-specific named pipes, retaining redirect checks and TLS backend fallback without caching those tunnels across destinations.
- Carry explicit CA certificates and client identities into HTTPS proxy TLS configuration, and apply the configured connect timeout when opening authenticated HTTP tunnels.
GitOrigin-RevId: ec7d9168972a7a640350e5348cfca40307390fbf
## What changed
Expose `connect_windows_proxy` for callers handling a CONNECT `407` response. Use libcurl with the signed-in Windows user's Negotiate or NTLM credentials, then return the authenticated connection with buffered data preserved.
Support HTTP and HTTPS proxies, with default or caller-provided TLS configuration. HTTPS proxies requiring authentication tokens bound to their TLS connection are unsupported.
Enable libcurl's NTLM support in Cargo and Bazel builds.
## Testing
Add a Windows-only test asserting that libcurl provides NTLM, SPNEGO, and SSPI support.
GitOrigin-RevId: 4ed7d2208766404b6125ce2f4b5f91de5616a37f
## Why
Callbacks from the shared stdio host need to retain each execution's trace context after it yields, without keeping the execution span open until its background cell closes.
## What changed
Capture a W3C trace-context snapshot for each execution and restore it as the parent of tool callback and notification spans. Instrument delegate invocation and its returned future so the context is available on entry and across awaits.
## Testing
Add an integration test with overlapping yielded cells completed in reverse order. Cover sampled, unsampled, and untraced callers, trace isolation, callback and notification parentage, and execution spans finishing while callbacks remain blocked.
GitOrigin-RevId: 57924123d2c2f31bcfaf725688d8bba775235fa0
## Why
Responses Lite includes retained reasoning in its reported token usage even without an `X-Reasoning-Included` header. Adding a separate reasoning estimate can trigger automatic compaction prematurely. Locally recomputed usage also already includes retained reasoning.
## What changed
Initialize reasoning accounting from `use_responses_lite` before each sampling or compaction request, while allowing server events to override it. Mark locally recomputed token usage as including reasoning.
## Testing
Update the compaction regression test to verify that Responses Lite usage without the header stays below an 81-token limit at 80 tokens and triggers compaction at 81 tokens.
GitOrigin-RevId: 06f9272cb3c259e927941f1d2a1d7096310ffb2c
## Why
Remote registry requests need to recover when credentials are rejected or rotated while `exec-server` is running. Concurrent requests should reuse refreshed credentials rather than refresh the replacement token again.
## What changed
- Add an `AuthProvider` recovery hook and retry registry registration, connection, and harness-key validation once after a `401` when the provider supplies replacement credentials.
- Use live ChatGPT authentication for remote `exec-server` and guard proactive and unauthorized refreshes so concurrent callers share the completed refresh.
- Allow executor credentials to retry when the parent has already supplied a valid replacement.
- Keep connection authentication, refresh, and retry within the original connection deadline.
## Testing
Add coverage for concurrent refreshes, stored and parent-supplied credential rotation, refresh failures, retry limits, and connection deadlines, plus a CLI integration test for refresh and credential reload in a running executor.
GitOrigin-RevId: 9ed533e410762c7ac8b7c0c30bb93fd20d96778b
## What changed
- Store `logs_2.sqlite` in `log_dir` when explicitly configured, unless a managed `sqlite_home` requirement applies. Keep other databases in `sqlite_home` and preserve the existing SQLite log location when `log_dir` is unset.
- Initialize and lock the selected log directory only when opening its database, so an unavailable directory does not prevent authentication commands from loading configuration. Share the cleanup lock across configuration clones and allow failed initialization to be retried.
- Leave existing logs in place. Read legacy logs for diagnostic submissions when the current query returns no logs, attaching them as `codex-logs-legacy.log`. Continue retention cleanup and page reclamation for both locations, with independent maintenance locks and filesystem checks.
## Testing
Add coverage for configuration routing and managed requirements, deferred initialization and lock lifetimes, authentication with an unavailable log directory, legacy log attachments, and cleanup of current and legacy databases.
GitOrigin-RevId: 48216694596d738eff1b3c5eef6c4e7ada6d0137
## What changed
Add `codex migrate-nfs-db` to replace configured SQLite WAL databases in place with validated copies in `DELETE` journal mode, retaining the originals and their sidecar files in backup directories. Stage conversion locally, check database integrity before installation, and support resuming an interrupted replacement.
Require confirmation and exclusive access, with all Codex instances stopped on every host until migration succeeds. The NFS mount must support cross-client POSIX locks. Reject unsupported file permissions and ACLs, and serialize migrations with a migration lock.
Support configuration overrides, `--profile`, and `--staging-dir`. Skip the logs database when `log_dir` is explicitly configured and leave databases already in `DELETE` mode unchanged. Reject remote mode.
GitOrigin-RevId: 7b48cc3a713113a3c09a2243e49b54258bef609d
## Why
`message_board_tools` always exposed all nine discussion tools. Consumers need a way to expose a subset, such as `read_thread` and `list_threads`, over the same board.
## What changed
Accept an ordered slice of the new `MessageBoardTool` enum and optional description overrides in `message_board_tools`. Use the enum for schemas and dispatch, and keep the agent extension's existing tool set and order through `MessageBoardTool::DISCUSSION`.
## Testing
Extend the local-board integration test to verify that a reader-only selection exposes exactly the requested tools in the supplied order and reads the existing discussion.
GitOrigin-RevId: 60026d73768330ff4b694c3a9e99480e3ed7af27
## Why
Local message boards use a fixed database filename. Callers need separate storage for boards that share a session identity.
## What changed
- Add `LocalBoardOptions` with a validated `storage_namespace`, and support it when opening, configuring, and permanently deleting boards.
- Allow `LocalBoard::with_options` to select storage for its connections.
- Store custom namespaces in `message_board_<namespace>_1.sqlite`, while preserving `agent_message_board_1.sqlite` for the default namespace.
- Have the agent message board extension select the same namespace for opening, setup, and deletion.
## Testing
Add coverage for namespace isolation, reopening saved posts, and avoiding collisions with other runtime databases. Update setup tests to exercise a custom namespace through reopening and deletion.
GitOrigin-RevId: a6ea8d171809625edeedc3a8ac3ee2e1bbf1682a
## What changed
Move the message-board API, types, permissions, storage backends, and tool implementations into the new `codex-message-board` crate. Keep extension registration in `codex-agent-message-board-extension`, and update core and the remote client to use the shared crate directly.
Move the existing board tests with the implementation and add the crate to the Cargo workspace and Bazel build.
GitOrigin-RevId: 33ce527c2bef7f862eba55cb9b32222902ccf4ad
## Why
Mailbox preemption can discard completed response items and lose the response ID needed for incremental continuation. Buffered tool calls must also wait for queued agent messages to be read.
## What changed
- Gracefully interrupt and drain Responses Lite streams, retaining completed items and the response ID for the next request.
- Cancel buffered tool calls before dispatch and suppress partial items that may be discarded during the drain.
- Keep user steering queued until draining finishes. Include pending mail in follow-up requests after disconnects and preserve content-filter recovery guidance.
## Testing
Add WebSocket regression coverage for buffered completion, interrupts without acknowledgement, buffered tools and partial items, user steering during draining, disconnects, and content-filter recovery.
GitOrigin-RevId: c9d3fb87a81850e5686e07f28103e2ff4378ddec
## Why
Failed cloud discovery previously retried before each model step. Apply a cooldown across steps and turns to avoid repeated discovery requests while a provider remains unavailable.
## What changed
- Share per-thread retry backoff logic between cloud skill and plugin catalogs, starting at 30 seconds and doubling to a maximum of 5 minutes.
- Scope cooldowns to authentication and MCP resource generations so account changes and reconnects permit immediate recovery.
- Clear backoff after successful discovery or configuration resets, while allowing cancelled refreshes to retry immediately.
## Testing
Update lifecycle and concurrency tests to cover shared cooldowns, reconnect recovery, and cancellation. Use paused time to verify exponential delays, the 5-minute cap, and backoff across turns.
GitOrigin-RevId: 30dab5ceba51e8ad64c0794c1729719e99f577ce
## What changed
- Add `application.storage.default_project_parent` and `application.storage.projectless_workspace_root` requirements for new Desktop workspaces. Require absolute paths and reject unknown storage keys.
- Apply managed paths to `desktop.defaultProjectParent` and `desktop.projectlessWorkspaceRoot`, overriding local preferences and rejecting writes to managed settings or their containing table while preserving unrelated Desktop settings.
- Expose storage requirements through `config/requirements/read` and update generated JSON schemas, TypeScript types, and Python models.
## Testing
Add coverage for storage validation, per-setting layer precedence, config overrides, write protection, and requirements API responses. Adjust daemon shutdown test deadlines and TUI test synchronization, paste handling, and future allocation.
GitOrigin-RevId: 44b93fe69e6a471c3f485cbf5f9c9ebd34ac3234
## Why
A failed provider catalog refresh could discard model metadata and disable tool search on subsequent turns, even after a successful discovery.
## What changed
Keep the last successful catalog, ETag, and original freshness when refreshes fail without an access rejection. Continue invalidating cached catalogs for HTTP `401`/`403`, rejected refresh tokens, and gateway credentials that require login. Distinguish `GatewayLoginRequired` from login in progress so only the former invalidates the catalog.
Successful empty responses remain authoritative, and cached catalogs remain scoped to the provider and authentication identity.
## Testing
Add regression coverage for catalog preservation across refresh errors and restarts, access rejection invalidation, identity isolation, and tool search and instruction stability on the next turn after a failed refresh.
GitOrigin-RevId: 9dd7c73ac2cc1fd49664efd93b839baa94e05c49
## Why
Authentication can change between model steps. Refresh cloud catalogs before the next step so plugin acceptance and skill instructions reflect the current auth scope, including plugins that are no longer available.
## What changed
- Add a `refresh_catalog_for_step` lifecycle hook and rebuild MCP when plugin catalogs change.
- Reuse successful discovery within a turn and auth scope, serialize concurrent refreshes, and allow retries after failures or cancellation. Limit cloud skill discovery warnings to once per turn.
- Skip cloud discovery during prewarm and other context-only captures, and check hosted plugins even when no executor roots are selected.
## Testing
Add regression coverage for mid-turn auth rotation and plugin removal, serialized refreshes and retries, stale discovery results, and prewarm boundaries.
GitOrigin-RevId: 355681e84dc765e8dc1a04608779d95c706c2188
## What changed
Represent collected fragments with `RetainedInstructionItem`, preserving their content, semantic kind, and source. Assign retention policy when composing context: user instructions, host notices, and framing remain required, while assistant context remains optional at commentary priority.
GitOrigin-RevId: 0080f3abd0d763bd402e78da0ac1b0fa0cd7678d
## Why
Installed-plugin inventory fetches could keep following pagination cursors
without an overall bound, delaying release of the plugin mutation gate.
## What changed
- Limit each inventory fetch to 60 seconds and 100 pages, and reject repeated
pagination cursors.
- Request 200 items per page for scoped inventory fetches as well as all-scope
fetches.
- Report overall fetch timeouts as `remote_catalog_timeout`.
## Testing
Add coverage for cursor cycles, completion at the page limit, and overall
timeout cancellation after successful pages. Verify that failed pagination
preserves the cached inventory and plugin files and releases the mutation gate.
GitOrigin-RevId: e8429dd804581041d102defa9d19db02bcb1739b
## What changed
Add a save RPC that accepts the existing report options and an absolute `savePath` on the server host. Save the report, selected attachments, and a manifest without uploading, and return `savedPath`, `threadId`, and `promptHash`. Local exports remain available when uploads are disabled by configuration.
Publish archives atomically without replacing existing files. Bound uncompressed contents to 200 MiB and attachments to 128, preserve duplicate attachment names safely, and fail without publishing a partial ZIP when a selected file cannot be read. Honor `includeLogs` and omit automatic rollouts for untouched threads without materializing them.
Update protocol schemas and generated TypeScript and Python types.
## Testing
Add coverage for local-only saves, disabled uploads, path validation, overwrite protection, log consent, archive budgets, duplicate names, missing files, and rollout selection before materialization and after writer shutdown.
GitOrigin-RevId: 222c0fe7ab923904de9c8a9f12ad27f7a0104875
## Why
Resuming a thread can replace saved model settings with global defaults. A selected service tier can also be lost when the current model does not support it.
## What changed
- Restore saved service tier, reasoning summary, and personality on resume unless the caller explicitly overrides them, respecting thread ownership of settings snapshots.
- Retain the selected service tier for persistence, settings notifications, and resume comparisons while filtering the effective tier against the current model's capabilities.
- Carry the runtime model, service tier, and personality into thread revert reconstruction.
- Persist initial thread settings before user input during MCP startup so an interrupted first turn retains its settings checkpoint.
## Testing
Extend regression coverage for cold resume and explicit overrides, unsupported service tiers, settings preservation during revert, and settings checkpoints before user input when MCP startup is interrupted.
GitOrigin-RevId: e5ec012a5678048c5eedd443e524f38c142bf9fd
## Why
A delayed conversation classification can reach execution after a newer request has already published a score and committed history, leaving it with no useful work to contribute.
## What changed
Pass the classification's sampling timestamp into conversation requests and check for newer scores and retained history before starting inference. Return `Superseded` only when both have advanced; keep requests that can still contribute either a score or history.
## Testing
Add regression tests covering supersession of scores and history separately and together, including a classification delayed by permissions. Verify skipped requests send no inference request and preserve cached scores, approvals, and score progress.
GitOrigin-RevId: ef78a59e2010a82ac4f1fb93568a562447187099
## What changed
- Extend app-server tests for account credential reuse, refresh races, stale bearer rejection, retry exhaustion, and step-up failures at the identity provider and resource authorization server.
- Assert token-exchange request fields and enterprise protocol negotiation while preserving the ordinary MCP handshake.
- Add core scenarios for permission and tool-policy refreshes, policy revocation, and thread removal, including cancellation of in-flight enterprise calls.
- Snapshot conversation history after revocation to verify that prior calls remain visible and ordinary tools continue working.
- Allow the existing startup-failure test to wait through the MCP startup deadline.
GitOrigin-RevId: 31cf71a28f91bd30746b74df8fa669edb6bc23a1
## What changed
- Prefer an eligible account grant over saved enterprise credentials, with a fallback to saved enterprise sign-in during initial connection if the account grant is rejected.
- Remember rejected account grants per enterprise MCP registration across threads and report authentication failures as requiring reauthentication. Keep administrator configuration errors distinct from sign-in failures.
- Compare account grants independently of access and ID token rotation, recheck the latest identity eligibility after exchange, and retry once when the refresh grant changes.
- Make MCP authentication status reflect usable account grants, saved enterprise credentials, and completed connection failures.
## Testing
Extend regression coverage for enterprise sign-in failures and recovery, MCP authentication status, account grant identity across token refreshes, and retries after transient credential failures.
GitOrigin-RevId: eeac65f668d5d2d55e9d8bf4547fbcbdd6a6c42d
## Why
Managed stderr logs previously retained a tail only before a new launch, allowing logs to grow during a long-running process. Direct stderr writes could also block lifecycle operations on disk I/O.
## What changed
- Route managed app-server and updater logs through a shared nonblocking worker with a bounded queue. Drop records when the queue is full or a record exceeds 16 KiB, and avoid waiting for the worker at shutdown.
- Prune live logs when they exceed 1 MiB, preserving the latest 256 KiB in `.log.previous` before truncating the current file. Keep inherited stderr handles usable after truncation.
- Pause predecessor pruning during Windows updater handoff and resume it after a failed handoff only when ownership has been restored.
- Keep ordinary stderr output when managed logging has not been initialized.
## Testing
Add regression tests for stalled workers and queue overflow, tail retention, inherited append and nonappend handles, and pruning exclusion during updater handoff.
GitOrigin-RevId: b6301962b1180fa6cdb231a56edce2429dc20b54
Pause Tokio time after the validation timeout in
`malformed_handshakes_preserve_active_streams_and_admission_recovers`
while inspecting diagnostics, since that inspection does not poll the
harness WebSocket for pings. Resume time before closing the WebSocket so
the close handshake and its timeout run with an active clock.
GitOrigin-RevId: 733a1fa35f80561424be97c4beb233780a7796cb
## Why
Dropped lifecycle facts can leave incomplete turns retaining analytics metadata indefinitely.
## What changed
- Limit retained turns to 256, evicting the oldest turn and its associated tool and review state when capacity is reached.
- Forward `ThreadClosed` notifications to analytics and discard the closed thread's turn and review state, including reviews received without turn lifecycle facts.
## Testing
Add regression tests covering metadata release after capacity eviction, continued reporting of new complete turns, cleanup isolated to the closed thread, and ignored approval responses after thread closure.
GitOrigin-RevId: eed0a1fef013381654345d535f03de42b99e8607
## Why
Guardian review metrics derived duration from wall-clock timestamps, while analytics measured elapsed time when tracking the result. Clock changes and delayed tracking could produce inconsistent latencies.
## What changed
Capture one monotonic duration at review completion and pass it to both metrics and analytics. Preserve wall-clock timestamps for event ordering, and saturate analytics latency conversion at `u64::MAX`.
## Testing
Add a regression test covering skewed wall-clock timestamps and delayed tracking, verifying that the duration metric uses the captured interval and completion timestamps are preserved.
GitOrigin-RevId: 19eafb8e1f1a3d13a751bdc72c218e4cafeae410
## What changed
Add `codex-utils-critical-path` to track shared dependency paths and attribute time to inference, automatic review, user waits, tools, overhead, and unknown work. Parallel joins retain the later-finishing branch, preserving its full history without double-counting overlapping work.
Clip attribution to the requested interval and count unrecorded gaps as overhead. Bound retained history by collapsing it into unknown work at the depth limit, with nonrecursive destruction and bounded debug output.
## Testing
Add unit tests for parallel joins, nested dependencies, ties, user waits, interval clipping, overlapping segments, history exhaustion, and safe formatting and concurrent release on small stacks.
GitOrigin-RevId: 2df7453162491d3ae97bf4bee942730c8a4d5e5c
## What changed
Store `RolloutItem::SessionMeta` and `RolloutItem::Compacted` payloads in `Box` values and update their producers, consumers, and test fixtures. Preserve the existing serialized rollout format.
GitOrigin-RevId: 59fcc4beb47d23160872dcb6e95ace9b142d3d11
## Why
Reducing code cell lifecycle events discarded their response payload references and lost the initial response status when the runtime status changed.
## What changed
Store the initial response status and separate raw payload IDs for the initial response and final result on `CodeCell`. Carry payload references through queued lifecycle events so they survive replay when the source item arrives later.
## Testing
Extend lifecycle tests to verify distinct initial and final payload IDs and the preserved initial status for yielded cells that complete and failed cells whose lifecycle events precede their source item.
GitOrigin-RevId: 8f49a3174ea487a5d91ebc2a6c97133aacaabfd2
## Why
Sessions using the same MCP catalog can repeat schema normalization and retain duplicate tool specifications.
## What changed
Split immutable MCP tool metadata and normalized specifications from per-handler caches, and reuse matching definitions within each `McpManager`. Require matching tool metadata, agent-plugin mode, and schema budget before reuse.
Keep shared definitions behind weak references and prune expired entries on cache misses so completed sessions do not keep their schemas alive. Build definitions outside the shared index lock.
GitOrigin-RevId: 02345359a1b0d297ca56f2f469de88aba89106f1
smurching-oai