## Why
Thread-owned Guardian context is always enabled, and `guardianv2.thread_context` is deprecated and ignored.
## What changed
Remove explicit `GuardianThreadContext` enables from retained-context and heartbeat tests. Parameterize the heartbeat skill-context test with optional feature lists so it covers the default behavior and `GuardianRootHandoffContext`, each with one or three heartbeats.
GitOrigin-RevId: 78e3159832bff08b79c7977d876d9acb587e9c71
## What changed
- Accept per-environment `skills.required` lists through `environment/add` and `environments.toml`.
- Before inference, require each named skill to be enabled and supplied by the selected environment. Missing skills fail the turn before a model request, including when environment setup fails.
- Stop applying an environment's requirements after it is deselected. Exempt isolated Guardian reviewers, which have no skill catalog.
## Testing
Add tests for available, missing, disabled, and wrong-environment skills, failed environment setup, recovery after deselection, and propagation of TOML requirements to the correct environment.
GitOrigin-RevId: 302ced4774697d2f414f745fe0535e71f747eda2
## What changed
Move base instructions from the top-level `instructions` field into a prefixed `developer` message in `input` for standard Responses requests, matching Responses Lite. Remove the field from HTTP and WebSocket request types and assign instruction messages stable IDs derived from the thread and instruction text.
## Testing
Extend standard and Responses Lite tests to cover instruction message serialization, stable IDs, and resume consistency. Update WebSocket tests to verify that changed instructions trigger a full request without `previous_response_id`.
GitOrigin-RevId: e2de4b8e30a9339377fac133ceda981abd1a112b
## What changed
Give the initial Guardian review and its checkpoint recovery attempt separate recovery flags instead of resetting a shared flag. Keep the single restart under the original review deadline.
Represent reviewer selection explicitly with `ReuseIfAvailable` and `FreshParentCheckpoint`, so recovery requires a fresh session and ignores reusable fork snapshots.
GitOrigin-RevId: 35f63f375e2a8bef10271944e222ffce1d1939c4
## Why
A Guardian review restarting from the parent checkpoint must bypass cached reviewer sessions and snapshots, even when another review is using the cached session.
## What changed
- Add `ReviewerRequest::requires_fresh_session` and enable it for parent-checkpoint recovery.
- Replace an idle cached session with a fresh reusable session, or start a fresh ephemeral session when the cached session is busy. Neither recovery path inherits a cached snapshot.
- Reserve the review permit before validating a cached session and before exposing a newly spawned session to concurrent requests.
## Testing
Add tests covering recovery with idle and busy cached sessions, verifying that recovery receives no cached snapshot and that subsequent reviews reuse the appropriate session.
GitOrigin-RevId: 894f685736da54312336932aa2037e74a20629d9
## Why
Reviewer compaction can invalidate previously admitted transcript evidence. When a parent checkpoint is available, restarting from that checkpoint preserves the captured parent context instead of relying on a lossy reviewer summary.
## What changed
- Restart an invalidated reviewer once from the action-time parent checkpoint, within the original review deadline.
- Preserve fresh checkpoints during startup and fit review input using the existing budget selection.
- Select transcript evidence from the checkpoint actually loaded by the reviewer. Track the transcript source so cursors remain valid across legacy-history eviction and reset when the source changes.
- Resume transcript deltas after successful recovery, including authorization added after the parent checkpoint.
## Testing
Add regression coverage for checkpoint recovery, post-checkpoint authorization, resumed deltas, and stopping after one restart. Extend history tests to cover both legacy retention and parent-checkpoint reuse, and update the checkpoint-migration snapshot.
GitOrigin-RevId: bc20fb9489afaad963303beff1c79f62fa5023f7
## What changed
Expose two JavaScript globals for consuming results as inputs settle:
- `as_settled(promises)` returns an async iterator of fulfilled and rejected records in settlement order, preserving queued results while the consumer is busy.
- `stream_settled(promises, emit)` delivers those records to a callback, awaiting each invocation and propagating callback errors.
Records include an `index`, a `status`, and either a `value` or a `reason`. Iterable inputs use positional indices; `Map` inputs retain their keys. Both helpers accept promises, thenables, and plain values.
## Testing
Add runtime coverage for ordering, input types, object identity, callback behavior, early iteration exit, and termination. Add stdio and gRPC host tests for partial results across waits, plus a scenario snapshot verifying fulfilled and rejected records reach the model.
GitOrigin-RevId: cddc9f3679ae04f2421e154caa0417afb15f550d
## Why
Incremental namespace declarations contain only changed or added tools. Their descriptions need to make clear that previously declared tools remain available unless explicitly marked unavailable, and that redefinitions replace earlier definitions.
## What changed
Append this guidance to updates of existing namespaces, preserving any namespace description. Leave initial declarations and catalog hashes unchanged.
## Testing
Extend unit and request-history coverage for the namespace hint, unchanged turns that append no tool definitions, and MCP tool removals that append a standalone notice without rewriting history or repeating the notice on the next turn.
GitOrigin-RevId: e32f3c9948efb44a1253868fe67967f117cc21e8
## Why
Manual, pre-turn, and post-turn compaction previously deferred rebuilding initial context until the next turn. Their saved replacement history could omit refreshed instructions and skill catalogs.
## What changed
- Install initial context during local and remote compaction, before the last real user message or summary, keeping the compaction summary last.
- Persist matching full world-state and turn-context baselines with every replacement instead of retaining only partial extension metadata.
- Build replacement context and recompute token usage for the current model even when the previous model performs compaction. Reuse that captured context for eligible current-model retries.
- Capture manual compaction context centrally and propagate preparation errors to the task runner.
## Testing
Update assertions and snapshots for context ordering, refreshed instructions, resume and fork history, current-model checkpoint baselines, and cloud skill catalog retention across compaction and resume.
GitOrigin-RevId: 8572a76c31c6884a3c580b3bbb350d49af876bd1
## Why
MCP elicitation reviews used turn-start context, which could omit remote environments that became ready before the tool invocation and use stale environment permissions.
## What changed
Capture the issuing step's Guardian review context with each live MCP invocation and reuse it for elicitation reviews. Fall back to turn context when no matching live invocation is available, and preserve the captured context for in-flight actions.
## Testing
Extend the deferred-executor regression test to cover MCP elicitations, checking that Guardian receives the newly ready remote environment and its denied paths, excludes stale local denied paths, and returns the denial to the parent model.
GitOrigin-RevId: 0cc58b8f30543a9af87d53e1a90f574a8e6a2f84
## Why
An already-cancelled delegate spawn can report a runtime admission or startup error instead of `TurnAborted`, including when runtime shutdown has been requested.
## What changed
Check the cancellation token before runtime admission and session startup so prior cancellation returns `TurnAborted` immediately after approval-policy validation.
## Testing
Extend the pre-cancelled spawn test to cover both normal operation and requested runtime shutdown, asserting `TurnAborted` within a one-second timeout.
GitOrigin-RevId: 5a25a211acdd5c246a3c638243ccc8b7240fd051
## Why
The warning assertions need tracing callsite interest to remain independent of untraced parallel tests, even when the test runner already owns the global subscriber.
## What changed
Run the test on a current-thread Tokio runtime and capture warnings with a thread-local subscriber and an explicit buffer. Keep a `NoSubscriber` dispatch alive to stabilize callsite interest, and assert the deferral warning and `client_echo` directly against the captured logs.
GitOrigin-RevId: 5bf22b2cbfefcf8f65016d2cc1444112d9f6a6c7
## What changed
Add `tools_change_count` to turn profiles and turn analytics events. Compare the full model-visible tool list before each sampling request, retaining the previous list across turns and connection resets. The first list establishes a baseline; subsequent changes increment the current turn's count.
## Testing
Add coverage for schema changes, tool additions, ordering changes, empty lists, and tracking across sessions and transport fallback. Extend the turn event serialization test to include `tools_change_count`.
GitOrigin-RevId: 44a15c42dc7d0b35c166666faeb9627b718b33dd
## What changed
Add the default-off `stable_environment_tools` feature flag. When enabled, advertise environment-backed tools before an executor is ready, keep environment selectors stable as readiness changes, and include the `shell` and `login` parameters. Execution still requires a usable environment and its policy.
When disabled, expose environment-backed tools only when a usable environment exists, derive selectors from usable environments, and condition shell parameters on the execution configuration. Restore tool-specific unavailable-environment errors and skip the environment readiness check for `write_stdin`.
## Testing
Parameterize the empty-environment tool availability test for both flag states, check that commands require a ready transport by default, and explicitly opt tests that rely on stable tool exposure into the feature.
GitOrigin-RevId: 68c6240fb3adf85d7f9b407804d4823c515abfdc
## Why
Malformed `deny_read_acl_state.json` previously caused deny-read ACL reconciliation to fail. Recovery must restore bookkeeping without removing unknown existing restrictions or modifying a linked file's contents.
## What changed
- Retry malformed reads for up to two seconds, then rebuild bookkeeping only after excluding active writers and successfully applying current deny-read ACLs. Preserve unknown historical denies and continue propagating I/O errors.
- Validate the opened state file before reading or writing, rejecting reparse points and files with multiple hard links.
- Write state in place without truncating before validation, preserving file identity and permissions. Retry sharing violations during writes while preventing replacement through delete access.
## Testing
Add Windows tests for corrupt-state recovery, linked-file rejection, concurrent writer and delete-handle behavior, file identity and ACL preservation, and subsequent permission-profile changes. Extend the elevated sandbox integration test to verify that denied reads remain blocked after recovery while allowed writes and deletes succeed.
GitOrigin-RevId: 4aca32df00d015df3763fdc447ffdfbe3fd6cec0
## Why
A review can fail before the UI receives the event that enters review mode. Failed-turn completion also needs to preserve the running indicator when a queued `/review` has already been submitted and is awaiting startup.
## What changed
- Emit the review-entry lifecycle before starting the review task, after aborting existing tasks and clearing connector selection.
- When a live failed-turn completion repeats an already handled error, refresh task state only if no user turn is pending startup, and continue processing queued input.
## Testing
Add regression tests for lifecycle ordering on a server-overload response and for preserving the running indicator while a queued review awaits startup after a failed turn.
GitOrigin-RevId: 0f91d3d1a1a983542c54c2686630de9c232321b1
## Why
Command Center previously reset to project grouping on startup, losing the user's selected grouping.
## What changed
Save grouping changes to `tui.agents_overview_grouping` in the client’s user configuration and restore the saved choice at startup. Support `project`, `status`, and `model`, with `project` as the default. Report save failures in Command Center while keeping the selected grouping active.
## Testing
Add tests for cycling through and persisting all grouping modes, preserving unrelated settings in a custom configuration file, displaying save errors, and restoring saved grouping during startup.
GitOrigin-RevId: fd56dfecb4b79cfa8439f54420ce58cc0ed66ec7
## Why
Environment readiness previously changed which tools and command parameters were advertised to the model, even when the selected environments stayed the same.
## What changed
- Keep enabled command, patch, image, and permission tools exposed when environments are absent, starting, or failed.
- Return a consistent message asking the model to wait when no usable execution environment is available, while preserving distinct errors for unknown environment IDs.
- Derive environment selectors from all selected environments rather than only ready ones.
- Always advertise `exec_command`'s `shell` and `login` parameters while continuing to enforce login policy at execution time.
## Testing
Expand integration coverage for empty, pending, failed, ready, and recovered environments, including stable command parameters and login-policy rejection. Update subagent and skill-catalog snapshots, with separate Windows expectations.
GitOrigin-RevId: bb08084066896a9f14b3b568c081c3eaac981db7
## Why
Third-party exposure settings can eagerly expose tool schemas or exclude tools from `exec`. Keeping eligible tools deferred lets MCP catalogs change between turns without changing the model's eager tool prefix.
## What changed
Add the disabled-by-default `code_mode_only_strict_3p_tools` feature, applied only when the effective model tool mode is Code Mode Only.
- Defer eligible MCP, app, and client-supplied dynamic tools while keeping them discoverable and callable through `exec`.
- Override `deferLoading`, MCP/app `omit_tools_from`, and direct-only or excluded namespace settings. Log conflicts without adding prompt warnings.
- Preserve disabled, app-only, policy-denied, and over-budget tool restrictions. Keep built-in tool exposure unchanged, including when a dynamic tool has the same name.
## Testing
Add unit and integration coverage for exposure overrides, effective model modes, tool eligibility, and app discovery and dispatch. Add an MCP refresh scenario that verifies the eager tool prefix stays unchanged as tools are added, replaced, and removed while `exec` tracks the current catalog.
GitOrigin-RevId: a6a8e401214119bb02ea4711473c8fa3ec85c81a
## Why
Tool availability can change between calls. Conditioning discovery guidance on whether deferred tools are currently present changes the `exec` description when the catalog changes.
## What changed
Always include the configured discovery guidance in the Code Mode `exec` description, even when no deferred tools are present. The default guidance explains how to find tools in `ALL_TOOLS` and notes that availability can change between calls. Custom guidance and explicit empty overrides remain supported.
## Testing
Add regression coverage for stable guidance with empty and populated deferred tool catalogs in both Code Mode configurations, including custom and empty overrides. Update tool-description assertions and request snapshots.
GitOrigin-RevId: 086ea19a48e3c51ec8c429a619676a936d658721
## Why
In `CodeModeOnly`, MCP resource helper definitions changed when the first server was added or the last server was removed.
## What changed
Always register `list_mcp_resources`, `list_mcp_resource_templates`, and `read_mcp_resource` in `CodeModeOnly`, even without configured MCP servers. This keeps their definitions stable across server configuration changes.
## Testing
Add a regression scenario that calls all three helpers before adding a server, with a server registered, and after removing it. Assert identical tool definitions across requests, empty resource and template lists, and the expected error when reading from an unknown server.
GitOrigin-RevId: c5777efac250b83dd4bb1c9100a6b6169470556f
## What changed
When `IncrementalTools` is enabled for Responses Lite, track tool definitions in world state and conversation history. Send the initial catalog once, then append only added or changed definitions while preserving earlier request history. Emit developer notices for removed tools and namespace instruction changes.
Use history-supplied definitions for normal requests and remote compaction, and omit empty `additional_tools` prefixes. Keep the existing tool delivery behavior for the Responses API and when the feature is disabled.
## Testing
Add unit and snapshot coverage for catalog and namespace transitions, empty catalogs, and stable hashing across JSON object key order. Add a request-history scenario covering unchanged tools, a model switch that changes one definition, persisted catalog hashes, and Responses API behavior.
GitOrigin-RevId: 76bde175973b81f8ad23873919da3c8ddcdf6065
## Why
Changing a deferred tool's output schema could add or remove the shared MCP type preamble from the `exec` description.
## What changed
Include the default or configured MCP TypeScript preamble whenever `code_mode_only` is enabled, regardless of the available tools' output schemas. Preserve an empty preamble override as a way to omit the section.
## Testing
Add a regression test that checks description stability when a deferred tool gains an MCP output schema, covering default, custom, and empty preambles. Update description expectations, request snapshots, and two Guardian context-budget test thresholds for the added preamble.
GitOrigin-RevId: 725035d6159e9da3c4beff5482c426c082e768b4
## Why
When `flush_transcript_tail_on_session_end` is enabled, remaining speech must be available in thread history as soon as realtime closure is reported, even while a model response is still running. Saving that transcript should not trigger another model request.
## What changed
Record the transcript tail directly in model and client history, attaching it to the active turn or creating a completed history turn when idle. Flush persistence before emitting `RealtimeConversationClosed`, and keep realtime context active until recording finishes so the return to text instructions follow the transcript.
## Testing
Extend core and app-server tests to cover immediate history visibility in paginated and legacy modes, idle and active turns, closure without extra inference, transcript reuse in subsequent text turns, and interruption after closure.
GitOrigin-RevId: ae323b711f22aa88db1432e24ad4bb3eb3f08f6b
## What changed
Add a scenario test and snapshot covering the request before remote `/compact`, the compaction request, the persisted replacement history, and the first subsequent user turn.
Capture the replacement history before the next turn can update its context, and assert that the compaction request contains one `compaction_trigger` and the next request includes the returned encrypted checkpoint.
GitOrigin-RevId: e22e12f16bdb11b1296e81a2e0ffd5f70e560ff6
## Why
Bedrock catalogs cleared all service-tier metadata, and requests omitted every selected tier. This prevented users from selecting `ultrafast` or using tiers defined in custom catalogs.
## What changed
- Advertise `ultrafast` for `openai.gpt-6-astra` on Mantle and the `global.openai.gpt-6-astra` and `us.openai.gpt-6-astra` Runtime variants, while retaining the implicit default tier unless users select otherwise.
- Preserve custom catalog speed tiers, service-tier definitions, and defaults while continuing to normalize Bedrock web search compatibility.
- Send selected service tiers only when the Bedrock model advertises them, including `flex`.
## Testing
Add coverage for `model/list` tier discovery, custom catalog preservation, configured and per-turn HTTP request tiers, and TUI slash-command selection, toggling, and model switching.
GitOrigin-RevId: 4793e7af7b9f604a6b7812611030da6adb8acd35
Register `incremental_tools` as an under-development feature that is disabled by default, and expose it in the configuration schema for top-level and profile feature settings.
GitOrigin-RevId: 9579822c745e9fce598085bdc5ef8da6aabb06d7
## What changed
Allow Responses-compatible providers to configure live web access and remote compaction under `model_providers.<id>.capabilities`:
```toml
[model_providers.custom.capabilities]
external_web_access = false
remote_compaction = "v2"
```
`external_web_access = false` keeps hosted web search from accessing the live web. `remote_compaction` accepts `v2` to send `compaction_trigger` items through the Responses endpoint, or `unsupported` to use local compaction. Omitted capabilities retain existing provider defaults.
## Testing
Add unit coverage for preserving unspecified defaults and a request-snapshot scenario verifying cached web search and V2 compaction for a custom provider.
GitOrigin-RevId: fbcc4870aa51c112af575d711679a3941d1e7878
## Why
Paginated thread history can persist large MCP results in completed tool-call items. Bound these stored results to avoid retaining multi-megabyte payloads.
## What changed
Apply a 64 KiB preview budget when persisting oversized MCP results in paginated history. Replace the result with a truncated text preview of its serialized JSON, preserving `is_error` and clearing `structuredContent` and `_meta`. Leave small results unchanged.
Share the truncation helper with the existing MCP tool-call event path.
## Testing
Add helper tests for unchanged small results and oversized results with structured content and metadata. Add a thread-history test verifying that persisted MCP results contain a bounded preview with the beginning, end, and truncation marker while preserving tool arguments.
GitOrigin-RevId: 09dce73c70bfd357c9b9630f456e9e5c1d8492cf
## What changed
- Remove `namespace_tools` from `ProviderCapabilities` and `namespaceTools` from `ModelProviderCapabilitiesReadResponse`, updating generated schemas and SDK types.
- Preserve tool namespaces and configured collaboration tool namespaces without a provider capability check, and always use Responses Lite tool serialization for Responses Lite requests.
- Gate tool search on `supports_search_tool` and retain the existing image generation and web search checks without requiring a namespace capability.
GitOrigin-RevId: ae1843563736765a40e03a02c41ff5e470018a93
## What changed
Strengthen `tool_call_timing_guard_ignores_code_mode_source` to capture tracing output and assert that direct and nested code-mode calls together emit exactly one `codex.tool_call` event. Disable ANSI formatting and keep callsite interest independent of untraced parallel test threads.
GitOrigin-RevId: cbf0c9bf6d4e5a72b6a1eb08e6aac43e257f2dcb
## Why
World-state sections could return only a single optional text fragment, with message separation controlled by the fragment itself. They could not contribute explicit `ResponseItem` values alongside instructions.
## What changed
- Let sections return ordered `WorldStateUpdate` values containing fragments or response items, with `Prefix`, `Standalone`, or `Mergeable` placement.
- Preserve explicit item boundaries when grouping context updates, and place prefix updates first when assembling initial context.
- Move standalone-message decisions from fragments to world-state updates and retain snapshot handling.
## Testing
Add tests for mixed fragment/item ordering, role and standalone boundaries, initial-context prefix placement, and snapshot restoration without repeated updates.
GitOrigin-RevId: c2514c3c98538da83f5140c10a9d89f531de6549
## Why
Streamed `response.failed` events can carry retry advice in `response.error.headers`. Ignoring it loses server-provided delays or uses a conflicting delay from a rate-limit message.
## What changed
- Parse nested `Retry-After` headers using the existing HTTP header validation and retry parser, and preserve them for rate-limit, overload, and generic retryable failures.
- Prefer valid headers over rate-limit message advice, falling back to the message when the header is absent or invalid. Keep terminal error classifications unchanged.
- Reject unrepresentable plaintext retry durations without panicking.
## Testing
Add regression coverage for header precedence, validation, and terminal errors, plus retry-loop coverage for SSE sampling, remote compaction, and WebSocket failures.
GitOrigin-RevId: 46433c2e9e6d5e804211ef96950fadd70e322b49
## What changed
- Remove `stdout`, `stderr`, and `formatted_output` from `CommandExecutionItem` and `ExecCommandEndEvent`, retaining `aggregated_output` as the output field.
- Update command event producers, legacy conversion, rollout traces, and the precomputed schema. Use aggregated output for terminal trace results.
- Clear `interaction_input` in emitted and migrated command completion items.
## Testing
Update command execution and trace tests to use aggregated output, and assert that legacy rollout migration preserves `aggregatedOutput`.
GitOrigin-RevId: ed59827d4a5ec9a767084ee4feac72c2042440f2
## What changed
Remove `initial_messages` from `SessionConfiguredEvent` and the corresponding schema. Stop attaching replay history to startup events for resumed or rejoined threads, and remove `InitialHistory::get_event_msgs`.
## Testing
Update the MCP resume test to verify persisted tool events through `load_history`, and retain coverage that resumed requests include prior conversation items.
GitOrigin-RevId: 54b72bc44423c9b1ef2051891ef392c0ede32fe5
## What changed
Add an MCP integration test that forwards sandbox state from tool-call metadata unchanged to `codex sandbox --sandbox-state-json` and runs a real subprocess.
Verify that workspace writes succeed, overwriting an existing file outside the workspace fails and preserves its contents, and relative paths resolve against `sandboxCwd` rather than the MCP server's working directory.
GitOrigin-RevId: 768513c74679010e806a62dffdde6b7cccbfb41f
## What changed
Add `tui.mouse_scroll_speed` as a finite positive multiplier for transcript mouse wheel scrolling, including transcript overlays and session previews. Change the default from three rows to one row per wheel event; set the multiplier to `3.0` to retain the previous speed.
Fractional values accumulate movement across events and discard the remainder when direction reverses. For example, `0.5` scrolls one row every two events. Following the latest output pauses only when a whole row moves.
## Testing
Add tests for configuration validation, integer and fractional scroll speeds, direction changes, decimal accumulation, and following behavior. Extend local settings coverage to include a command-line override for the new setting.
GitOrigin-RevId: d355e246fd943415f0ce5a6affe849f77f3615e6
## What changed
Add `CodexThread::current_turn_model` to return the model slug selected for the named running turn's next step, independently of settings for future turns. Return `None` when the turn ID does not match, the task is cancelled, or the turn is no longer running.
## Testing
Extend the step-settings test to verify that the accessor ignores model changes for future turns, reflects active-turn model updates, and returns `None` for a mismatched ID or a completed turn.
GitOrigin-RevId: 2cae7c07fa761ed4d64befaa3419a2f9bce4b161
## What changed
Add the disabled-by-default `guardianv2_decisions_comparison` feature to run Decisions alongside Guardian V2 snapshot classification using the same policy and evidence. Initialize the comparison sampler with `CODEX_GUARDIAN_DECISIONS_API_KEY`.
Publish the baseline score before awaiting comparison results, and keep comparison results and failures from changing approvals. Apply Guardian's input token budget, skip conversation evidence, and cancel comparison work when the baseline classification is superseded. Record success, failure, and skip outcomes with bounded diagnostic reasons, including setup failures.
## Testing
Add lifecycle tests covering disabled comparison, disagreeing classifications, backend failures, unchanged published scores, matching policy and evidence, and rejection of over-budget requests before either backend sends them.
GitOrigin-RevId: 743d34c8ee440362da94f001058366027b39bd11
## Why
Reloading an unloaded root from a surviving child can supply the shared instruction wrapper already registered for that root. Replacing its underlying provider with that same wrapper makes instruction loading recurse into itself.
## What changed
Reuse the existing wrapper without replacing its provider when both are the same `Arc`. Continue to accept distinct replacement providers.
## Testing
Extend regression coverage to reload the root from a surviving child and verify that shared instructions still update. Check wrapper reuse, replacement with a distinct shared provider, and freezing the shared snapshot when switching to a provider that does not share with subagents.
GitOrigin-RevId: 7b6b2f9031f206e203f6fa14e1478a36b305ec39
## Why
Unread queue-only messages prevented idle agents from unloading, and sending a
message to an evicted agent reloaded its session. Pending mail should not require
an idle recipient to occupy a loaded thread slot.
## What changed
- Retain queue-only mail in an in-memory runtime mailbox when a session unloads.
Queue new messages for locally evicted agents without reloading them or
starting a turn.
- Add `AgentControl::take_mailbox` and `AgentControl::watch_mailbox` so sessions
can consume retained mail and react to host mailbox notifications.
- Preserve message order across eviction, reload, and subsequent follow-ups.
Keep agents with outstanding durable sleep or turn-triggering mail loaded,
and discard retained mail when an agent is removed or its tree shuts down.
## Testing
Add integration coverage for retaining mail across eviction, sending while
unloaded, and ordering messages around a follow-up after reload. Add a host
controller test verifying that mailbox notifications wake `wait_agent` and
that the message is delivered exactly once.
GitOrigin-RevId: 4b55647f9ee9d1aaf3753f0688863bda81617327
## Why
Subagents spawned without forked history do not receive their parent's client-defined dynamic tools, preventing them from using those tools for delegated work.
## What changed
Add the disabled-by-default `multi_agent_v2_dynamic_tools` feature flag. When enabled, fresh V2 subagents inherit the parent's dynamic tool definitions and persist them in session metadata.
## Testing
Add coverage for V1 and V2 feature gating, tool persistence, and a fresh V2 subagent calling an inherited tool and receiving its result without inheriting parent history. Update TUI assertions and a snapshot to reflect blank-line separation and question text in recovered answers.
GitOrigin-RevId: a17a16bf5570f54ce922c1521cffb883022354f2
## What changed
- Centralize `windows-sys` at version `0.61.2` in the workspace and migrate dependent crates to it.
- Adapt Windows API calls to the updated handle, boolean, and type definitions. Replace custom handle owners with `OwnedHandle` and use `BorrowedHandle` where ownership stays with the caller.
- Add explicit wrappers for non-owning Windows handles passed between threads.
## Testing
Enable the MCP refresh cleanup test on Windows. Retain process handles to verify that a superseded MCP server stays alive during an in-flight call, exits after cancellation, and leaves the replacement running until shutdown.
GitOrigin-RevId: 86f53d9b300d13edff7154f4a7c32e34dacd2919
## What changed
Update `post_sampling_token_estimate_is_disabled_by_always_on_sinks` to register explicit event metadata with the subscriber and assert that its interest is `never`. This replaces the scoped default subscriber and global interest-cache rebuild with a direct check of the configured logging layers.
GitOrigin-RevId: 7c4ecd9c4adef148fd222394ad1ad326e8033d95
## Why
Environment changes and overlapping catalog refreshes can cause a step to use another step's selected plugins or executor skills. Model context and skill tools need to agree on the capabilities captured for that step.
## What changed
- Resolve selected plugins from each step's ready capability roots, honoring disabled plugins, and retain the snapshot with the step's extension data.
- Use that snapshot for skill attribution in World State and the first step's active-plugin telemetry.
- Return executor skill catalogs directly from refresh so overlapping requests render their own results while retaining thread caches.
## Testing
Add regression coverage for overlapping executor catalogs and environment changes during step preparation. Verify that `skills.list` retains the original step's skills, the next step advertises the new skills, and disabled executor and plugin skills remain absent from model context.
GitOrigin-RevId: 37d3cbe8c7e062fd06290aa985ada4d3b3462d43
## Why
Legacy MCP tool discovery discarded pagination cursors, leaving tools on later pages unavailable.
## What changed
Pass `tools/list` pagination cursors to the existing collector in both legacy and modern protocol modes. Legacy discovery now follows subsequent pages and applies the same pagination limits and cursor validation as modern discovery.
## Testing
Extend integration coverage to both protocol modes, checking that tools from later pages reach the model and can be called successfully, while catalogs with oversized cursors are rejected. Update the legacy discovery test to expect an error for a repeated cursor.
GitOrigin-RevId: a6daa8034d2a403c1f2a5846e385789d8a620af7
## Why
A user restriction such as “Cancel the deployment” can fall outside the recent-message window after unrelated status questions. Guardian reviews need that restriction when assessing delegated actions.
## What changed
- Preserve root user messages during handoff relevance filtering, subject to the existing shared message cap.
- Apply handoff filtering during heartbeats while retaining saved heartbeat instructions and the active turn's trusted skills.
- Mark assistant context as incomplete when handoff filtering omits it, so ordinary user replies are not interpreted without their missing context.
## Testing
Add regression coverage for restrictions followed by status questions, bounded user history, omitted assistant questions, and trusted skills across first and repeated heartbeats. Update the delegation request-history snapshot.
GitOrigin-RevId: 67f77313572dc54225713cd7cdb90f06b0fbe030
## Why
Concurrent Linux test spawns can inherit writable descriptors for executable fixtures, preventing those fixtures from being executed even after the writing test closes its descriptor.
## What changed
Replace direct executable fixture writes and copies across the Rust tests with `codex_utils_cargo_bin::write_executable` and `copy_executable`. These existing helpers complete writes and copies in separate processes on Linux.
Remove the empty temporary inode before creating fake bubblewrap executables, and stage tampered bubblewrap bytes in a separate data file before copying them into the executable fixture. Add the required test dependencies.
GitOrigin-RevId: c875c83088731a4933ef1574e8d79d8cecc12f54
## Why
Short sender replies such as `👍` need preceding assistant context to be understood during a delegation review. User messages alone omit that context.
## What changed
- Capture preceding assistant context alongside up to three recent sender user messages at delivery, for Guardian only. Label assistant messages as untrusted context rather than authorization or verified questions.
- Share the existing 900-byte budget per exchange, prioritize user evidence, and report unavailable assistant context.
- Recognize `cloud_threads.send_message` deliveries and accept both compact and indented delegation envelopes.
## Testing
Expand integration coverage for app, TUI, and cloud-thread deliveries, contextual replies, and budget-driven omissions. Assert that sender user and assistant context stays out of the delegated agent's prompt.
GitOrigin-RevId: 4eca369cbb30241635273f729dec81b78ea57faf
## Why
`codex exec` and the TypeScript SDK need a way to request a Cyber access program for an individual turn. Resuming with an API key can also leave compaction using a saved program that the current credentials cannot access.
## What changed
- Add `--cyber-access-program` with `standard`, `daybreak_blue`, and `daybreak_red`, and expose the equivalent `cyberAccessProgram` turn option and `CyberAccessProgram` type in the TypeScript SDK.
- Apply explicit selections to new, resumed, and prompted fork turns. Override automatic Daybreak selection for that turn without changing the saved choice. Reject the flag for review and forks without a prompt.
- Filter Cyber programs by the built-in OpenAI provider when starting turns and restoring previous-turn settings for compaction.
- Allow previous-model compaction fallback with API-key authentication, comparing both model and program so a program change can trigger a retry with the current pair even when the model is unchanged.
## Testing
Add CLI help, argument parsing, exec, and SDK coverage for per-turn selection and unsupported commands. Extend core tests for API-key compaction retries, denied explicit selections, custom-provider filtering, remote compaction, and WebSocket response reuse.
GitOrigin-RevId: 586f25b621924f3bc509c9a39b39fe2020ff9cd2
felixxia-oai