Commit Graph
2981 Commits
Author SHA1 Message Date
tongzhou wang c8c3f70893 Account for synchronous MCP post-hook costs in rollout budgets (#61580)
## What changed

Allow successful synchronous `PostModelRequest` MCP hooks to return `extra_codex_rollout_budget_units` in `structuredContent`. Add these charges to rollout spending while preserving provider usage separately, including completed responses without token usage and incomplete responses with reported usage.

Validate prices as finite and nonnegative. Preserve valid sibling charges when another hook reports an invalid price, and surface accounting failures without retrying inference. Carry hook charges and overflow information through host accounting and enforce local budget limits.

## Testing

Extend coverage for hook-driven budget exhaustion, failed hooks, overflow, interrupted responses, and foreground and background compaction accounting.

GitOrigin-RevId: dc54ba53478d82848bcafb55ecef88215019d09f
2026-11-14 23:53:53 +00:00
tongzhou wang c8e4cbcd94 Support fractional rollout budgets and configurable context reminders (#61579)
## What changed

- Add a positive, finite `features.rollout_budget.limit_budget` allowance while retaining `limit_tokens` compatibility. Reject configurations that specify both.
- Preserve fractional balances in reminders and compare thresholds without rounding the balance down.
- Accumulate charges with compensated summation and allow spending exactly the allowance, including floating-point rounding tolerance for cases such as `0.1 + 0.2` against `0.3`.
- Add `reminder_at_context_start`, defaulting to `true`. Set it to `false` to suppress reminders at thread start and after compaction; combine it with an empty threshold list to disable budget reminders entirely.

## Testing

Extend configuration and rollout-budget tests to cover fractional allowances, legacy integer persistence, invalid limits, threshold-only reminders, and enforcement at decimal equality and after overspending.

GitOrigin-RevId: b8e7a2faa8d39e562bcae196364de419698bc67d
2026-11-14 23:53:23 +00:00
tongzhou wang 92c1a4dc2e Add response processing after completion and usage accounting (#61577)
## Why

Streaming can dispatch tools before a response finishes and its usage is accounted for. Allow callers to wait for completion and synchronous post-hooks so budget enforcement can block tools from the response that exhausts the budget.

## What changed

- Add `responses.process_mode = "after_complete_response"` to buffer response items and release them in order after completion, synchronous post-hooks, and usage accounting. Keep `streaming` as the default and deliver safety-buffering notifications immediately.
- Track whether response usage has already been charged to avoid duplicate accounting, including during compaction.
- Stop new tools across the agent tree when the rollout budget is exhausted. Coordinate steering with accounting and prevent buffered tool calls from dispatching after steering or cancellation.
- Bound buffered responses to 65,535 events and reject overflow.

## Testing

Add configuration coverage and extend integration tests for buffered output ordering, synchronous hooks, interrupted responses and resume, steering, invalid usage, and budget exhaustion preventing tool dispatch.

GitOrigin-RevId: bd06811cfd50cb701babe6353465c57dfed2a37a
2026-11-14 23:50:12 +00:00
tongzhou wang 81b995a6c9 Unify model request hook lifecycles across sampling and compaction (#61562)
## What changed

- Pair `PreModelRequest` and `PostModelRequest` hooks per sampling or compaction attempt, with distinct lifecycle IDs and cancellation reporting.
- Add `thread_id` and `kind` to hook inputs, distinguishing `sampling`, `foreground_compaction`, and `background_compaction`. Include the provider's `response_id` in post-hook inputs when available.
- Wait for synchronous post-hooks before releasing terminal response events, accounting for usage, or installing compaction results. Allow streaming output and tool execution to proceed before the hook finishes, and cancel synchronous post-hooks when their owning request is interrupted.
- Preserve incomplete-response, budget-exceeded, and hook-denied outcomes across sampling retries and compaction model fallback.

## Testing

Add coverage for streaming and tool ordering, interrupted post-hooks, per-attempt metadata, and foreground and background compaction accounting and installation ordering.

GitOrigin-RevId: 3e808c5b3bf2d5a6a326d9f5be5bbd0359b2d16e
2026-11-14 19:40:12 +00:00
goodman-oai 65a5410f72 Add trusted request-country metadata to Responses requests (#61535)
## What changed

- Add `RequestCountry` for host-supplied country codes. Normalize two-letter ASCII codes to uppercase, omit the field by default, and serialize explicit unknown (`XX`) as `null` to prevent fallback to request provenance.
- Carry `request_country` through thread defaults and captured step settings into Codex backend HTTP and WebSocket requests. Allow live country updates without enabling `step_model_switching`, and preserve WebSocket continuation reuse when the country changes.
- Propagate the captured country to subagents, reviews, and compaction requests.

## Testing

Add coverage for country validation and serialization, thread and live-turn update lifetimes, prewarm settings capture, agent inheritance and reloads, compaction and reviews, WebSocket country changes, and HTTP fallback.

GitOrigin-RevId: 787cd1fa7ba5d465a4eb64a92f13c79e93d9bca7
2026-11-14 15:44:38 +00:00
sayan-oai b1573f1024 Describe context and permissions per environment behind a feature flag (#61494)
## Why

Selected environments can have different filesystem and network permissions, but context previously described only the primary environment's rules.

## What changed

- Add the disabled-by-default `per_environment_context` feature.
- Group filesystem and network access under each environment ID, including when only one environment is selected. Label shared network requirements separately.
- Scope permission instructions and incremental updates to each environment. Withdraw stale descriptions and use labeled thread defaults when no environment is ready.
- Track approved command prefixes separately so new approvals do not require repeating unchanged permission instructions.

## Testing

Extend rendering and permission snapshots for scoped context, removed scopes, thread defaults, and format transitions. Update the active-turn scenario to cover two environments with different permissions, changing only the second environment's workspace and permissions and restoring its defaults on the next turn.

GitOrigin-RevId: c948089c12da7b15eee378d8c5a23a233d24502b
2026-11-14 09:43:57 +00:00
Zahan Malkani 5943d7acdf Persist timestamped realtime frames before acknowledging completion (#61473)
## What changed

- Record voice transcripts as `frame` items with a `voice_transcript` channel for eligible threads, while retaining legacy transcript support and the existing app-server live and mixed timeline representation.
- Assign first-observed timestamps that never move backward, retain them through completion, and restore the clock floor from history and compaction checkpoints.
- Flush realtime records before completion notifications, including file and initial directory durability barriers for local rollouts.
- Generate a distinct default session ID for each new call, use the call ID when attaching to an existing call, and derive opening record IDs from session IDs.
- Add protocol types for turn-input acceptance and session-end input receipts, excluding those receipts from live notifications and mixed timelines. Update generated schemas and SDK types.

## Testing

Add coverage for frame serialization and required voice payloads, clock rollback and checkpoint restoration, durability barrier failures and retries, per-call session IDs, and persisted transcript availability immediately after completion.

GitOrigin-RevId: 3f6632b517d563861aee085b723475b8945c391c
2026-11-14 06:14:36 +00:00
Ruslan Nigmatullin c2c1ebba0a Add optional tool output rendering for context window hints (#61452)
## What changed

Add the token-budget setting `context_window_as_tool_output`, defaulting to `false`. When enabled, render agent identity, context window IDs, and associated hints as an `agent.thread_hint` tool output. Use the same format for agent identity updates in forked threads.

Preserve this setting when applying model token-budget defaults without treating it as an explicit budget override.

## Testing

Update fork scenario snapshots to cover tool output rendering, including a child agent's window hint after its inherited parent context and a prediction fork above the compaction threshold.

GitOrigin-RevId: e1a7089935bd0aed44ffd15f0ae3ec41a2607218
2026-11-14 02:55:41 +00:00
Anthony Tafoya d26406987e Add explicit cloud catalog refresh before turn admission (#61449)
## Why

Hosts need to refresh cloud plugin and skill catalogs without starting a turn or waiting for step preparation.

## What changed

- Add `CodexThread::refresh_cloud_plugins()` and a turn-independent extension callback that bypass cached catalog reuse and failure cooldowns.
- Run catalog contributors concurrently while serializing publication with step snapshot capture. Invalidate MCP resource caches before explicit discovery and schedule MCP reconciliation afterward.
- Accept complete and partial catalogs, report discovery errors, and keep failed or cancelled refreshes retryable.
- Allow hosts to gate cloud plugin discovery through effective configuration.

## Testing

Add coverage for discovery before turn admission, catalog replacement and reuse, immediate retry after failure, concurrent contributors, and cancellation-safe publication and invalidation.

GitOrigin-RevId: 83e847fcab6562581af2c9e831b1d2a310f2b270
2026-11-14 02:45:51 +00:00
dob-oai 8afe928a76 Avoid starting dormant REPL servers for cleanup hooks (#61437)
## Why

Bundled `turn_ended` hooks can start an unused REPL server solely to clean it up, even though its deferred startup means there is no process to clean up.

## What changed

Allow bundled cleanup calls for `node_repl` and `cua_repl` to return an empty successful result when startup has never been requested. Limit this behavior to configured, unshared local stdio connections using the legacy MCP protocol. Other calls continue to wait for server startup.

## Testing

Add coverage for the connection eligibility checks and extend subagent hook tests to verify that bundled cleanup leaves unused REPL servers dormant while ordinary hooks can still start cached tools hidden from the model.

GitOrigin-RevId: 0a07772fccd670c53348ff7a702b4bee353e263c
2026-11-14 01:34:12 +00:00
Anthony Tafoya 6d919e65e2 Refresh cloud plugin and skill catalogs during step preparation (#61429)
## Why

Installing a plugin without connector tools must still refresh a cached cloud skill catalog so its skills are visible to the next inference.

## What changed

- Rely on step preparation for cloud plugin and skill discovery, removing their turn-start callbacks and the `requires_mcp_runtime` lifecycle hook.
- Request an MCP runtime refresh after every completed plugin install, including skills-only installs.

## Testing

Add regression coverage for skills becoming visible immediately after a connector-free install. Update lifecycle tests to cover discovery after MCP reconciliation, prewarm skipping discovery, cancellation allowing subsequent preparation, and reuse of unchanged connections. Use deterministic DNS lookups in network proxy authorization tests.

GitOrigin-RevId: 76036bd29253fb8987e606c3c28411c689590fc4
2026-11-14 01:11:17 +00:00
Alex Kira 27e2227b2b Enable incremental tools by default for Amazon Bedrock (#61426)
## What changed

Enable `incremental_tools_legacy` by default so both Amazon Bedrock providers publish tool declarations through `additional_tools` input items. Users can still opt out by setting `incremental_tools_legacy = false`.

## Testing

Update the incremental tool scenarios to cover the default behavior and explicit opt-out for both Bedrock providers. Adjust the Bedrock multi-agent test to verify collaboration tools in `additional_tools` declarations.

GitOrigin-RevId: 7a54aed24039d178f9b7d7f5a3f49fe4d588ffa7
2026-11-14 00:56:00 +00:00
Abhinav 4598f33790 Use thread-scoped model catalogs for model lookups (#61421)
## Why

Model lookups through the startup catalog can use the wrong provider's capabilities or instructions after a default provider switch or when a thread has its own catalog.

## What changed

- Resolve model metadata for sampling, prediction forks, and memory writes through the owning thread's catalog.
- Make the thread's models manager available to extensions before startup callbacks and use it for guardian sampler configuration.
- Import external sessions using the destination provider and originator's catalog, loading its offline cache only when no refreshed catalog is available.

## Testing

Add regression coverage for sampling after a default provider switch, parent-specific prediction fork capabilities, thread-scoped model lookups, and imported sessions retaining destination provider instructions and originator.

GitOrigin-RevId: c86081b9e55c5e2b42a7f082966485d7fb5139d4
2026-11-14 00:32:02 +00:00
Anthony Tafoya 0b297ec5ee Retry incomplete cloud catalogs with exponential backoff (#61400)
## What changed

- Keep partial cloud skill and plugin catalogs, and catalogs with warnings, usable while retrying discovery with exponential backoff. Start retries after 5 seconds and cap the delay at 5 minutes; clear backoff after a complete, warning-free result.
- Replace previous catalog entries with each accepted result, including partial results, and clear cached cloud skill resources when publishing a replacement.
- Check catalog freshness across turns and steps using auth and resource identity. Auth or resource changes bypass retry cooldowns, including within the same turn.

## Testing

Add and update tests for retry delays and recovery, concurrent refreshes, auth and resource invalidation, partial catalog replacement, and cloud skill read-cache replacement.

GitOrigin-RevId: 0ce5a9d5596e8a73d7084abbbaeb4fb8cda84aa0
2026-11-13 23:14:47 +00:00
Matthew Zeng e8f5d2bb55 Restore shared connector metadata for Codex Apps tools (#61398)
## Why

When an Apps `tools/list` response supplies connector names and descriptions in shared `openai/apps` metadata, tools need those fields restored to expose the correct namespaces and descriptions to the model.

## What changed

Match tools to shared connector metadata by `connector_id` on each response page and copy nonempty names and descriptions into tool metadata before conversion. Restrict this handling to Codex Apps; regular MCP servers use the standard tool listing path.

## Testing

Add a regression test with two tool pages that verifies both connectors' namespaces and descriptions appear in the model's tools state.

GitOrigin-RevId: d4c8b44007773ff65d3e366669ee5c705a480cb1
2026-11-13 23:05:04 +00:00
ioanbudea-OAI f626b8fde9 Remove legacy sandbox-policy accessors from configuration (#61395)
## What changed

- Remove `Config::legacy_sandbox_policy` and `Permissions::legacy_sandbox_policy`.
- Update core and TUI tests to use permission profiles and filesystem/network policies directly, with explicit conversion where legacy wire formats are required.
- Report the active sandbox label in app-server resume override mismatch diagnostics.

GitOrigin-RevId: cbc5417944e23404523e69dc5516d66d6bfafefe
2026-11-13 23:03:49 +00:00
Hung 5158e9780e Add opt-in file-backed tool audio (#61394)
## What changed

- Add the disabled-by-default `file_backed_tool_audio` feature to prepare fresh function and custom tool audio through the attachment store before persistence. Store returned file IDs in history and preserve them on replay without uploading again.
- Validate and budget mixed text and audio output before uploading retained clips, preserving audio token charges and producer limits. Resolve tool audio durations before continuation compaction checks.
- Keep inline audio on backend upload failures and use text placeholders for invalid or rejected clips. Replace unavailable tool audio files with placeholders in prompts while retaining the original history, so restored files can be used again.
- Treat file-backed tool audio as a competing source when attributing audio or unknown-file errors to user uploads.

## Testing

Add coverage for direct tool and Code Mode audio across follow-up turns, file expiration, restoration, and resume in both legacy and paginated history. Extend tests for truncation before upload, upload failure handling, bounded error placeholders, and error attribution.

GitOrigin-RevId: 66c1e7699c7d92f98c3a2d0f5cecbbd63147eb6c
2026-11-13 23:02:16 +00:00
Alex Zamoshchin f941a5fb92 Add optional release IDs to cloud plugin source locations (#61393)
## What changed

Add `release_id: Option<String>` to `PluginSourceLocation::Cloud` to carry the opaque release ID selected by the cloud catalog for a plugin declaration. Update catalog and discovery test fixtures to supply the new field, including a populated release ID in catalog reconciliation tests.

GitOrigin-RevId: 1d60e25bd5fb8efd84e6e825305ad68b847fc1ee
2026-11-13 22:44:55 +00:00
Dylan Hurd 347143f140 Preserve active additional context across compaction (#61390)
## Why

Compaction can drop active client context while deduplication prevents unchanged values from being sent again on subsequent turns.

## What changed

Track additional context by history item ID and restore active items when installing complete or prefix compaction results. Preserve their history order, refresh retained items with their original bounded payloads, and leave queued updates to arrive with their input.

## Testing

Add regression coverage for local compaction with message context and remote compaction with tool-output context. Verify that current context survives exactly once, stale values are absent, and restored context precedes the preserved user prompt.

GitOrigin-RevId: a3114f387c6a4ae1a2d1651993d0f7a101be7dd5
2026-11-13 22:19:12 +00:00
Alex Zamoshchin 42f950cd11 Refresh plugin recommendations when tooling state changes (#61387)
## Why

Cached recommendations can keep offering a plugin after installation. Refreshing the shared catalog during step preparation can also leave the recommendation prompt and install handler with different candidates.

## What changed

- Check `tooling_changed_at` before reusing recommendations, with a one-second reuse window and a 250 ms deadline that includes authentication. Refresh the catalog when the marker changes, preserving cached recommendations on failure and retrying on a later poll.
- Scope recommendation and featured-plugin caches by product, backend, and account identity, and share concurrent recommendation refreshes.
- Use the same recommendation snapshot for a step's prompt and install handler, while allowing later steps to observe updated recommendations.

## Testing

Add coverage for marker changes, missing markers, refresh failures, concurrent and cancelled requests, and the authentication deadline. Extend integration tests to verify consistent step candidates and stop offering an installed plugin while preserving the original recommendation block in conversation history.

GitOrigin-RevId: 44fa6723a623d9604f0c768118bc6889d2d3b693
2026-11-13 22:16:32 +00:00
Anthony Tafoya 826b6e8199 Reuse cloud plugin catalogs across turns until invalidated (#61386)
## Why

Cloud plugin discovery previously ran each turn even when authentication and Apps resources were unchanged.

## What changed

- Reuse complete, warning-free catalogs across turns, including empty catalogs, while authentication and the Apps resource cache key match.
- Retry partial catalogs and catalogs with warnings on subsequent turns, preserving failure backoff.
- Reject discovery results if authentication or Apps resources changed while discovery was in flight. Hosts must invalidate Apps resources when the remote plugin catalog changes.

## Testing

Extend lifecycle tests to cover populated and empty cache reuse, resource and auth invalidation during discovery, and reuse of the empty catalog after plugin revocation.

GitOrigin-RevId: f0f0f4fd5a83179d722ba3cb917deec7eef18458
2026-11-13 22:01:26 +00:00
Ruslan Nigmatullin 4aac2eacc5 Make clock reads synchronous and remove client clock requests (#61382)
## What changed

- Make `TimeProvider::current_time` and `CodexThread::current_time` synchronous, with clock reads required to avoid yielding or calling back into the session.
- Remove the experimental `currentTime/read` app-server request and its client-backed clock and sleep polling implementation.
- Require a host-provided `TimeProvider` for external clocks. App-server thread startup now fails when `features.current_time_reminder.clock_source = "external"` has no provider.

## Testing

Add coverage for external provider selection and rejection of app-server thread startup without an external provider.

GitOrigin-RevId: 09479d6eb28453b343dee78b3673092e73ae5d73
2026-11-13 21:51:53 +00:00
Alex Zamoshchin a46c5be77e Delegate installed plugin recommendation filtering to the endpoint (#61374)
## What changed

Request plugin suggestions with `includeInstalled=false` and use the endpoint results without filtering against loaded plugins or the local installed-plugin cache. Retain disabled-tool and client eligibility filtering, and remove plugin loading that was only needed for recommendation filtering.

## Testing

Update tests to verify the query parameter and confirm that endpoint recommendations survive local installed-cache filtering while disabled tools remain excluded.

GitOrigin-RevId: a1b199f9ac8da7dae40abe46b6ccfaf1d8230a5c
2026-11-13 21:00:04 +00:00
Krish Chainani ae12a9d645 Add deferred MCP reprojection for the next regular turn (#61356)
## What changed

Add `CodexThread::reproject_mcp_on_next_turn()` to request an MCP reprojection after `TurnStarted` and before tools are bound. Repeated requests coalesce, and idle reconciliation or prewarming leaves the request pending until the next regular turn.

## Testing

Add coverage for deferred requests surviving prewarm and idle reconciliation, coalescing into one reprojection, and exposing refreshed tools in the turn's first model request. Verify that discovery starts after `TurnStarted`.

GitOrigin-RevId: 260b2381013291a47eb9a7f736c129de76c5897f
2026-11-13 18:52:09 +00:00
ioanbudea-OAI 6919e2cd8e Apply elastic entry caps to Guardian classifier evidence (#61346)
## Why

Trusted classifier evidence needs to fit configured entry limits while preserving complete records and metadata that reflects only delivered reviews.

## What changed

Apply elastic caps before rendering trusted tools, trusted skills, and previous reviews into developer messages. Omit oversized tool attestations and review fragments whole, retain only complete skill paths that fit, and record omissions for the request's omission notice. Keep review IDs aligned with retained fragments and preserve fixed-budget behavior.

## Testing

Add coverage for whole-record omissions, retained review IDs, and omission notices. Exercise orchestrator connector evidence with both fixed budgeting and an elastic tool entry cap.

GitOrigin-RevId: 729db452026f93b59765a3936730b9b6dddfcf9a
2026-11-13 18:47:15 +00:00
ioanbudea-OAI a655b50de9 Add opt-in elastic budgeting for asynchronous Guardian requests (#61345)
## What changed

- Add `async_elastic_budgeting` and `async_elastic_sections` under `[features.guardianv2]`, independently of synchronous budgeting and disabled by default.
- Apply section limits and elastic allocation to new classifier evidence after accounting for setup and retained conversation history. Preserve the configured elastic policy when applying model defaults.
- Keep permission context whole during elastic allocation and record truncation observations when image preparation rejects an image.

## Testing

Add coverage for configuration validation, fixed and elastic conversation budgeting, retained history and cursor preservation, and complete classifier requests that retain evidence beyond the fixed entry limit.

GitOrigin-RevId: 2955f0482f7c3e18325e296fc2ccb6592ffcb73a
2026-11-13 18:46:45 +00:00
Vivian Fang c33002259e Add per-item retention policies for injected history (#61344)
## What changed

Allow `thread/inject_items` to accept wrapped items with `retention: "default"` or `retention: "always"`, while preserving support for raw response items. Persist explicit policies separately from legacy client-authored metadata.

Carry explicitly retained items through local and remote compaction, context-window changes, and child-agent creation, subject to retention budgets. Preserve history order and keep retained tool exchanges together; require matching calls and outputs to be adjacent in the injection request and both marked `always`.

Update Quinn dependencies and record exceptions for additional Hickory DNSSEC advisories where DNSSEC is disabled.

## Testing

Update injection tests for explicit retention metadata and extend the compaction scenario to cover a retained tool exchange across repeated compaction and restart.

GitOrigin-RevId: 1195a4ed62dc162a62e58bc2ad733262fbabea27
2026-11-13 18:46:13 +00:00
Max Johnson 72be450854 Refactor session window access and async compaction test setup (#61263)
## What changed

- Read response window metadata through `current_window_from_state` with an explicit session state lock, releasing it before assembling response metadata.
- Name the token budget feature check used when restoring a forked child window.
- Pass an explicit `TestEnv` into the async compaction test helper and build with `build_with_environment`.

GitOrigin-RevId: 7ad852dbd65689e814f62cad645330ac5dac80f4
2026-11-13 05:15:59 +00:00
cooper-oai b05266d61b Update realtime model and honor provider-specific voice defaults (#61259)
## Why

An unset voice preference should use the call destination's default. Changed or cleared preferences should also take effect on the next V3 call, even when the thread retains an older setting.

## What changed

- Use `gpt-live-1.1-codex` as the default V3 realtime model.
- Preserve an unset voice until request serialization, defaulting V3 API calls to `marin` and ChatGPT calls to `cove`, while honoring explicit voice selections.
- Reload current voice settings when starting a new V3 call through the app server, including cleared preferences.
- Populate the TUI voice picker from the V3 catalog, add `marin` to that catalog, and remove voice-catalog lookups from call startup and preference saving.

## Testing

Add regression coverage for API and ChatGPT voice defaults, explicit overrides, changed and cleared settings, and TUI call requests and voice selection.

GitOrigin-RevId: d0dd4dcd4bcad9ab869d7bf524e7095a88ffcfaa
2026-11-13 05:04:45 +00:00
sayan-oai 6fcfa75a9f Preserve MCP registration scope across client operations (#61257)
## Why

MCP servers registered under the same name in different environments need distinct inventory, routing, and status reporting.

## What changed

- Keep MCP inventory and runtime status keyed by registration scope so same-named servers remain separate.
- Add optional `environmentId` selectors for status lookup, tool calls, resource reads, and OAuth login. Omitting the selector preserves existing name-only routing.
- Carry environment identity through startup notifications, elicitation requests, and tool-call history, and display it in the TUI.
- Support `/mcp login <name> --environment <id>` and update generated protocol schemas and SDK types.

## Testing

Add integration coverage for two environments with identically named MCP servers, tools, and resource URIs. Verify separate inventory, explicit environment routing, legacy routing when the selector is omitted, and errors for unknown environments. Update protocol and TUI tests for environment identity.

GitOrigin-RevId: 074a3db10549438cb6aff33eb7bed10791adae4d
2026-11-13 04:51:06 +00:00
cassirer-openai b8dde1eb40 Move rollout trace recording to a background worker (#61253)
## Why

Recording inference starts synchronously serializes prompts and writes trace payloads before model work can proceed. Defer that work so inference dispatch can continue while the trace is recorded.

## What changed

- Use one worker per trace bundle to write events in submission order, with up to eight pending jobs.
- Share prepared prompts and encoded requests with the worker, and build remote compaction snapshots there. Write payloads before publishing events that reference them.
- Add `TraceWriter::flush` and `ThreadTraceContext::flush` for live reads. Drain accepted jobs and seal the bundle on shutdown; suppress the closing marker after a recording failure.

## Testing

Add tests for deferred prompt serialization, ordered start and completion events, shutdown draining, and serialization failures that preserve earlier evidence without publishing a closing marker. Verify compaction snapshots are built on the worker and skipped when tracing is disabled.

GitOrigin-RevId: 32fd787bb1d281f1814df2f08c73b667628f8add
2026-11-13 04:36:03 +00:00
cassirer-openai e73744b3f4 Store rollout trace payloads in compressed shards (#61222)
## What changed

- Replace individual JSON payload files with numbered Zstandard shards in trace schema v3. Add `PayloadReader` to resolve payload references, cache decoded shards, and validate payloads during replay.
- Capture exact outgoing HTTP bodies and WebSocket messages alongside prepared prompts. Preserve full formatted context for continuations from unrecorded WebSocket warmups while retaining the actual incremental request.
- Flush payload records before publishing references and seal the final shard before recording completion. Replace `recording_complete` with `recording_status` to distinguish unclosed recordings, closed recordings with partial payloads, and complete recordings.

Reduce trace bundles with the binary that recorded them.

## Testing

Add coverage for shard rotation, oversized records, live reads, cache eviction, and malformed or corrupt payloads. Extend HTTP and WebSocket integration assertions to verify captured requests and prepared prompts, including warmup continuations.

GitOrigin-RevId: 40ce3c11383f48e08f52e7120b706c61a449df9c
2026-11-13 01:22:19 +00:00
goodman-oai 83a23c0936 Support model catalog Apps usage instructions (#61213)
## What changed

- Add the disabled-by-default `catalog_apps_instructions` feature to use the selected model's `model_messages.apps.usage_instructions`, subject to existing Apps availability and model gates.
- Preserve valid catalog text literally, allow an empty string to clear guidance, and fall back to bundled instructions for missing, whitespace-only, or oversized values. Overrides are limited to 800 UTF-8 bytes.
- Reconcile guidance with the latest retained Apps fragment, emitting replacement or removal notices when needed and avoiding repeated updates across resumes. Preserve compatibility with bundled guidance and legacy checkpoints.
- Recognize catalog Apps fragments as contextual developer messages in event mapping and rollout rollback.

## Testing

Add resolver, world-state, client, and scenario coverage for literal text and byte limits, feature and availability gates, model switches, clearing guidance, resumes, and legacy history recovery.

GitOrigin-RevId: 6240d407ada8f22fa101753ff3ddb338632ae283
2026-11-13 00:29:24 +00:00
ioanbudea-OAI caf0d30005 Upgrade the Rust toolchain to 1.99.1 (#61205)
## What changed

- Update the workspace, Bazel toolchain and lockfile, CI and release workflows, and Windows setup script from Rust `1.95.0` to `1.99.1`.
- Adapt Rust code to the updated toolchain and Clippy lints, including `try_update` for atomics, `as_chunks` for fixed-size slices, and targeted allowances for generated tonic services.
- Refresh the halfwidth sound mark rendering snapshot to use literal characters in debug output.

GitOrigin-RevId: 5852029f687436dc814c2bca68141ee1f4a67d49
2026-11-12 23:51:52 +00:00
Bryan Ashley fae87361f4 Add graceful agent-tree suspension and recovery (#61198)
## Why

Suspending a locally controlled agent tree must preserve unfinished turns and child results that arrive during teardown so they remain available after recovery.

## What changed

- Add `ThreadManager::suspend_agent_tree` to stop durable trees, close their writers, and return unfinished root and named V2 child turn IDs plus any pending root wake.
- Persist terminal child results before suspension succeeds, including results queued for evicted parents. Report failed or unconfirmed writes.
- Add `load_suspended_agent` to restore named V2 descendants through their recorded ancestors without starting model turns.
- Allow empty `continue_turn_if_idle` input to wake on saved results, guarded by the preceding turn ID. Keep mail queued until task registration so suspension can preserve it.

## Testing

Add coverage for completion and wake races, child errors, parent writer failures, evicted parents, descendant restoration, repeated suspension after caller cancellation, and ephemeral roots remaining active.

GitOrigin-RevId: 3041791f51bff2555210f6b2d2443f9f05f68330
2026-11-12 22:43:14 +00:00
Adam Perry @ OpenAI cdfcc18dba Clarify remote message board and clock documentation (#61197)
Describe remote board connection settings, host ownership, and configured
clocks in general terms. Update the remote board test credential to match
that terminology.

GitOrigin-RevId: c2c32808c98afad849f6e4f59f75d8d204f51694
2026-11-12 22:41:50 +00:00
ioanbudea-OAI 9908717ddb Preserve permission profiles across working directory updates (#61195)
## What changed

Stop rebuilding permission profiles through the legacy sandbox policy when only the working directory changes. Preserve the profile, its active selection, and profile workspace roots while updating runtime workspace roots for the new directory.

## Testing

Extend regression coverage to check that directory-only updates preserve the permission profile snapshot and expected thread settings. Verify that workspace-write permissions allow ordinary files in the new workspace while denying writes to `.git/config` and the previous workspace.

GitOrigin-RevId: a4f4b2dc4ffa16cbf33ddb124eed4b759b2aae78
2026-11-12 22:15:06 +00:00
neerajrj-oai d0221bc947 Keep plan item IDs unique after completion (#61186)
## Why

Plan items used a single ID derived from the turn ID, so multiple completed plans within one turn shared an ID across retries or tool continuations.

## What changed

Generate UUIDv7 plan item IDs and carry an unfinished plan's ID across model requests within a turn. Once a plan completes, assign a fresh ID to the next plan.

## Testing

Add unit and integration coverage for unfinished retries, retries after plan completion, and tool continuations, including matching streamed events to saved completions. Update the app-server test to check plan IDs against the start notification.

GitOrigin-RevId: 87b4b3e6bd74224ea96f4308f739bcd2efb62299
2026-11-12 21:40:33 +00:00
ioanbudea-OAI d4242793f7 Defer Guardian context limits to request preparation (#61185)
## What changed

Preserve original root conversation evidence, retained instructions, assistant context, and verified answers until request preparation. Apply shared section preparation to both synchronous and asynchronous reviews, preserving fixed limits and omission notices while allowing elastic preparation to use the originals. Cache the fixed root view for reuse.

Split classifier setup from final input assembly and normalize new images before deduplication and token accounting, preserving committed history.

## Testing

Extend tests for fixed and elastic preparation across both review profiles, original root evidence after handoff, and image normalization during retained-request preparation.

GitOrigin-RevId: 86c83384c1f5be3707622d675fc54b4c8e1ddc0a
2026-11-12 21:40:00 +00:00
swastik-oai e61a08f9ec Reuse Guardian reviewer history for direct reviews (#61163)
## What changed

Route direct reviews through the reusable Guardian session so sequential `thread/guardianReview` calls on the same loaded thread can share review history with live approvals. Keep the reviewer conversation separate from the actor and follow the parent thread's persistence settings. Retries reassess the final call under the current policy.

## Testing

Extend app-server coverage to verify history reuse within a parent thread and isolation between parents. Extend the core scenario and snapshot to cover sequential direct reviews without executing the recorded commands.

GitOrigin-RevId: 5f10a04aec43a9351e38575265cb6b5ff06530a5
2026-11-12 18:30:45 +00:00
ioanbudea-OAI ce3031018e Remove the legacy sandbox policy setter from Config (#61144)
## What changed

- Remove `Config::set_legacy_sandbox_policy` and update session, approval, and Guardian test setup to set permission profiles and workspace roots explicitly.
- Reuse `PermissionProfile::from_legacy_sandbox_policy_for_cwd` for legacy policy validation and updates in `Permissions`.

GitOrigin-RevId: 5ee92c46290ee6a3d7e2d2db7b02641e80bdc44f
2026-11-12 15:35:27 +00:00
Abhinav c0cc9fc820 Add refreshable headers for OTLP HTTP exporters (#61028)
## Why

Allow collector credentials to refresh without rebuilding OpenTelemetry providers.

## What changed

Add `otel.headers_helper`, a command run from `CODEX_HOME` that returns a JSON object of header strings. Share its results across explicitly configured HTTP log, trace, and metric exporters. Refresh every 29 minutes by default, configurable through `otel.headers_helper_refresh_interval_ms`.

Helper headers override matching static headers. Failed refreshes block affected exports until the helper recovers, and HTTP redirects are disabled when helper headers are used. Repository-local configuration cannot set the helper; exporters set by admin requirements, gRPC exporters, and built-in analytics do not use it. Helper execution respects network policy.

## Testing

Add tests for shared execution, credential refresh and recovery, blocked exports after failure, network restrictions, and configuration validation. Collector tests cover header precedence across all three signals using JSON and binary HTTP encodings, including exclusion of an admin-managed exporter.

GitOrigin-RevId: 89018ac09c2c99523d3f92bf77048e7ed9dae5f2
2026-11-12 05:21:30 +00:00
Alex Kira 64b8d019c8 Record reasoning replay removal metrics (#61017)
## What changed

Parse reasoning removals from response `input_transformations` and carry the counts through completion events to sampling telemetry.

- Add `codex.responses.reasoning_replay` to distinguish unavailable reports, no reported removals, prefix mismatches, and other removals.
- Add `codex.responses.reasoning_items_removed` histograms by bounded removal reason, grouping unknown or missing reasons under `other`.
- Include zero counts for usable reports; exclude missing, null, or malformed reports from histograms without failing the response.
- Use the sampling step's session telemetry when recording response events.

## Testing

Add SSE and WebSocket integration coverage for report availability, removal counts, unknown reasons, zero observations, and `product_sku` tags in exported metrics.

GitOrigin-RevId: 83c70ca6067fa4b34bf5c463958b4632a7a0a188
2026-11-12 04:31:25 +00:00
Ruslan Nigmatullin 77cbcdb4a6 Preserve thread-local model ID state across replay and compaction (#60942)
## What changed

- Initialize item and context-window ID generators per thread, reuse the stored window on resume, and give history reconstruction a separate item generator.
- Recover item ordinals from checkpoint replacement history, guardian history, and retained context without lowering the cursor established by earlier records.
- Stamp compaction resume metadata with the next item ordinal in append order, tracking the cursor in local and in-memory thread stores.

## Testing

Extend the history ordinal test to verify that a checkpoint records the cursor at its position in a batch, excluding later allocations.

GitOrigin-RevId: 7121b747a039dc5477235078702137d1ff74e789
2026-11-12 00:27:01 +00:00
felixxia-oai c95b72ed6f Reserve Guardian context budget for selected Code Mode exec results (#60914)
## What changed

Add `transcript_selected_exec_results` to synchronous Guardian elastic budgeting so selected Code Mode `exec` outputs can receive a separate token reservation and entry cap from ordinary tool evidence.

Configure selection with `nested_tool_names`. Match results against the host-recorded inventory of completed nested tool calls, including normalized tool names, even when the original `exec` call has left review history. Leave the new section's reservation at zero when omitted.

## Testing

Add coverage for recorded-inventory matching and results whose calls are missing from history. Extend the elastic budgeting scenario to verify that selected output is retained while ordinary tool output is capped.

GitOrigin-RevId: 6a8de250e36b18f2acfdab993c83c254f101ce87
2026-11-11 23:49:59 +00:00
sayan-oai 561f0f5cf0 Keep same-named MCP servers separate across environments (#60905)
## Why

MCP servers discovered on different executors can share a name. Resolving them by name alone exposes only one server and prevents tools and resources from being addressed independently.

## What changed

- Add the disabled-by-default `multi_environment_mcp` feature to expose environment-scoped servers, including multiple installations of the same plugin.
- Carry exact `McpServerKey` identities through catalogs, connections, tool calls, resources, and approval tracking. Give environment-scoped tools stable namespaces derived from their environment.
- Include `environment_id` in scoped resource listings and accept it in resource requests. Require it for environment-owned servers when multiple environments are selected.
- Preserve server ownership through hooks and use the owning environment's permissions for scoped elicitation review.

## Testing

Add coverage for same-named servers across environments, independent calls and approvals, resource selection, plugin discovery, and rejection of hook changes to the tool environment.

GitOrigin-RevId: 5593372b06a15c16aa3220faa3b19f5d3b20e67e
2026-11-11 23:36:19 +00:00
psachdeva-oai 8a4b226ad5 Version answer display instructions and restrict unsupported payloads (#60904)
## What changed

- Replace free-form `AgentMessageDisplay` actions with a version `1` contract using `show`, `hide`, and `block` modes, with typed annotations, replacements, block reasons, and trusted-contact handling.
- Discard invalid or unsupported display instructions and set `moderated` to `true` when reading moderation data in Rust and the Python SDK.
- Update generated JSON schemas, TypeScript types, and Python SDK models to match the contract.

## Testing

Add coverage for display variants, tolerance of extra fields, and unsupported-version or unknown-mode fallback. Update the final-answer moderation test and add Python coverage for thread history and item notifications.

GitOrigin-RevId: ec64070002da67e8e377dccb3728806ac8b290e7
2026-11-11 23:35:05 +00:00
Alex Kira b117b2d6fe Enable incremental tool declarations by default for GCP (#60848)
## What changed

- Add the stable `incremental_tools_legacy_gcp` feature, enabled by default, to control incremental tool declarations on regular GCP Responses independently of Amazon Bedrock.
- Mark `incremental_tools_legacy` stable and keep it disabled by default for Amazon Bedrock.
- Expose the new flag in the configuration schema.

## Testing

Update integration coverage for GCP defaults and opt-out, independent provider flags, and incremental tool declarations during remote compaction.

GitOrigin-RevId: 7acf085789d44122fd2a5da97932a8b9e93e240e
2026-11-11 19:18:14 +00:00
Charlie Marsh ed7ca22bb6 Honor advertised WebSocket message size limits (#60841)
## Why

Requests that exceed the server's advertised WebSocket message size limit can fall back to HTTP without first uploading the oversized message.

## What changed

Read `x-codex-websocket-max-message-bytes` from the handshake response and check the serialized request's byte length before sending. Route oversized requests through the existing HTTP fallback. Ignore missing or invalid limits, and remove the close-code-specific wording from the size-limit error.

## Testing

Add a fallback test that advertises a one-byte limit and verifies zero WebSocket generation requests and one HTTP request.

GitOrigin-RevId: 03b42012ca77d6b046e937fef8e6335d855a1c10
2026-11-11 18:58:56 +00:00
Charlie Marsh bc3ccba6e1 Fall back to HTTP immediately for oversized WebSocket requests (#60840)
## Why

A WebSocket close with code `1009` before any text arrives indicates that the request exceeds the peer's message size limit. Retrying the request over another WebSocket does not address that limit.

## What changed

- Classify these early size rejections as requiring HTTP fallback and switch transports without exhausting WebSocket retries, including during prewarming.
- Preserve ordinary stream retry behavior for size closes after any text frame and for other close codes.

## Testing

Add integration coverage for early size rejections during generation and prewarming, size closes after metadata, rate limits, or response creation, and ordinary closes.

GitOrigin-RevId: 4550d289bb414fec1faae2a195bcc635d394117b
2026-11-11 18:58:21 +00:00