Commit Graph
383 Commits
Author SHA1 Message Date
jif 02d60fb5cc Preserve client attribution when spawning subagents (#62072)
## Why

Delegated work should retain client attribution such as `cost_center` from the turn that requested it, even when a yielded code-mode cell spawns its child after a newer parent turn starts.

## What changed

Capture client metadata from the originating turn and apply it to the child's initial turn in both legacy and v2 agent spawning. Children and grandchildren retain the captured values when the parent is later steered with updated metadata.

Exclude `misalignment_override`, `parent_response_id`, and `mcp_request_meta` from inheritance, while preserving each child's own thread and turn identifiers.

## Testing

Add coverage for metadata filtering and snapshot isolation, legacy and v2 spawning with fresh or forked history, grandchild inheritance, and spawning from a yielded code-mode cell after the parent advances to another turn.

GitOrigin-RevId: 8c378c49239ccfac03deb2b793fc69aba1d41644
2026-11-17 16:15:55 +00:00
jif 58627e6fea Fix turn interrupt reply ordering and shutdown handling (#62066)
## Why

A delayed terminal event from an earlier turn could acknowledge interrupts for a later turn. Interrupts could also remain pending after the thread listener shut down.

## What changed

- Associate pending `turn/interrupt` requests with their target turn and reply only when that turn completes or aborts.
- Track the current Core turn through stop hooks and resolve review delegate IDs to the owning turn when validating interrupts.
- Settle pending interrupts with an error on listener shutdown and reject further interrupts against the retired runtime.

## Testing

Add regression tests for delayed terminal events, queued turn starts, review delegates, and interrupt handling during thread revert and runtime replacement.

GitOrigin-RevId: 9695fc3f718baa6b792a14d54136bb87fc003100
2026-11-17 15:33:21 +00:00
Adam Perry @ OpenAI d8d840d32a Capture local attachments before queuing turn input (#61962)
## Why

Queued input can outlive its local attachment files. Reading those files only when rendering a model message can lose media submitted while turn admission is pending.

## What changed

- Introduce `ResolvedUserInput` to carry captured image bytes, audio data URLs, or read errors through session queues and agent input.
- Resolve client attachments on the submitting host before queuing, using a blocking task for local file reads.
- Render messages from captured payloads while preserving original paths and text spans for display, along with image position mappings.

## Testing

Add coverage for rendering and serialization after source files are removed, steered audio retained during pending admission, and local audio errors and display content.

GitOrigin-RevId: 4bbb6bda44167439a8cfdae15d3c476a7b1d336b
2026-11-17 04:08:00 +00:00
jif f8e58064fc Preserve remote compaction summaries when interrupting child agents (#61916)
## Why

Child agents configured to interrupt at item boundaries should retain the summary when interrupted during remote compaction. An interrupt during post-turn compaction should also preserve the answer already completed.

## What changed

- Keep sampling interrupt registration active through summary installation, then honor the pending interrupt without running post-compaction hooks.
- Deliver an already completed answer when post-turn compaction is interrupted.
- Keep queued follow-ups for a new turn instead of continuing the interrupted turn.

## Testing

Add a streaming scenario test and context snapshot covering interrupts during mid-turn and post-turn remote compaction. Verify that resumed requests retain the summaries, the completed answer is delivered, and each queued follow-up starts a separate turn.

GitOrigin-RevId: 939eb8b3ed24dbdbff10c9f62226e38a0198fd6f
2026-11-16 23:58:12 +00:00
viyatb-oai c9df641eac Preserve native sandbox executable paths and PTY dispatch names (#61910)
## Why

Sandbox helpers with non-UTF-8 paths can lose their exact executable path when converted into command requests. PTY launches with a custom `argv[0]` can also try to execute the dispatch name instead of the configured program.

## What changed

- Carry native executable paths through sandbox requests, local execution, shell snapshots, and sandbox integrity checks.
- Use native Unix PTY spawning when `argv[0]` differs from the executable, preserving the dispatch name and controlling terminal setup.
- Preserve PTY executable lookup through the child environment's `PATH` and fallback to its home directory when the requested working directory is missing.

## Testing

Add regression coverage for non-UTF-8 sandbox helper paths, PTY dispatch names, child `HOME`/`PATH` lookup, controlling terminals, and exit-code preservation.

GitOrigin-RevId: 6a2af0486220d98195f5d67fde2597e3c4515e11
2026-11-16 23:26:42 +00:00
jif cff061fca9 Add opt-in item-boundary interrupts for subagents (#61909)
## What changed

Add the multi-agent v2 setting `interrupt_at_item_boundary`, disabled by default. When enabled, `interrupt_agent` lets a subagent finish and retain its current output item while cancelling tools immediately. Later output items are discarded, and retained tool calls that have not run receive an aborted result.

Wait for the interrupted turn's terminal event before starting follow-up input, and prevent idle contributors from restarting the interrupted child.

## Testing

Add streaming tests for retaining an assistant message or tool call, discarding later output, and preserving history and follow-up input when the child resumes. Update configuration and interrupt-routing tests.

GitOrigin-RevId: 8b2b9b1383bfb4eefea7369a40012bebc5fff07b
2026-11-16 23:26:07 +00:00
jif e5839cbc59 Preserve aggregate turn diffs across steering (#61820)
## Why

A steer can restart sampling within the same turn. Recreating the diff tracker loses earlier edits, and changing the working directory can change the meaning of relative diff paths.

## What changed

Keep the diff tracker for the lifetime of the turn and retain each environment's first display root. Refresh cached diff headers when new environments are added. Emit the final aggregate `TurnDiff` before `TurnComplete` instead of at sampling completion.

## Testing

Extend regression coverage for steering during the final flush, with and without additional patches and with a changed working directory. Check final diff ordering across continued sampling and header updates when adding an environment.

GitOrigin-RevId: fbd1ebef072d8486a324190c9125fc1446510802
2026-11-16 14:18:24 +00:00
tongzhou wang ce82e38b16 Gate new rollout work on shared budget exhaustion (#61582)
## Why

A budget can be exhausted while another thread waits to start work. New requests, tools, and automatic turns must observe that stop without discarding an already-consumed response whose cost exhausted the budget.

## What changed

- Share a work gate across the rollout tree while preserving native and host accounting ownership.
- Recheck the budget before HTTP and WebSocket requests, hook execution, tool dispatch, and compaction. Prevent automatic wakeups and late input from restarting stopped work, leaving pending input queued.
- Record response costs without rejecting the consumed response solely because its bill exhausts the budget, and stop new work when enforced accounting overflows.
- Expose a thread-registration receipt that remains committed after unload.

## Testing

Extend rollout-budget coverage to verify that accepted response text remains in history after exhaustion and that waiting tools observe the shared stop while completed tools retain their results.

GitOrigin-RevId: e583d657c6f29e374d374cdb014bf23fd6c28614
2026-11-15 00:03:23 +00:00
tongzhou wang 2e61e8ec4e Record neutral interruption markers for rollout budget stops (#61581)
## Why

Turns stopped by an exhausted rollout budget need a history marker without attributing the interruption to the user.

## What changed

- Record and flush a `<turn_aborted>` marker before the terminal event for `SessionBudgetExceeded`, respecting the interrupt-message setting.
- Use neutral guidance in a `developer` message for automatic budget stops.
- Centralize tool-start budget checks across parallel dispatch, tool handlers, and MCP handoff, retaining the check after tool-start observers finish waiting.
- Share interruption classification and history recording across turn and compaction paths.

## Testing

Extend the rollout-budget integration test to assert that both the budget-crossing turn and a later denied turn persist `developer` interruption markers.

GitOrigin-RevId: 2a49ac9824a4e95e4799f2d0f638f8c46889e8e8
2026-11-15 00:02:01 +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
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
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
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
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
sayan-oai c1934fc99e Resolve saved Daybreak choices for each subagent turn (#60773)
## Why

A subagent can use a different model from its parent, with different supported access programs. Saved Daybreak choices need to resolve against the subagent's model and remain consistent throughout its turn, including compaction.

## What changed

- Resolve saved choices before native subagent turns start, preserving explicit turn overrides and applying saved toggles on later turns.
- Distinguish inherited parent programs from explicit child selections. Adapt inherited Daybreak Blue/Red selections to the child model while retaining the saved choice as the default for later turns and nested subagents.
- Validate the account associated with automatic selections and reject incompatible saved choices before sending model requests, while retaining rejected user input in conversation history.
- Require model changes to wait until the next turn when a saved Daybreak choice is present.

## Testing

Add coverage for both multi-agent spawn versions, model-specific selection, automatic compaction, WebSocket response reuse, account changes, unsupported choices, and rejected-input persistence across resume.

GitOrigin-RevId: 11971138d656600d9878187225e4d33217e68ad4
2026-11-11 10:10:37 +00:00
jamy-OAI d9a32c1789 Capture saved Daybreak preferences for native helpers (#60749)
## What changed

Add the disabled-by-default `daybreak_program_propagation` feature to capture a parent's saved Daybreak choice when spawning native helpers. Retain the choice and account identity for helper turns and nested spawns, while preserving explicit `cyber_access_program` overrides. A helper's own saved preference takes precedence on later turns.

Read pending metadata before persisted metadata so preferences survive thread materialization, with stored choices taking precedence. Share this resolution logic with `thread/read` and support refreshing the feature through user configuration.

## Testing

Add coverage for saved and pending preferences, unset choices, explicit overrides, failed metadata reads, nested spawns, parent unloading, feature withdrawal, and preference changes between helper turns.

GitOrigin-RevId: b3b349da8cb3d48154cddf058890587385efcda2
2026-11-11 07:39:04 +00:00
Dmitriy Kovalenko 3f797555cd Refresh Apps tools outside the MCP runtime gate (#60639)
## Why

Holding the session MCP gate during `tools/list` blocks other runtime access. Releasing it requires preventing stale refresh results from publishing after configuration or authentication changes.

## What changed

- Capture the Apps client under the session gate, then release the gate before fetching tools.
- Validate runtime and authentication state before publishing the shared cache and client catalog; retry a superseded refresh once.
- Wait for in-flight Apps refreshes before capturing model tools, preparing Apps calls, or shutting down the runtime.

## Testing

Add regression coverage for stale refreshes after runtime invalidation or logout, publication validation after call preparation drains, and a model step that waits for refreshed Apps tools while MCP status queries remain available.

GitOrigin-RevId: 42d4be8c4a9218933f8e3d07fc7f32f2da9cbd9d
2026-11-10 18:33:49 +00:00
Charlie Marsh aa39ded153 Trace MCP refresh causes and plugin metadata changes (#60505)
## What changed

- Track and log `McpRefreshReason` values for runtime refreshes, including invalidations received during thread startup and forced refreshes. Restore claimed reasons when a refresh is cancelled before publication, while preserving concurrent invalidations.
- Correlate installed-plugin metadata invalidations with downstream MCP refreshes using a shared change ID. Log changed field names by plugin ID, additions, removals, and reordering without logging field values.
- Record whether background installed-plugin bundle updates were requested by startup, plugin listing, or installation.

## Testing

Add coverage for metadata change attribution across membership changes and reordering, and for retaining refresh causes across cancellation and concurrent invalidation.

GitOrigin-RevId: 3d4485ccfcc1d8788eb9909a060ac200c467af13
2026-11-10 02:45:30 +00:00
rhan-oai f69c660a83 Add host reporting for terminal agent turn outcomes (#59725)
## What changed

Expose `AgentTurnOutcomeReporter` for hosts to install through `StartThreadOptions::thread_extension_init`. Await its callback after emitting the terminal turn event, independently of native parent notifications, so hosts can handle result delivery and retries.

Track outstanding outcome handoffs and wait for them during graceful shutdown before running thread-stop hooks. Prevent new turns from starting once handoff registration closes.

## Testing

Add integration coverage for completed and errored outcomes, turn lineage, queued follow-up turns, and shutdown waiting for handoffs. Extend Guardian tests to verify that strict safety stops reach the reporter while default interruptions remain silent.

GitOrigin-RevId: 0d825f6208f0250969734488881f32fd1dcd7509
2026-11-05 19:20:56 +00:00
rhan-oai f3074e35c0 Render host-owned V2 subagent rosters in environment context (#59531)
## What changed

- Expose `AgentSessionReader` and `AgentRef` for read-only access to host-owned agent session membership.
- Read direct children when rendering V2 environment context, using the existing roster size limits. Refresh membership when rebuilding context after compaction, and omit the roster if the provider is absent or the read fails.
- Preserve unrecorded initial input across suspension or interruption while waiting for a roster read. Resolve the roster before advancing the compaction window.

## Testing

Add coverage for roster refreshes after compaction, unchanged environments avoiding extra reads, unavailable rosters, recoverable input during blocked reads, and stale compaction results rejected after a concurrent history replacement.

GitOrigin-RevId: 3c7e3822af09e9295d405e53dabb87b020c8f00c
2026-11-05 00:39:39 +00:00
Henry Levy d78f68f8d1 Allow per-turn overrides for direct-only tool namespaces (#59481)
## What changed

- Add `direct_only_tool_namespaces` to `TurnStartOptions`. Omission inherits the thread configuration; an empty list clears direct-only routing for that turn. Steering ignores the option.
- Persist explicit overrides in turn context and restore them when recovering an interrupted turn, including turns started by standalone tool output.
- Keep subsequent turns and spawned or resumed child agents on the thread's routing defaults.

## Testing

Add coverage for routing changes across turns in Responses and Responses Lite, recovery without carrying overrides into subsequent turns, and child-agent inheritance.

GitOrigin-RevId: cea1a997424572d4cd372cd15b098626904c6796
2026-11-04 19:58:49 +00:00
Rohan Mehta 41cbd5b9b2 Add per-turn prompt cache TTL options (#59480)
## What changed

- Expose `PromptCacheTtl` through `TurnStartOptions` and `RecoverTurnRequest`, supporting `30m` and `12h` minimum cache lifetimes.
- Carry the selected TTL through turn execution and compaction into `prompt_cache_options.ttl` on Responses HTTP and WebSocket requests. Omit the field when unset and keep the selection scoped to the turn.
- Include prompt cache options in WebSocket request reuse checks.

## Testing

Add coverage for both TTL values and omission on a subsequent turn without an override, and extend WebSocket serialization coverage to include prompt cache options.

GitOrigin-RevId: 8529306e6ad6030fe77c5e6bf19f3810717c0c5e
2026-11-04 19:58:24 +00:00
Ruslan Nigmatullin d84c872464 Route response item IDs through a shared generator (#59312)
## What changed

Introduce `ItemIdGenerator` and pass the runtime-owned generator through sessions, model clients, history, hooks, tool outputs, and memory requests. Keep UUIDv7 generation as the default and replace direct calls to `ResponseItemId::new` with the shared callable.

Preserve existing item IDs and backfill missing or empty IDs through the same generator, including forked history before reconstruction and persistence.

## Testing

Extend coverage for UUIDv7 IDs on user messages, agent messages, and additional tools, and assert that forked model context matches the backfilled items persisted to history.

GitOrigin-RevId: 0dd9ccf0f46f0f1005a6acfc740133b03ce82259
2026-11-04 04:02:00 +00:00
taohong-oai db14d7508d Report auto-review overhead in turn critical-path analytics (#59260)
## What changed

Track Guardian compaction and review turn execution within each tool's auto-review interval, and expose `critical_path_auto_review_overhead_ms` in turn analytics. This measures the portion of `auto_review_ms` outside observed Guardian turn execution; it is a subset, not an additional duration.

Freeze review observations when the review or tool invocation ends. Report unknown overhead as `null`, including when denial reconsideration runs an unobserved follow-up turn.

## Testing

Add coverage for shared compaction and review observations, cached reviews, unfinished reviews, reused tool call IDs, and execution timing with child critical-path capture enabled or disabled. Extend analytics and app-server tests to check the new field.

GitOrigin-RevId: a63dec2a75571a3167c8c3f0c1d21e11e4a8d050
2026-11-03 23:49:01 +00:00
taohong-oai d5830f6fc6 Expand per-turn critical path capture and split inference timing (#59245)
## What changed

- Report sampling and compaction separately as `critical_path_sampling_ms` and `critical_path_compaction_ms`, retaining `critical_path_inference_ms` as their combined compatibility total.
- Enable critical path capture for subagent and Guardian turns when `CriticalPathTiming` is enabled, while continuing to exclude memory consolidation sessions.
- Preserve capture across incoming agent messages and support direct `spawn_agent` and `wait_agent` calls. Attribute these calls to the parent's tool time and keep child work on the child's own turn.

## Testing

Add runtime coverage for independent parent/child timing, incoming agent messages, Guardian turns, feature gating, and memory consolidation exclusion. Update phase accounting and analytics tests to cover the new fields and millisecond truncation.

GitOrigin-RevId: 09d4c0b853e817216bed173e4404d0d782de97ec
2026-11-03 22:29:33 +00:00
Hung f1b0721c88 Enable uploaded audio with durable rejection recovery (#59234)
## Why

Uploaded audio was rejected even for audio-capable models. Invalid or expired recordings also need to be excluded from later requests so users can continue the conversation.

## What changed

- Accept uploaded audio when the selected model supports audio; otherwise, preserve valid sibling input and report the unsupported attachment before inference.
- Track successful audio decoding by model and provider, and attribute provider rejections against the exact submitted request. Preserve attachments when attribution is ambiguous.
- Persist confirmed rejected file IDs across compaction and resume, omitting them from future requests without rewriting retained user history.
- Check audio availability before remote compaction and retry once without an unambiguously rejected recording. Include guidance to remove or replace affected attachments in audio errors.

## Testing

Add unit and scenario coverage for uploaded audio across follow-up turns, compaction, cold and paginated resume, model switching, ambiguous rejections, expired files, and queued or interrupted input.

GitOrigin-RevId: e701901e7b1c0865c9f2a5918197215f549b06c0
2026-11-03 21:24:16 +00:00
bkotsopoulos 66e3b5a02c Record turn critical path metrics and trace fields (#58980)
## What changed

- Add critical path capture status and duration fields to `codex.turn_profile` traces.
- Emit capture counts for complete and incomplete profiles, and total and component duration histograms for complete profiles, tagged by product SKU and reasoning effort. Record metrics once per turn timing state and skip disabled captures.
- Extend critical path histogram boundaries up to two hours to cover long turns.

## Testing

Add coverage for duplicate reporting, disabled and incomplete captures, product SKU normalization, trace fields, component durations, and extended histogram boundaries.

GitOrigin-RevId: 5b1af9255f1c4a376b61c81eb68c7a74b01ae9ae
2026-11-02 20:57:23 +00:00
rhan-oai d893b14475 Add host-owned turn admission to the agent runtime (#58963)
## What changed

Add `AgentRuntimeOwnership::Host` and an `AgentTurnAdmission` provider so hosts can acquire execution capacity before runtime tasks run, including root turns. Require the provider at startup and retain its execution guard for the task lifetime. Controller ownership remains the default; host ownership skips controller rollout-budget accounting, budget reminders, and subagent roster context.

Keep pending admission interruptible without blocking input submission. Report admission failures through the turn lifecycle, record accepted input before exposing the error, and unwind task setup such as review mode. Preserve recoverable input during suspension. On interruption, skip prompts that still require a `UserPromptSubmit` hook rather than recording them without inspection.

## Testing

Add integration tests for missing-provider startup cleanup, denial and immediate retry, steering during admission, cancellation, execution guard lifetime, submit-hook handling, suspension recovery, and review-mode cleanup. Add a context snapshot verifying retained input reaches the next admitted turn.

GitOrigin-RevId: 9ee85fad2ecfbfb38735bac8ef88888fb179a967
2026-11-02 20:10:37 +00:00
Charlie Marsh ba0d0e1f74 Handle steers accepted during the final transcript flush (#58877)
## Why

Input accepted while a regular turn was flushing its transcript could miss the final pending-input check and remain unprocessed when the turn completed.

## What changed

Check pending input atomically with closing steer admission. If input was accepted, resume sampling in the same turn while preserving MCP startup requirements, and complete the turn with the updated answer.

## Testing

Add a regression test that pauses the final flush, submits a steer, and verifies that the model receives it exactly once before the same turn completes with the answer to the steer.

GitOrigin-RevId: 193db3ab6916a86e97c3845590598fd8fbe47081
2026-11-02 12:24:38 +00:00
Hung 2546265e3e Add audio file references with explicit runtime rejection (#58430)
## What changed

- Allow audio inputs to carry uploaded file references via `fileId` in the app-server API and `file_id` in model content, preserving existing inline audio shapes and Python class names.
- Preserve mixed inline and file-backed audio ordering through history reconstruction and migration, and budget file-backed audio conservatively when duration is unknown.
- Keep uploaded audio disabled at runtime: reject it before inference or raw history injection, preserve accompanying text, and prevent rejected attachments from reaching follow-up or resumed requests.

## Testing

Add coverage for wire-format round trips, SDK compatibility, mixed audio history ordering, and rejection during turn start, steering, task completion, and raw injection.

GitOrigin-RevId: 70895c47a2494c0c9936348ad1250eb1b7b892a6
2026-10-30 20:53:25 +00:00
Dmitriy Kovalenko 7a030b9769 Fix context token estimates for responses without usable usage (#58404)
## Why

Retained output from failed responses or responses without usage data can be omitted from context token estimates, delaying automatic compaction even as reasoning accumulates.

## What changed

Track the history boundary covered by the last usable usage report and estimate all items added after it. Spending-only updates preserve that boundary, while successful usage reports cover the request input even when the response contains no output items. Avoid counting unreported reasoning twice and align recomputed context counts and remote compaction estimates with the boundary.

## Testing

Add unit coverage for initial input, empty responses, spending-only updates, and reasoning accounting. Add regression coverage verifying that accumulated unreported reasoning triggers compaction for incomplete responses with or without usage and completed responses without usage.

GitOrigin-RevId: 80c0b800c7236ff40ae53fdd18320e0d06e026ab
2026-10-30 18:12:39 +00:00
andrewgu-oai c62e9df39f Preserve reasoning token accounting across resume and fork (#57839)
## Why

Resuming or forking a thread loses the session flag indicating that reported token usage already includes retained reasoning. Counting that reasoning again can trigger automatic compaction too early.

## What changed

Persist `last_usage_includes_reasoning` with `TokenUsageInfo` and use it when estimating context usage. Record the flag with completed sampling and compaction usage, and mark local estimates as including reasoning. Older records without the field retain the existing accounting behavior.

## Testing

Add coverage for saved inclusive, exclusive, and legacy token counts. Extend the encrypted-reasoning compaction test to verify the compaction threshold for live, forked, and cold-resumed threads.

GitOrigin-RevId: 9f22262a693111ec0406dce11804eb4844ad8983
2026-10-28 21:22:06 +00:00
taohong-oai f371f9701f Capture direct-tool critical paths behind a feature flag (#57201)
## What changed

Add the default-disabled `critical_path_timing` feature to capture dependency paths for root-agent turns. Connect direct tool dispatch, execution-lock handoffs, and completed tool batches to turn timing so overlapping work shares its common history without double counting.

Omit captures for unsupported dependencies, including CodeMode, agent communication, approval review, and user waits. Settle aborted tasks before freezing timing, invalidate captures on forced teardown, and apply runtime feature changes only to future turns.

## Testing

Add coverage for serial and overlapping tool dispatch, execution-lock dependency handoffs, cancellation, unsupported paths, frozen timing after completion, and configuration refresh boundaries.

GitOrigin-RevId: 71458b654b1c4a05799886928b449c948ea33109
2026-10-26 19:19:10 +00:00
bkotsopoulos b82d542aef Add turn, tool, and skill attribution to traces (#56524)
## What changed

- Emit `codex.turn_profile` events with finalized sampling request and retry counts when turns complete or abort, and preserve the caller's tracing span for detached extension interrupts.
- Add conversation and turn identifiers to turn and tool dispatch spans, identify tool sources, and include dispatch-to-result-ready duration on tool completion events.
- Emit `codex.skill_read.source_selected` before cache or provider access, recording the selected cloud or executor source, skill and plugin identity, resource kind, and continuation status.

## Testing

Add regression coverage for trace correlation, tool completion timing, turn profile counters on completion and interruption, detached interrupt span propagation, and skill source attribution including cached cloud reads.

GitOrigin-RevId: db78aa270bbcb8dc9e4558d202c109ad54141811
2026-10-23 16:55:32 +00:00
bkotsopoulos 810501d60d Add trace instrumentation for turn startup and response streaming (#56515)
## What changed

- Record time spent waiting for each stream event in `codex.response.receive_wait_ms` on the existing `handle_responses` span.
- Add trace spans for startup prewarm consumption, turn-start lifecycle callbacks, assistant text flushing, and response stream error handling.
- Include the turn ID and lifecycle phase on the turn-start span and the request type on the error-handling span.

GitOrigin-RevId: 7792508c830ef05b03aed01eb840ca7e6577ad0b
2026-10-23 16:09:06 +00:00
alishobeiri-oai 8531184b49 Make agent-message wakes respect interrupts and Plan mode (#56330)
## Why

A pending agent-message wake can outlive an interrupt or a switch to Plan mode while model discovery is waiting. It must recheck whether it can start before consuming queued messages.

## What changed

- Revalidate reserved wakes after discovery and build the turn with the configuration selected during admission.
- Mark interrupted wake reservations as stopped so later agent messages cannot restart automatic work, while preserving queued messages for explicit turns.
- Keep finishing turns marked as running until completion is persisted, protecting their rollout writer from capacity eviction.

## Testing

Add regression tests for interrupts and Plan-mode changes during discovery, queued-message preservation, explicit follow-ups with `wake_on_agent_messages` disabled, and completion status and persistence with the feature enabled and disabled.

GitOrigin-RevId: afe1d57063a5ec491e8b16d6c62135caa69c0a3c
2026-10-23 02:21:49 +00:00
Dmitriy Kovalenko 07cc649c65 Avoid double-counting retained reasoning in token usage (#55991)
## Why

Responses Lite includes retained reasoning in its reported token usage even without an `X-Reasoning-Included` header. Adding a separate reasoning estimate can trigger automatic compaction prematurely. Locally recomputed usage also already includes retained reasoning.

## What changed

Initialize reasoning accounting from `use_responses_lite` before each sampling or compaction request, while allowing server events to override it. Mark locally recomputed token usage as including reasoning.

## Testing

Update the compaction regression test to verify that Responses Lite usage without the header stays below an 81-token limit at 80 tokens and triggers compaction at 81 tokens.

GitOrigin-RevId: 06f9272cb3c259e927941f1d2a1d7096310ffb2c
2026-10-21 21:23:01 +00:00
ioanbudea-OAI 60d78332fc Measure Guardian approval blocking time per turn (#55591)
## What changed

Record `codex.guardian.turn.blocking_ms` when a turn completes or aborts, with an `outcome` label. Count overlapping approval waits once, exclude idle gaps, and split pending reviews across the active turns they span. Use each turn's telemetry so samples retain the servicing turn's model attribution.

## Testing

Add unit coverage for overlapping waits, idle gaps, late turn completion, and duplicate stops. Extend integration coverage to verify attribution across turns and model changes for yielded command reviews and MCP elicitation, including aborted turns.

GitOrigin-RevId: a3e6a0c32a06fe58ea93119ecd941076100fb82c
2026-10-20 14:43:08 +00:00
alishobeiri-oai 8c076942cb Allow agent messages to wake idle root sessions behind a feature flag (#54961)
## Why

A worker can finish after its parent has returned a final answer. Agent messages without `trigger_turn` previously needed an outstanding durable sleep to wake an idle session.

## What changed

Add the disabled-by-default `wake_on_agent_messages` feature to start a new root turn when agent mail arrives through direct delivery or the host mailbox. Reuse the previous turn context for messages without `trigger_turn`. Keep these automatic wakes disabled for interrupted sessions, Plan mode, and child agents.

## Testing

Add coverage for both delivery paths waking a completed root exactly once, child completion waking a parent after its final answer, and the interruption, Plan mode, and child-agent boundaries.

GitOrigin-RevId: f1a4fc7c22be8ca46b2173a85445126cbdd6c4b6
2026-10-18 01:58:02 +00:00
Benjamin Carlsson 7eed30ad1b Preserve running turn recovery across thread settings updates (#53879)
## Why

Standalone thread settings updates cleared the running turn's continuation lineage, preventing recovery after a daemon restart.

## What changed

Preserve continuation lineage while a task is registered. Continue invalidating pending continuations for idle, reserved, and finishing turns, and record started-turn lineage atomically with task registration.

## Testing

Extend settings invalidation coverage to idle, reserved, and finishing turns. Exercise an in-flight reasoning effort update before compaction, then verify cold resume preserves the turn's frozen settings and allows continuation. Wait for thread idle before testing idle continuation.

GitOrigin-RevId: d7c316af6a4c871790bb57f0af9892bb30871ac8
2026-10-14 06:03:02 +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
jif 816fa702ec Account for token usage from incomplete responses (#53670)
## Why

Terminal `response.incomplete` errors discard reported token usage. Counting that usage as a new context baseline can also double-count input and trigger unnecessary compaction on the next turn.

## What changed

- Carry the response ID and valid reported usage with incomplete-response errors.
- Record usage for sampling and foreground compaction in cumulative totals, rollout records, budgets, and token-usage contributors without replacing the context baseline. Preserve the original terminal error, including when accounting exhausts the budget or cancellation races the error.
- Charge incomplete background compaction requests to the rollout budget without publishing foreground token usage.
- Ignore missing or malformed usage rather than estimating it.

## Testing

Extend coverage for SSE and WebSocket incomplete responses, rollout usage records, compaction accounting, deferred token-count events, and token-usage contributors. Add regressions for unnecessary compaction on the next turn and background compaction publishing foreground usage.

GitOrigin-RevId: bf1ee6df3912052ab7c19e1c6af8ad3dc4b07675
2026-10-13 13:36:24 +00:00
Eric Traut 7a0a8ac9c4 Close input admission while a turn finishes recording (#53452)
## Why

Inputs arriving during turn finalization could be queued after the final batch had already been drained, or recorded ahead of that batch.

## What changed

Mark active turns as finishing before draining pending input. Reject injection into finishing turns and make late client responses and notifications wait until the final batch is recorded before appending to history. Continue accepting input into turns reserved but not yet started.

## Testing

Extend the task-completion test to block final recording, verify that input admission is closed, and confirm that a late notification is recorded after queued input while preserving the notification truncation budget. Strengthen the retained-tool-output scenario to assert tool output names and ordering.

GitOrigin-RevId: a57e44bd974e60e134183dab989795e2c3538596
2026-10-12 18:54:22 +00:00
pakrym-oai 5ef96ab278 Add opt-in output token replay for OpenAI requests (#52742)
## What changed

- Add the disabled-by-default `output_token_replay` feature to request `output.encrypted_content` from OpenAI providers.
- Preserve encrypted message and tool-call output, message and function-call status, and output-text `annotations` and `logprobs` through history and replay. Update the generated protocol schemas and SDK types.
- Strip output ciphertext from requests to other providers while retaining it in stored history, including after replay is disabled. Remove encrypted content from memory extraction inputs.

## Testing

Add integration coverage for the feature/provider combinations, metadata replay across follow-up requests, and history preservation when resuming with replay disabled.

GitOrigin-RevId: 77729a021b6771933402aebe74ece8eb39c057e2
2026-10-10 01:25:00 +00:00
pakrym-oai e79c498b5e Preserve tool declaration mode across resumed context windows (#51480)
## Why

Changing the incremental tools setting on resume should not move tool declarations and base instructions within an existing context window.

## What changed

- Determine the current window's tool declaration mode from its history, and apply the incremental tools setting when building a new window after compaction or a context reset.
- Use that mode consistently for sampling, remote compaction, startup prewarm, prompt debugging, and Guardian input budgeting.
- Preserve tool declarations alongside their comparison baseline in full-history agent forks, including compacted histories; remove them together when rebuilding context.

## Testing

Add request snapshots covering legacy resume, migration after remote compaction or a context reset, disabling incremental tools, and tool updates after restart. Extend agent fork and WebSocket startup prewarm tests to cover retained declarations and incremental tools.

GitOrigin-RevId: 7c46849a1788027a8435082e4d1f7e7f48d2700c
2026-10-06 21:28:13 +00:00
jif 0ef50c9f66 Prevent invalidated wakeups from starting a turn (#51427)
## Why

An interrupt or turn replacement can invalidate a pending-work wakeup while its turn context is being constructed. Task startup could then recreate a turn whose reservation was no longer active.

## What changed

- Recheck that the wakeup still owns the active turn reservation immediately before calling `start_task`, and return if it does not.
- Allow the remote-control re-enrollment test up to 45 seconds for the connection following reconnect backoff.

## Testing

Add a regression test that pauses wakeup context construction, processes an interrupt, and then resumes construction. It verifies that no additional turn starts and only the initial model request is sent.

GitOrigin-RevId: c857fdd9f0fac3790d35a95a1f21a6d41977bb29
2026-10-06 17:26:19 +00:00
Eddie Chen f1dd04959c Preserve turn attribution when queued mail wakes durable sleep (#51419)
## Why

Queue-only agent mail can wake a sleeping session without starting a new task. The wake turn needs to retain the previous turn's attribution so delegated work reports completion to the initiating agent, including after a reload.

## What changed

Capture the previous turn's attribution while reserving the pending-work turn. Reuse its trigger, parent and root turn IDs, and initiating agent path for queue-only wakes. Mail that explicitly starts a new task continues to supply its own attribution.

## Testing

Extend regression coverage for live and reloaded sleep wakes, delegated and unknown attribution, settings changes while asleep, and subsequent task-triggering mail. Verify that a reloaded sleeping worker routes completion to the initiating turn.

GitOrigin-RevId: 28e38474028bdbf35a0538c1202165f056bb3dc3
2026-10-06 17:09:56 +00:00
Eddie Chen cfc946f4e1 Expose and persist turn lineage across the app server (#51415)
## Why

Clients starting delegated work need to attribute descendant turns to the original turn and retain that attribution through completion and history reads.

## What changed

- Accept optional `parentTurnId` and `rootTurnId` in `turn/start`, and expose `rootTurnId` on returned turns and lifecycle notifications.
- Default new turns to their own root when none is supplied. Preserve the active turn's root when steering, ignoring incoming lineage fields.
- Carry root IDs through completion, interruption, recovery, and fork boundaries, and persist them for paginated history reads. Leave unknown roots null in older history.
- Update protocol schemas and generated TypeScript and Python bindings.

## Testing

Extend coverage for default and explicit roots, steering without changing lineage, interrupted and recovered turns, history reads after restart in both history modes, nested fork pagination, and compatibility with older database writers.

GitOrigin-RevId: f1a570a12bd11742929fd2e8b017e5d35aa07dcb
2026-10-06 16:56:39 +00:00
Eddie Chen 551bd409eb Preserve turn attribution across recovery and compaction (#51402)
## Why

Recovering an interrupted turn replaced its original trigger with `retry` and restored only its root lineage from model context. Turns suspended before that context was written could lose their attribution entirely.

## What changed

- Persist regular-turn attribution in `TurnStartedEvent` and compaction resume metadata, including the trigger, parent and root turn IDs, and initiating agent path.
- Restore matching attribution during recovery, keep unknown triggers absent, and retain the model-context root fallback for older records.
- Reconstruct attribution from surviving turns across rollback and late completion or abort events, and clear inherited attribution when spawning a forked agent.

## Testing

Add coverage for repeated cold recovery, suspension before context persistence, legacy rollouts, fresh attribution on subsequent turns, compaction checkpoints, rollback, and late terminal events.

GitOrigin-RevId: 85b5763e8fbd48a1ea2c07cfda8196e63c2cd33a
2026-10-06 15:50:27 +00:00
Charlie Marsh 685270a56a Make apply_patch preserve line endings unconditionally (#51203)
## Why

Updating a CRLF file with `apply_patch` previously normalized its line endings to LF unless preservation was explicitly enabled. Preserve the file's existing line endings without requiring an opt-in.

## What changed

- Use line-ending-preserving updates for both tool calls and standalone `apply_patch` invocations, removing the legacy normalization mode.
- Mark `apply_patch_preserve_line_endings` as removed.
- Always set `CODEX_APPLY_PATCH_PRESERVE_LINE_ENDINGS=1` in child environments for compatibility with older patch executables, overriding stale values.

## Testing

Update CLI, tool-call, shell-heredoc, and app-server tests to check CRLF preservation by default. Keep coverage for overriding stale compatibility environment values.

GitOrigin-RevId: 96d0b0e5350faa3fc7a45d63627a132657b44523
2026-10-05 23:52:06 +00:00