Commit Graph
94 Commits
Author SHA1 Message Date
felixxia-oai fad8b3d5bc Initialize Guardian evidence cursors from thread history state (#60670)
## What changed

Expose history initialization and the effective history mode to thread-ready contributors. Add `EvidenceCursor::at_thread_ready` with paginated-history validation:

- Start new or cleared threads at position zero without requiring an existing rollout.
- Start resumed, forked, or supplied-history threads at the durable tail to avoid backfilling existing evidence.
- Skip the initial `SessionMeta` header when capturing evidence from position zero.

## Testing

Extend evidence tests to cover initialization before a rollout exists, capture of the first original record, rejection of resume without a rollout, and exclusion of existing records on resume.

GitOrigin-RevId: 32aa8e2afb71eb4e882b5f1fb107ac8616b33c8e
2026-11-10 22:00:08 +00:00
Dmitriy Kovalenko 76dcb8199b Honor permission grant scopes in executor capability discovery (#60461)
## Why

Approved filesystem access should make executor skills discoverable and readable for the lifetime of the grant. Turn-scoped access must not authorize persistent plugin servers or hooks, or bypass disabled-plugin policy.

## What changed

- Apply environment-specific session grants to capability discovery and refresh MCP after session grants change.
- Use a separate discovery inventory for skills with turn-scoped grants, keeping plugin servers and hooks on the session inventory. Reuse discovery results across unchanged model steps.
- Preserve disabled-plugin checks for unreadable manifests and symlinked manifest files or directories, including ancestor markers, while retaining standalone skill roots.

## Testing

Add coverage for V1 and V2 skill discovery before approval, after approval, and on the next turn; disabled-plugin ownership checks; discovery cache reuse; and session hooks surviving turn-scoped discovery failure.

GitOrigin-RevId: 929a7f8cb01f4bcfb4bee2c50ae02cd762066b2d
2026-11-09 22:20:07 +00:00
ioanbudea-OAI 86ac091659 Preserve turn configuration in Guardian scoring and approvals (#59642)
## Why

Guardian scoring and cached approvals need to use the requesting turn's captured policy even when the live thread configuration changes. Newly tightened mandatory review requirements must still prevent cached approvals.

## What changed

- Pass captured turn configuration through tool lifecycle and approval review inputs, and use it to resolve Guardian policy and classifier instructions.
- Apply current reviewer and model requirements before scoring and cached approval, including rechecking restrictions after collecting score authorization.
- Count computer-use JavaScript execution attempts even when background scoring is disabled, so a synchronous call consumes the initial-call allowance.
- Retain runtime configuration snapshots when applying agent role overrides, and share Guardian policy text resolution between configuration loading and runtime publication.

## Testing

Add and extend regression coverage for captured policy instructions, reviewer restrictions, stale cached evidence, and a synchronous computer-use call followed by an adaptive turn.

GitOrigin-RevId: 24ed4a1b6ce1733ad775790f30e440da614f9cd6
2026-11-05 11:41:09 +00:00
rhan-oai 2fa67f03f3 Allow volatile World State sections to refresh after compaction (#58956)
## Why

Host-provided context such as budget reminders can become stale while compaction runs. Replaying a captured value or retaining its comparison snapshot can prevent the next step from emitting a fresh reminder, even when the remaining budget has changed.

## What changed

- Add `WorldStateSectionContribution::without_full_context()` to omit a section's fragment and comparison snapshot from initial and compacted context rebuilds while preserving ordinary step updates.
- Reset comparison state after prefix compaction for sections excluded at either capture or installation, including when their messages survive in the live suffix.
- Include step World State updates when building debug prompt input.

## Testing

Add regression coverage for budget reminders refreshing after remote compaction within the same threshold, unchanged request prefixes within a context window, refreshes with a retained live suffix, and step reminders in debug prompts.

GitOrigin-RevId: ea4456755e16c93b5d16e8c762b26a33f8571a86
2026-11-02 19:38:59 +00:00
jif 9def09afa7 Use serving assessments for adaptive Guardian review (#58075)
## What changed

- Match final serving assessments to their response and output item, then retain them with tool invocations and code cells across waits.
- Use matching assessments instead of asynchronous scoring for adaptive review: low-risk actions can skip synchronous review, while elevated-risk actions require it.
- Revalidate captured authorization and policy before approval. Require fresh review for stale or incomplete authorization, preserve explicit review requirements, and discard assessments when hooks rewrite tool input.
- Keep asynchronous scoring as the fallback when an assessment is missing or does not match.

## Testing

Add integration coverage for direct and nested tool calls, hook rewrites, explicit escalation, interrupted responses, catalog-driven policy, incomplete authorization, and assessment retention across code-cell yields and waits.

GitOrigin-RevId: 4346a227204137f018c03c34b302d6f1e9592c96
2026-10-29 12:54:39 +00:00
Francis Chalissery d416d8a0ff Support window prefixes for extension world-state sections (#57276)
## What changed

Add `WorldStateSectionContribution::with_window_prefix()` to render extension developer fragments as separate messages before retained history when building a context window, including after compaction. Updates within an existing window still append to history. Ignore fragments with other roles for sections that opt in.

## Testing

Add a unit test for prefix placement and duplicate avoidance, plus local and remote compaction scenarios covering live updates, persisted replacement history, and rejection of non-developer fragments.

GitOrigin-RevId: 8621457c6eec17f1585ca8e6fa154b1bf3257330
2026-10-26 21:46:26 +00:00
Anthony Tafoya 9e12b14891 Refresh cloud plugin and skill catalogs after mid-turn auth changes (#55957)
## 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
2026-10-21 18:37:30 +00:00
Ruslan Nigmatullin 87f3c3284e Track code-mode cell lifecycle and result buffering time (#55653)
## What changed

- Add `CodeModeCellTiming` notifications for yielded executions, terminal waits, and unavailable cells, with thread, call, and cell identifiers.
- Measure how long completed results remain buffered before retrieval and carry `result_buffered_for` through the runtime, stdio protocol, and gRPC response metadata. Immediate delivery reports zero; waits without a completed result report no buffering measurement.
- Default missing buffering data to zero for compatibility with older hosts.

## Testing

Extend runtime tests to verify buffering duration with a paused clock and retrieval through termination. Add gRPC conversion tests for nanosecond precision, optional metadata, and the distinction between immediate delivery and no completed result.

GitOrigin-RevId: 60c0c5a3affa69ac3779bf1812fca50f5748af76
2026-10-20 18:53:06 +00:00
Anthony Tafoya 9b865ab31e Prepare skill catalogs before building step tools (#55357)
## Why

Skill tools depended on World State rendering to initialize their catalogs and could return `skill catalog is not ready for this step` before rendering completed.

## What changed

- Add an asynchronous `ToolContributor::prepare_for_step` hook and await all contributors before constructing tools.
- Prepare immutable skill catalogs, aliases, plugin policy, and configuration for each step. Tools, invocation detection, and World State rendering consume the prepared state without relying on rendering to perform discovery.
- Apply configuration reloads to subsequent steps while preserving live authorization checks for skill reads.

## Testing

Add coverage for preparation completing before tool construction and World State rendering, skill tools working before rendering, catalogs remaining stable across steps, and configuration reloads during discovery.

GitOrigin-RevId: 1fb95c58c412488353f3bd4403824d6d1476d3d0
2026-10-19 19:33:11 +00:00
Anthony Tafoya a309f9206a Integrate discovery V2 with step-scoped skill catalogs (#55322)
## What changed

- Build executor skill catalogs from discovery V2 inventories without reopening host paths or falling back to legacy discovery.
- Merge Cloud and executor skills by remote plugin identity and skill name, retaining Cloud metadata and accepted provider read routes. Apply plugin selection and disablement before merging.
- Share a stable catalog across each step's rendering, list/read tools, and explicit and implicit skill invocations.
- Preserve unmatched executor entries when Cloud catalogs are unavailable or partial; only complete Cloud catalogs disable unmatched remote entries.
- Bind discovery caches to captured environments and executor sessions, and refresh MCP when V2 inventories change. Preserve pending invalidations when forced MCP refreshes fail.

## Testing

Add regression coverage for V2 inventory sharing and refresh, environment isolation, catalog merging, partial Cloud results, disabled plugins, and catalog consistency across steps and read routes.

GitOrigin-RevId: 842bdcbc9e19364dd69a9abf9e48330b7425a76c
2026-10-19 17:02:20 +00:00
Anthony Tafoya f7a135ec61 Reconcile V2 executor plugins with cloud catalogs (#55301)
## What changed

- Use V2 discovery inventories for plugin MCP contributions when supplied, including empty or failed inventories; retain legacy discovery when no V2 snapshot is supplied.
- Merge plugins by canonical identity, using cloud metadata and enablement for remote plugins and attaching executor installations only when versions match. Preserve configured executor-local plugins and executor enablement when cloud data is unavailable.
- Keep installation order stable across refreshes and run merged MCP declarations on the first retained executor installation while preserving attribution for the others.
- Carry local and remote identities through selected-plugin snapshots, skill attribution, and active-plugin telemetry. Track disabled plugin IDs alongside disabled roots.
- Preserve a complete cloud catalog when a partial refresh arrives, and cap tool-description plugin attribution at 1,024 bytes while retaining all plugin associations.

## Testing

Add coverage for catalog parsing and reconciliation, enablement and discovery-mode selection, installation refresh ordering, executor-specific MCP normalization, selected-plugin snapshots, and bounded UTF-8 attribution text.

GitOrigin-RevId: a8ad352de6a2b8dc4bfd1ff6a5c62ee8ed1d1664
2026-10-19 15:15:17 +00:00
pakrym-oai ce0323f2f2 Support typed content in contextual prompt fragments (#54776)
## Why

Contextual fragments and extension-owned world-state sections currently return strings, limiting them to text content.

## What changed

- Return `ContentItem` from fragment bodies and rendering, preserving classification and attribution. Apply text markers only to `InputText`; pass other content through unchanged.
- Allow extension prompt and world-state contributions to carry typed content, and preserve media in context-window hints.
- Hash non-text world-state fragments by their serialized content and role while retaining the existing text hashing behavior.
- Add shared content-to-tool-output conversion that preserves image references, image detail, and audio.

## Testing

Add coverage for typed world-state hashing and preservation of image content and attribution. Expand code-mode output tests to cover input text, output text, file-backed and inline images, and audio.

GitOrigin-RevId: 2ed00d466a3f169b151202f3eef71e667ea64be1
2026-10-17 04:02:21 +00:00
tongzhou wang 1b6e0d8145 Track content provenance through prompt assembly and compaction (#53831)
## Why

Message roles and content classifications do not identify who supplied each context part. Source attribution needs to survive prompt assembly and history truncation, including messages that combine text from different producers.

## What changed

- Add per-part `ContentItemMetadata` for user input, developer instructions, `AGENTS.md`, skills, harness context, command hooks, and tool context with optional callable namespaces.
- Carry attribution alongside content and classifications through contextual fragments, extension prompts, hook output, and agent messages. Keep tool-provided context-window hints separate from harness text.
- Preserve the source of retained text slices during compaction and attribute generated truncation notices to the harness. Default missing historical attribution to unknown.
- Attribute host MCP calls using cached tool namespaces only when catalog revisions remain stable, without starting servers or refreshing catalogs for attribution.
- Send the new metadata to Responses Lite and strip it from requests to endpoints that do not support it.

## Testing

Extend compaction tests to cover mixed sources, missing or extra metadata entries, and discarded media. Update Responses Lite request assertions to verify endpoint-specific attribution handling.

GitOrigin-RevId: b0ebc929b7c60f3ec49d1affbf897312d63edcc8
2026-10-14 02:01:49 +00:00
Anthony Tafoya cf0ed4c600 Acquire and cache opt-in executor capability inventories via V2 (#53736)
## What changed

- Add the disabled-by-default `executor_capability_discovery_v2` feature to acquire `capabilities/discoverV2` inventories from supporting remote executors using each selected environment's working directory and sandbox.
- Share snapshots with MCP and world-state contributors and skill providers without activating discovered plugins or publishing discovered skills to the model.
- Cache inventories by environment instance, executor session, working directory, and sandbox. Coalesce concurrent requests, refresh the cache at turn start, retry transient failures, and refresh MCP projections after recovery.
- Resolve the system skills root in the executor and pass it explicitly to capability location discovery.

## Testing

Add coverage for feature gating, unsupported executors, inert inventories, cache reuse and eviction, turn refreshes, failure recovery, executor replacement, and system skill locations.

GitOrigin-RevId: 1bdca28fa7f17622b27e2713ab3dcc0c13426889
2026-10-13 18:15:50 +00:00
Adam Perry @ OpenAI 806d2d1cc1 Remove per-content source attribution metadata (#52329)
## What changed

- Remove `ContentItemMetadata`, provenance and namespace types, and the `content_item_metadata` protocol field while retaining `content_item_kinds` classifications.
- Simplify context fragments, extension contributions, and hook output to carry content without source attribution. Return MCP tool results directly without an attribution wrapper.
- Render context-window hints together in a single text fragment.

## Testing

Update Responses and Responses Lite request assertions to expect content-kind classifications without per-content attribution, and adapt context, hook, and extension tests to the simplified types.

GitOrigin-RevId: 201894cd4c93c6d4b4c41c845e7a9af1dd3c95ca
2026-10-08 23:57:36 +00:00
tongzhou wang ad54b75574 Track source attribution for individual context parts (#52195)
## What changed

- Add `ContentItemMetadata` to distinguish user input, developer instructions, `AGENTS.md`, skills, harness context, and tool-supplied context independently of message role and content classification.
- Keep attribution attached to content through prompt assembly, history processing, and hook-output spilling. Distinguish command hooks from MCP hooks and retain tool namespaces when known.
- Treat missing historical attribution as unknown and send per-content metadata only to Responses Lite.

## Testing

Update request and history assertions to cover attribution persistence and Responses Lite metadata handling.

GitOrigin-RevId: 24e56005f5bdb98bf82f2b3943f29c4c81f3d587
2026-10-08 17:56:54 +00:00
mtsui-oai 044da98b4a Expose selected environments to MCP contributors (#51503)
## Why

MCP contributors need the complete executor selection to distinguish an unavailable primary executor from a ready secondary. Ready capability roots alone do not preserve pending or failed selections.

## What changed

Add `with_selected_environments` and `selected_environments` to `McpServerContributionContext`, and pass the captured selections during thread MCP resolution and model-step plugin selection. Preserve priority order and pending or failed entries. Use `None` for threadless discovery and `Some(&[])` for an explicitly empty selection.

## Testing

Add regression coverage for executor ordering with pending and failed selections, environment context in model-step plugin selection, and the distinction between absent and explicitly empty selections.

GitOrigin-RevId: 2c1222fae72c7e2c75b2f40abfad3b2048cf9316
2026-10-06 23:07:48 +00:00
jif 57d57df608 Let late low-risk scores complete pending Guardian reviews (#51396)
## Why

A synchronous Guardian review could keep an action waiting even after an asynchronous score supplied valid low-risk evidence for approval.

## What changed

- In adaptive mode, race the synchronous review against score updates that satisfy the existing cached-approval checks, unless a fresh review is required.
- When asynchronous approval wins, cancel and drain the pending review, then revalidate the action and history lifetime and enforce cancellation and timeout limits before approving.
- Emit an approved review completion without a synchronous assessment or risk level, and use the cached-approval path for the final decision.

## Testing

Add an app-server regression case that holds the synchronous review pending, releases a low-risk score, and verifies approval completes without a risk level. Sequence existing scenarios to finish synchronous reviews before releasing classification results when they test synchronous evidence.

GitOrigin-RevId: 04f5ad6b960b819b565fccb1b85c8039ccbd0479
2026-10-06 15:33:14 +00:00
sayan-oai 9561a34531 Keep plugin and skill snapshots scoped to each step (#50050)
## Why

Environment changes and overlapping catalog refreshes can cause a step to use another step's selected plugins or executor skills. Model context and skill tools need to agree on the capabilities captured for that step.

## What changed

- Resolve selected plugins from each step's ready capability roots, honoring disabled plugins, and retain the snapshot with the step's extension data.
- Use that snapshot for skill attribution in World State and the first step's active-plugin telemetry.
- Return executor skill catalogs directly from refresh so overlapping requests render their own results while retaining thread caches.

## Testing

Add regression coverage for overlapping executor catalogs and environment changes during step preparation. Verify that `skills.list` retains the original step's skills, the next step advertises the new skills, and disabled executor and plugin skills remain absent from model context.

GitOrigin-RevId: 37d3cbe8c7e062fd06290aa985ada4d3b3462d43
2026-10-01 17:44:52 +00:00
Adam Perry @ OpenAI da2e174a66 Scope extension filesystem access to callback permissions (#49898)
## Why

Executor skill reads need to use the environments and permissions captured for the current step, including newly approved grants. Cached continuation pages must remain bound to the filesystem and permissions that produced them.

## What changed

- Supply borrowed `EnvironmentAccess` to turn-input contributors and extension tools, with session and originating-turn grants already applied.
- Require executor skill reads to use filesystem access from an environment available to the callback, and replace separate optional read contexts with `SkillReadContext`.
- Bind skill pagination snapshots to `EnvironmentAccessKey`, which identifies the source filesystem and callback permissions.
- Route image-generation filesystem operations through the same accessor.

## Testing

Extend skill-read integration coverage to verify that a denied reference becomes readable after a turn-scoped grant is captured by the next model step. Update Windows coverage to verify that restricted reads fail when the executor cannot enforce a sandbox, and add a test requiring turn environment selection before exposing capability roots.

GitOrigin-RevId: 825b49302319965835d61c77fc5bcdb951e10543
2026-10-01 06:29:03 +00:00
pakrym-oai b44ca872b1 Persist world-state snapshots alongside rendered context (#49847)
## What changed

Return snapshots and model-visible fragments together from world-state rendering, and use those snapshots for history baselines, step updates, compaction, and new context windows.

Update `WorldStateSectionContribution` so its renderer returns an optional replacement snapshot and an optional fragment. A missing snapshot preserves stored section state, even when retained history requires a fresh render; a null snapshot removes the section. Derive step patches from the history baseline rather than a separately captured live world state.

Adapt skills and Git attribution contributors to the new contract.

## Testing

Add coverage for skipped snapshot updates, explicit section removal, changed-state patch replay, unchanged-state deduplication, and matching live and persisted checkpoints when starting a new context window.

GitOrigin-RevId: af71606a278358ae937853cb76e8a019a501408e
2026-10-01 03:38:50 +00:00
Bryan Ashley f53f5a6fed Expose original error details to turn lifecycle contributors (#49138)
## Why

Turn error hooks receive only a `CodexErrorInfo` category, which omits backend metadata such as usage-limit reset times and rate-limit snapshots.

## What changed

Add borrowed `CodexErrorDetails` to `TurnErrorInput` and pass the original details through turn and compaction error notifications, alongside the existing error category.

## Testing

Add coverage for usage-limit metadata during sampling, manual compaction, and post-turn compaction, including preservation of completed answers and existing error events. Extend lifecycle tests to verify that an `InvalidImageRequest` retains its details when surfaced as `BadRequest`.

GitOrigin-RevId: 6f2058992126da0f69508ef51998e7581557deea
2026-09-29 02:40:51 +00:00
vkg-oai e72da2b538 Stabilize skill catalogs across executor availability changes (#48353)
## Why

Changes in executor availability can alter cloud skill catalog rendering under a shared budget and repeat an unchanged executor catalog when it reconnects.

## What changed

- Cap cloud skills at three quarters of the metadata budget initially, leaving the remaining budget and unused cloud allowance for filesystem skills. Reduce the cloud cap only when doing so allows all skill entries to fit; description truncation alone does not trigger rebalancing.
- Persist the allocation across disconnects, compaction, and thread resume. Recompute it when the cloud inventory or total budget changes.
- Emit a short availability update when an unchanged selected-environment catalog returns and its full text remains in history. Reinject the full catalog when that text is missing.

## Testing

Add coverage for allocation stability, catalog and budget changes, omission handling, and full catalog reinjection after compaction and resume. Extend the selected capability stack test to verify reconnection avoids repeating the catalog.

GitOrigin-RevId: 839fe98024e00c4082a46e6a40675e15fd2c16c6
2026-09-26 04:19:46 +00:00
rreichel3-oai c7e80f873f Preserve executor MCP credential boundaries across reconnects (#48143)
## Why

Cached executor MCP declarations must not gain access to host environment credentials when reconnecting to an executor that lacks environment credential resolution.

## What changed

- Retain credential policy through MCP registration, catalog materialization, and runtime setup. Derive selected HTTP plugin policy from the plugin's source environment.
- Reject host bearer-token fallback and transports requiring host environment headers or helpers for executor-owned declarations. Preserve host-configured credential resolution.
- Include credential policy in connection and tool catalog cache identities, and avoid reading host environment values for executor-only identities.

## Testing

Add regression coverage for legacy executors, cached declarations after a capability downgrade, policy preservation through catalog rebuilds, and connection and tool cache isolation.

GitOrigin-RevId: 2d5b85e4e98668c637c538cbe528914c936bd831
2026-09-25 17:15:29 +00:00
sayan-oai e8098eb406 Separate selected plugin identities from MCP contributions (#47939)
## Why

Plugin identity and skill ownership need to be available independently of MCP data, including for plugins without servers or connectors. Hosted plugins have no executor root and should not own executor skills.

## What changed

- Add `McpServerContributor::selected_plugins` to declare executor plugin identities with deferred MCP data. Let the host assign server attribution and precedence across contributors.
- Represent hosted connectors with `HostedPluginConnectors` and make `SelectedPluginIdentity::selected_root_id` optional instead of assigning synthetic roots to hosted plugins.
- Resolve executor contributions before reading hosted connectors, and clear stale hosted state on account changes before executor loading can wait.
- Preserve connector declarations from disabled plugins so Apps can hide those connectors.

## Testing

Add coverage for account changes while executor loading is pending. Update tests to verify hosted identities have no executor root and disabled plugins hide their connectors even when plugin IDs differ from root IDs.

GitOrigin-RevId: e7f511264c5874c1b56d5f3e4a46820724046db0
2026-09-24 21:03:50 +00:00
jif 61e23bc6a1 Bind Guardian async scores to target environment permissions (#47830)
## Why

Asynchronous Guardian scoring omitted the target environment's denied-read restrictions, and cached score authorization did not include that permission context. A score from one environment could therefore be reused for an action in another environment with different restrictions.

## What changed

- Share permission resolution between synchronous review and asynchronous scoring, including resolving `write_stdin` targets from the owning terminal session.
- Include environment identity and denied-read paths/globs in async scoring context and cached score authorization.
- Fall back to synchronous review when permission evidence cannot be resolved or exceeds the async limit of 3,000 bytes, rather than omitting restrictions.

## Testing

Add regression coverage for rejecting cached scores with different environments or denied paths, passing the target environment's restrictions to the classifier, and rejecting oversized async permission evidence.

GitOrigin-RevId: 14b6f3454bbea458be037560eb1da47d22874df1
2026-09-24 12:34:04 +00:00
Ruslan Nigmatullin a79a18cd0e Expose tool dispatch and timing observations to extensions (#47662)
## Why

Tool lifecycle observers need visibility into calls rejected before execution starts and timing that remains available on cancellation, independently of tracing settings.

## What changed

- Add `on_tool_dispatch` and `on_tool_timing` callbacks to `ToolLifecycleContributor`, with thread, turn, and call IDs for correlation.
- Report direct calls before readiness, dispatch waiting, and hooks. Report handler duration when dispatch completes or is dropped, excluding readiness and gate waiting and returning zero if the call never passes the gate.
- Expose code-mode host durations as a separate `HostOperation` timing boundary. Exclude nested code-mode calls from dispatch and handler timing observations to avoid double counting.

## Testing

Add tests for callback ordering on completed and rejected calls, timing on cancellation, zero handler duration before gate admission without tracing, and code-mode host timing without nested dispatch or handler timing observations.

GitOrigin-RevId: 97998b5094564c13cb2cfad089bb0a9d6615eb5c
2026-09-23 20:49:48 +00:00
Anthony Tafoya 394dadb93d Add turn-start cloud plugin discovery to the MCP extension (#46572)
## What changed

- Support an optional cloud provider alongside executor plugins, refreshing its thread-scoped catalog at each regular turn when plugins are enabled.
- Use cloud connector metadata for app attribution without starting cloud-declared MCP servers.
- Scope cached catalogs to authentication and Apps availability. Retain the previous catalog on discovery failure within the same scope, clear stale catalogs on auth changes, and reject results completed under outdated credentials.
- Add a cancellable regular-task startup phase that reconciles MCP before discovery and reapplies contributor state afterward without invalidating resource caches.

## Testing

Add regression coverage for per-turn refresh, empty catalogs, discovery failures, plugin disabling, auth changes during discovery, MCP reconciliation and cancellation, and cloud app attribution alongside executor MCP tools.

GitOrigin-RevId: 8469eb109ea316e21cbc4122c52ad8000caeef03
2026-09-19 02:07:58 +00:00
papayo-oai b244e9bafe Add a command-start callback for tool lifecycle extensions (#46539)
## What changed

Add `ToolLifecycleContributor::on_command_start` with `CommandStartInput` carrying the resolved command, working directory, selected executor's filesystem, and lifecycle context. Await contributors after command hooks and environment resolution so they can prepare the filesystem before `exec_command` runs, including calls from code mode.

Reserve process IDs only after preparation completes so cancelled preparation does not consume an ID. Keep deterministic process IDs from being reused after removal.

## Testing

Add tests for direct and code-mode calls that verify callbacks receive hook-rewritten commands and the executor's working directory, filesystem preparation completes before execution, and interrupted preparation prevents execution without consuming a process ID. Add a regression test for deterministic process ID reuse after removal.

GitOrigin-RevId: d10bcafcc583a6e3e8b5924d004806727e3763bb
2026-09-19 00:46:06 +00:00
teddywyly-oai 6d75b525ea Allow hosted Apps MCP contributions to override the protocol mode (#46019)
## What changed

Add an optional `protocol_mode` to `McpServerContribution::HostedApps` and apply it when registering the server. Preserve the hosted Apps default when no override is provided.

## Testing

Add an integration test that selects MCP `2026-07-28` with the global feature disabled and completes a native user verification flow, including resuming the tool call with the verification response. Extend capability tests to confirm that the override preserves hosted Apps user verification without granting it to ordinary extension registrations.

GitOrigin-RevId: e8662c45f964c00b0bb4503bbcfd8432d7e2fa70
2026-09-16 20:32:52 +00:00
felixxia-oai 1fd392f6b2 Retire the unused Guardian extension prototype API (#45679)
## What changed

- Mark `guardian_ext` as a removed compatibility flag.
- Remove `InternalSessionSpawner`, `InternalSessionSpawnFuture`, `ApprovalReviewInput`, and `ApprovalReviewError` from the extension API, along with the spawner test.
- Remove `NodeReplReviewEvidence::review_inputs` and make the core re-exports of `NodeReplReviewEvidenceMode` crate-private.

GitOrigin-RevId: 1c532bc57264644cf521ef8077285a9b8ed67391
2026-09-15 11:39:20 +00:00
jif 18d7ace221 Move Guardian reviewer lifecycle into the extension (#45537)
## Why

Guardian reviews must stop when their parent shuts down or their history is reset. Shutdown must also finish reviewer cleanup before closing the parent's persistent history, including when a review is waiting to retry after a rate limit.

## What changed

- Let the Guardian extension own reviewer startup, prewarming, and shutdown through `ThreadManager`, tracking and joining outstanding work during teardown.
- Cancel reviews on history reset or parent shutdown and reject decisions returned after cancellation, including cached extension decisions.
- Keep background prewarm previews from overwriting the active turn's model metadata.

## Testing

Extend regression coverage to verify reviewer cleanup after parent shutdown, prompt shutdown during a 60-second Guardian retry with the network request denied, and preservation of active model review requirements during prewarming.

GitOrigin-RevId: 05efc369b9a1642a10365eac8a09da77f3f6e28c
2026-09-14 22:58:51 +00:00
thomas 6baa076eb6 Allow extensions to select MCP protocol mode per HTTP server (#44571)
## Why

Extension-owned HTTP MCP servers previously inherited the default protocol mode. Extensions need to select a mode for their own server independently of other HTTP servers.

## What changed

- Add `McpServerContribution::SetWithProtocolMode` and re-export `McpProtocolMode` through the extension API.
- Carry the winning registration's protocol override through catalog resolution and materialization, and apply it to Streamable HTTP connections.
- Preserve existing defaults when no override is present. Selecting a protocol mode does not grant host-owned Apps cache access or environment authority.

## Testing

Add coverage for registration precedence and materialization, and extend cache isolation tests to cover explicit protocol selection. Add an integration test verifying that extension servers can select either the legacy or newer protocol while other HTTP servers retain the default mode.

GitOrigin-RevId: 606cecc03094a93258ffc433849f575020b81473
2026-09-10 16:02:59 +00:00
felixxia-oai 9c4879f3a5 Preserve complete actions in Guardian approval reviews (#44569)
## Why

Truncating action arguments can leave approval reviewers evaluating incomplete actions. Large actions need complete review input and explicit handling when they exceed the review budget.

## What changed

- Remove action truncation and the fixed synchronous action byte limit. Admit complete actions against the whole-request budget, splitting long text losslessly into bounded transport parts and accounting for their framing.
- Route actions exceeding the asynchronous action budget to synchronous review. Prevent cached scores from covering oversized calls, including expanded approval arguments, while allowing later small actions to recover score reuse.
- Request user approval when optional review exhausts its local input budget. Keep required review and compaction service failures closed to approval, and retire exhausted review sessions.

## Testing

Add coverage for complete large-action delivery, optional user fallback, required-review denial, subsequent review recovery, async overflow through MCP approval routing, and lossless text splitting with budget accounting.

GitOrigin-RevId: 08f06e68a94b2f779480dd6b6cf5bc30b241f6dd
2026-09-10 16:01:02 +00:00
jif 2617ed2e1c Move synchronous Guardian orchestration into the reviewer extension (#44252)
## What changed

Move the synchronous review loop, outcome mapping, deadline helpers, and session pool into `codex-guardian-reviewer`. Core supplies adapters for evidence capture, authorization checks, session creation, and event publication through the new host interfaces.

Have `guardian-v2` initialize the reviewer pool and host separately, with the pool managing prewarming, session reuse, concurrent forks, invalidation, and shutdown.

## Testing

Extend the app-server reviewer lifecycle test to cover interrupted concurrent reviews as well as completed reviews, including resuming a reviewer after parent shutdown. Adapt core reviewer tests to use the pool and factory interfaces.

GitOrigin-RevId: 2d69bd20c169b20534764b98cbdb63f38564a530
2026-09-09 17:59:10 +00:00
Abhinav 8ff4aa8ee4 Use captured step model settings for extension context (#44202)
## Why

Model switches within a turn can leave extension context using stale model metadata. Skill catalogs, context windows, and metric attribution need to match the model captured for each sampling step.

## What changed

- Pass captured model metadata and step-specific telemetry to world-state contributors.
- Use that metadata for skill catalog budgets and usage instructions, preserving configured budget overrides.
- Supply the captured model's usable context window to turn-context contributors, including when rebuilding context.

## Testing

Add regression coverage for model switches during skill discovery, catalog budgets and metric attribution, and extension context windows after `new_context` and retained-step context rebuilds.

GitOrigin-RevId: c51b40b739a380b9767a54ea8c8ee6de45bf01c8
2026-09-09 16:48:49 +00:00
jif 1e66885a16 Discount an approval's own code-mode wrapper from Guardian score lag (#43527)
## Why

A code-mode `exec` wrapper and its nested tool calls each advance Guardian's score lag. Counting the unscored wrapper against its own nested approval can trigger a synchronous review sooner than needed.

## What changed

Track originating response items and approval tool-call IDs so cached approvals discount their own wrapper when it is newer than the latest score. Preserve wrapper identity across yields and waits, and retain the full lag when provenance is missing or evicted. Other wrappers and scoring failures still count.

Use the current `write_stdin` call's identity for approval freshness, and omit tool-call identity for `execve` intercepts that only retain the launch ID.

## Testing

Add coverage for wrapper-specific discounts, already-scored wrappers, missing and evicted provenance, identity across code-mode resumes, and approval identities for stdin writes and `execve` intercepts.

GitOrigin-RevId: 4002a06d64c4e1e0a13c072303503d6ddfacad09
2026-09-07 17:35:23 +00:00
jif db0568dbbb Remove legacy Guardian approval review paths (#43462)
## What changed

- Remove the `fast_decision` and `full_review` extension hooks and `ApprovalAssessment`, leaving `ApprovalReviewContributor::decide` as the approval decision interface.
- Remove the duplicate Guardian V2 fast-approval implementation and `StrictReviewReason` state.
- Require an explicit `GuardianReviewReason` for synchronous reviews.

## Testing

Migrate Guardian reviewer tests to the production `decide_approval` path and cached-score tests to the registry's decision interface.

GitOrigin-RevId: 1f9092c0ff5bb1eed0c63fa557fb3807e2e71564
2026-09-07 13:02:46 +00:00
jif 0ae02915bd Add request-scoped Guardian approval decisions (#42807)
## What changed

- Extend `ApprovalReviewContributor` with a `decide` hook that can allow an
  action, return a synchronous review result, or hand the request to the user.
- Pass contributors the request's approval state and a callback bound to the
  existing synchronous reviewer. Review results satisfy the review gate while
  the host continues to enforce permissions.
- Add structured reasons for requesting a fresh Guardian review and select the
  first contributor decision in registration order.

GitOrigin-RevId: ec84474b3fcbcf1027c8502aa1c2947c10149f9d
2026-09-04 17:38:39 +00:00
jif 4e48cd02da Honor model-provided Guardian review policies (#42744)
## What changed

- Add optional per-model Guardian modes for computer use, shell, code mode,
  file changes, MCP, network, and permission requests.
- Apply disabled, synchronous, or adaptive review by action category while
  preserving legacy configuration when model policy is absent.
- Bind cached risk scores to the active model policy, thread settings, and
  environment selections so outdated scores cannot approve later actions.
- Keep unknown review modes on the synchronous path and ignore unknown policy
  fields for forward compatibility.

## Testing

- Cover policy serialization, legacy fallback, scoped tool classification,
  disabled categories, and score invalidation after policy changes.

GitOrigin-RevId: f334a5f4fffb198da66cde3a400233e0df650516
2026-09-04 11:59:18 +00:00
jif 86b7d12742 Add Guardian V2 analytics events (#42144)
## What changed

- Emit `codex_guardian_v2_classification` events with the outcome, risk
  level, duration, model, and thread attribution.
- Emit `codex_guardian_v2_fast_decision` events when Guardian V2 approves an
  action without a full review.
- Add `guardian_v2_enabled` to turn analytics, accounting for review scope
  and model requirements.

## Testing

- Extend the Guardian V2 app-server coverage to verify event payloads,
  attribution, enablement state, and scoped event emission.

GitOrigin-RevId: 0805975f8e8f9d954aa1058ea8c60545a090ae74
2026-09-01 22:02:42 +00:00
Eddie Chen 60fc699560 Give Guardian classifications distinct turn identities (#41385)
## What changed

- Generate a UUIDv7 turn ID for each Guardian classification and preserve it across retries.
- Record the owning turn as `parent_turn_id` and propagate its trusted `root_turn_id` when known, including through the tool lifecycle extension API.
- Omit ambiguous or unknown root lineage and keep classification identities distinct when reusing Guardian sessions and WebSocket connections.

## Testing

- Cover classification metadata across retries, connection reuse, concurrent samples, auth refresh, and consecutive owning turns.

GitOrigin-RevId: 78757f2ad031844e4b545802f634c678a9ee2298
2026-08-28 18:23:47 +00:00
Evan Fannin 5bf0ba3dd6 Let extensions process MCP tool results (#41202)
## What changed

- Add `ToolLifecycleContributor::on_mcp_tool_result` with access to the
  executed MCP tool context, rewritten arguments, extension data stores, and
  mutable server result.
- Run contributors before publishing MCP completion and before preparing the
  result for the model, so extensions can inspect or replace successful and
  error results.

## Testing

- Cover unchanged, replaced, and error results for direct MCP calls, plus
  replaced and error results through Code Mode.
- Verify completion waits for result processing and the processed result flows
  to both completion events and subsequent model input.

GitOrigin-RevId: ece7acfe9c5eed4778cb78a29c3ffd06a74e1bcf
2026-08-27 20:39:39 +00:00
felixxia-oai aa89cf62bb Freeze plugin roots in MCP tool attribution (#41117)
## Why

Plugin MCP trust decisions must describe the exact host-discovered server
registration. Looking up a plugin again by ID can resolve a different root.

## What changed

- Retain the host plugin root in MCP catalog attribution and connection identity.
- Reconnect an MCP server when its attributed plugin root changes.
- Pass the retained root through tool lifecycle attribution and use it directly
  when building trusted tool context.

## Testing

Add coverage for catalog identity, connection reconciliation, lifecycle
attribution, and rejecting a different root associated with the same plugin ID.

GitOrigin-RevId: bc92743dd40ab44065a07c3ea972b4956ef8351f
2026-08-27 11:27:50 +00:00
Adam Perry @ OpenAI 81e180044d Scope extension capabilities to invocation lifetimes (#41020)
## What changed

- Add callback lifetimes to extension `ToolCall`, `ToolEnvironment`, turn-input context, and skill-read request types.
- Require extension tool executors to handle calls for any invocation lifetime and tie their returned futures to that lifetime.
- Update built-in extensions, adapters, and tests to use the lifetime-scoped APIs without retaining host-owned environment capabilities.

GitOrigin-RevId: aca9ce3a1b6870df550bc84ec9308cc65533f161
2026-08-27 02:00:31 +00:00
felixxia-oai 21ff2e802c Expose MCP provenance to tool lifecycle extensions (#40976)
## What changed

- Add optional `McpToolContext` metadata to `ToolStartInput`, exposing the
  model-visible MCP tool details and its source classification without exposing
  the executable client.
- Classify MCP calls as connectors, configured servers, plugin servers,
  executor-selected plugins, or other registrations based on the prepared call.
- Prepare each MCP call before notifying tool lifecycle contributors and reuse
  that same call for execution so the callback describes the call that runs.

## Testing

- Cover host-owned connector calls and extension-owned Apps server calls,
  including their distinct provenance and executed tool names.

GitOrigin-RevId: efd23f511b1045ffdd96349d621ad62365569b2b
2026-08-26 21:30:23 +00:00
hefuc-oai 5ce04789fb Honor managed plugin disablement for executor capabilities (#40640)
## Why

Selected executor plugin roots could still expose capabilities when managed
requirements disabled the `plugins` feature.

## What changed

- Suppress MCP servers, skills, apps, and connectors from selected executor
  plugins when `plugins` is disabled, while preserving the selected-root
  identity used to filter those capabilities.
- Apply the policy consistently to direct selected-root discovery and batched
  executor capability discovery.

## Testing

- Add contributor-level and app-server coverage that verifies disabled plugin
  capabilities are absent and their MCP servers never start.

GitOrigin-RevId: e30bd8a936c4b0e083908b89a65f4d02e01747de
2026-08-25 16:47:13 +00:00
pmccrary-oai 3a469a297d Route extension hints into context-window metadata (#40533)
## What changed

- Add a `ContextWindow` prompt slot for thread-scoped extension contributions.
- Include fragments in that slot in the token-budget context-window message.
- Keep the MCP `notes/thread_hint` bridge when the native history-notes extension is disabled, without falling back to it after a native request failure.

GitOrigin-RevId: befcb7cc8570b846094b2b542a112b186b39a30a
2026-08-25 03:38:24 +00:00
Won Park 333beecd41 Add structured full approval reviews to extensions (#40472)
## What changed

- Split the approval review contributor into a `fast_decision` path for existing evidence and a `full_review` path for structured reviews.
- Add host-owned full-review inputs with action evidence, conversation history, thread and turn identifiers, approval context, and extension data.
- Add structured assessment and error types, and resolve full reviews using the first contributor that claims the request.

## Testing

- Cover full-review input propagation, first-claim ordering, and short-circuit behavior in the extension registry tests.

GitOrigin-RevId: 6f763d874af8c0b66a0d760b7b7c7dfe32251c9e
2026-08-24 20:25:26 +00:00
Tamir Duberstein e3609f2d02 Expose invocation sources to extension tools (#40382)
## What changed

- Add `ToolCallSource` to extension `ToolCall`s and propagate direct or Code Mode
  invocation metadata, including the runtime cell and nested tool-call IDs.
- Add `ToolCall::response_byte_budget` so extension tools can bound direct-call
  responses by the host truncation policy while allowing Code Mode responses up
  to the tool's own limit.
- Share the skills extension's response-size limit between its `list` and `read`
  tools.

## Testing

- Verify that Code Mode source metadata is preserved when core constructs an
  extension tool call.

GitOrigin-RevId: b10aef01be30d35b67bee3af24a1a4175c9ee112
2026-08-24 11:02:14 +00:00