## Why
MCP requests that need human input, including browser sign-in, must be handled by the root thread. Subagents need guidance to hand these blockers to their parent without prompting the user or automatically accepting requests that require input.
## What changed
- Reject interactive MCP elicitations and tool approval prompts in subagents with guidance to ask the parent and wait before retrying.
- Recognize browser authentication and `codex_requires_user_input` metadata even when the form schema is empty, while preserving automatic permission approvals and review decisions.
- Carry user-interaction eligibility through MCP runtime creation and connection reuse, and guard prompt registration.
- Preserve connector authentication diagnostics alongside handoff guidance, subject to normal tool-output limits.
## Testing
Add unit and integration coverage for blocked subagent prompts, root browser authentication, automatic approval and review, connection reuse, and authentication diagnostic preservation and truncation. Add a request-history snapshot for browser-auth handoff guidance.
GitOrigin-RevId: 83e146b6e5c1a2a22a34681ad45615f504b3c96c
## Why
Compaction prompts omitted recorded Code Mode tool calls, leaving them without the nested tool inventory available to normal sampling requests.
## What changed
- Attach pending and retained Code Mode tool metadata to local and remote v2 compaction prompts, including updated result metadata and completeness markers.
- Bound Code Mode metadata separately while preserving captured direct tool records.
- Keep observations omitted from shortened compaction retries available if compaction fails and normal sampling resumes.
## Testing
Add regression coverage for metadata attachment, retry retention, independent budgeting, and binding validation. Extend local and remote compaction tests to check tool inventories with metadata enabled and disabled.
GitOrigin-RevId: 1dc012bd5838797c605d981a8fc1b0378a142227
## Why
Server-level tool exposure settings apply to every connector on the apps server. Individual connectors need a way to opt out of deferred discovery without changing exposure for other apps.
## What changed
Add `apps.<connector_id>.omit_tools_from` with support for `code_mode`, `deferred`, and `direct`. Combine connector omissions with server omissions so app settings preserve server restrictions.
For example, `omit_tools_from = ["deferred"]` makes a connector's tools available without tool search, through direct calls or Code Mode as permitted by the active tool mode and remaining restrictions.
Expose the setting in the app-server protocol, JSON schemas, and generated TypeScript and Python types.
## Testing
Add integration coverage for connector-specific exposure and MCP dispatch across tool modes, server restrictions, and direct-only namespaces. Extend config tests to cover populated, absent, and empty omission lists, and add a Code Mode request-history snapshot.
GitOrigin-RevId: cc99ab290c8a878d06d8d7ccf2d1dbd7c46bfa8a
## What changed
- Add `ResolvedModelMessages` to resolve catalog text and bundled defaults while preserving explicit empty overrides and their source.
- Move base-instruction rendering, Guardian prompt composition, multi-agent role rendering, and `update_plan` guidance filtering into `codex-prompts`; migrate consumers to the shared APIs.
- Separate permission-profile resolution from prompt composition, and annotate Guardian policy and classifier instructions with content kinds.
## Testing
Add and update coverage for missing versus empty templates, literal overrides, multi-agent role composition, Guardian policy substitution and truncation, and preservation of permission path spellings and order.
GitOrigin-RevId: 52335bb7acec0f432d5c57acb2accd5f0276056e
## What changed
Capture result metadata for host-owned app calls when analytics and executed tool call metadata recording are enabled. Respect both `analytics.enabled = false` and host-disabled analytics, using the prepared call's ownership to determine eligibility.
Disable incremental WebSocket request reuse when raw result metadata or its call binding changes, so late results attached to an already-sent output are included in the next request. Continue allowing reuse when only other metadata changes or result metadata is filtered out for the endpoint.
## Testing
Extend coverage for direct and code-mode calls, analytics settings and host overrides, runtime recording changes, and prepared call ownership. Add metadata comparison and WebSocket reuse tests, and verify that raw response item notifications continue to omit result metadata.
GitOrigin-RevId: 411843fbff3749ee96fd6f0ebc591362942a4555
## Why
Clock provider failures can abort a turn while preparing time context or running clock tools. Allow turns to continue with an explicit indication that the current time is unavailable.
## What changed
- Add `features.nonfatal_clock_read_errors`, disabled by default, to report clock failures to the model without failing the turn.
- Emit a generic `failed to read current time` notice for context reads and tool errors, without exposing provider error details. Deduplicate context notices per turn and compaction window, and remove inherited notices from forked subagent context.
- Omit unavailable environment dates and explicitly clear previously visible dates with `<current_date status="unavailable" />`.
- Return external sleep clock failures to the model when the feature is enabled, preserving sleep item completion notifications.
## Testing
Add coverage for continued inference after clock failures, notice deduplication across compaction, subagent notice filtering, environment date removal and recovery, and sleep failures during initial and polling reads.
GitOrigin-RevId: a39c3723c06a6f786d8ad59d667e8b9f626c973e
## What changed
Reject `request_plugin_install` calls from non-root agents with an error returned to the model before parsing arguments or prompting for installation. Apply the restriction to both legacy install requests and recommended plugin requests.
## Testing
Add unit and integration tests for both request formats, verifying that subagent calls return the root-only error and emit no installation elicitation request.
GitOrigin-RevId: f23bbc6dd34d69366b81e4033bd554bba1fb816a
## Why
Clients need widget presentation details to render MCP Apps without waiting for the full MCP catalog, including when replaying saved history.
## What changed
- Add `mcpToolCall.mcpAppUi` with the invoked descriptor's `resourceUri` and `preferredModelDisplayMode`, and carry it through tool-call events and saved history.
- Support `fullscreen` and default to `inline` when the display preference is missing or unsupported.
- Keep existing resource URI fields for compatibility. Leave `mcpAppUi` null for older history and tools that declare widgets only in result metadata, where clients still use catalog discovery.
- Update protocol schemas, TypeScript and Python types, and app-server documentation.
## Testing
Add parameterized integration coverage for fullscreen, missing and unsupported preferences, legacy URI metadata, and result-only widgets, verifying consistent tool events and preservation across session resume.
GitOrigin-RevId: 2a9bed804dcea8c2b4903b8406fa6710681f7f5e
## What changed
- Accept `fileId` alongside the existing `url` form for app-server image inputs, and forward file references to the Responses API as `file_id`. Update generated schemas and client types.
- Preserve file references, image detail hints, and mixed inline/file image ordering through user-message events, thread history, and rollout migration. Retain file images when truncating tool output.
- Pass file references through image preparation without resolving them, while keeping resize-notice numbering correct. Omit them from unsupported TUI display and Guardian image context.
- Reject image-edit requests whose recent-image window includes a file reference, preventing selection of an older inline image instead.
## Testing
Add coverage for serialization, request and rollout preservation, mixed-image history ordering, incomplete ordering metadata, tool-output truncation, and rejection of unsupported image-edit selections.
GitOrigin-RevId: 6ca20a8577155cc934b720803c3b7b3bffdf972a
## Why
Older compaction checkpoints can lack a producer model hash. Enabling thread-owned Guardian context must accommodate those checkpoints without losing user instructions or verified answers.
## What changed
- Select review policy from each history snapshot independently of the session's evidence capture policy.
- Keep legacy review for checkpoints with missing or empty producer hashes when the transcript preserves the evidence. Preserve strict compatibility checks when review depends on retained-only evidence.
- Continue capturing retained instructions and answers during migration, while supplying legacy reviews with runtime answers and subagent reviews with root authorization evidence.
- Record producer model hashes on new compactions regardless of context mode.
## Testing
Add coverage for checkpoint migration across compaction and resume in both history storage modes, snapshot policy stability, retained-evidence preservation, and authorization revalidation for owning sessions and subagents.
GitOrigin-RevId: ec72b4a51d40729173e24ac1b53b7d516ddaa825
## Why
MXC is a sandbox implementation, not a restricted-token sandbox level. Executor requests need to represent that choice separately from `WindowsSandboxLevel`.
## What changed
- Introduce `WindowsSandboxSelection` for executor sandbox contexts and remove `Mxc` from `WindowsSandboxLevel`.
- Preserve the `windowsSandboxLevel` wire field and its serialized values for compatibility.
- Share sandbox selection between executor process launches and filesystem helpers, and use the new selection in capability discovery and skill reads.
- Disable Windows sandbox selection for executor paths that do not use Windows path conventions.
## Testing
Extend coverage for MXC wire serialization, Windows skill-read sandbox checks, and capability discovery with distinct permissions. Exercise remote filesystem write restrictions with both restricted-token and MXC sandboxes, including rejection when native MXC is unavailable.
GitOrigin-RevId: 266211377bcb138a0dc75861e9ff2225fa37a53d
## What changed
- Add `AllowedTools` to the extension API, captured once at thread startup to restrict which tools can be advertised or executed. An empty list disables all tools; an absent value preserves ordinary tool setup. Callers must supply the list again when resuming a thread.
- Filter tool registration and hosted tool specifications before Code Mode and discovery. Match names with their namespaces and require generated tools such as `exec` and `wait` to be explicitly allowed.
- Move Guardian reviewer tool selection into the allowlist while retaining feature and sandbox restrictions, with a fallback for older reviewer sessions.
## Testing
Add coverage for selected, empty, and absent allowlists across tool sources and Code Mode. Add managed-thread tests verifying advertised tools, rejection of excluded calls, and that replacing extension state after startup cannot change the captured allowlist.
GitOrigin-RevId: 33c1722c4241f368963d8385fa0415112e16a72d
## What changed
Extract V2 interruption validation and dispatch into `AgentControl::interrupt_spawned_agent`, returning the agent path and previous status with typed validation and runtime errors. Update the `interrupt_agent` tool handler to delegate to it while retaining tool error mapping and activity emission.
Preserve rejection of root and self targets and successful handling of unloaded or already-dead runtimes without reloading them.
GitOrigin-RevId: 893bbc4f464596d0fee032a71702a4a4f6becf5c
## What changed
Extract target validation, runtime reloading, and message delivery from the V2 tool handler into `AgentControl::deliver_message`. Keep target resolution, analytics, and tool-facing error mapping in the handler.
Represent plaintext and encrypted payloads with `AgentMessage`, sharing communication construction with agent spawning. Preserve queue-only and follow-up turn semantics, target checks before reload, and turn metadata propagation.
GitOrigin-RevId: a88bcede015a17f6c11c7cad95ee58add5dc2820
## What changed
Move shared spawn and resume configuration helpers into `agent::child_config` and route both multi-agent versions through `prepare_agent_spawn_config`.
Preserve each version's role and model precedence, full-history inheritance, service-tier selection, runtime policy, and default-role metadata for reloads. Return configuration errors as strings and convert them to `FunctionCallError::RespondToModel` at tool-handler boundaries.
GitOrigin-RevId: 9d539050aa44f94043a6487f1f6f1037de1d15ca
## What changed
Represent images in `ContentItem` and `FunctionCallOutputContentItem` with `ImageReference::Inline`, flattened to preserve the existing `image_url` wire format. Update image producers and consumers and regenerate app-server schemas and SDK artifacts.
Preserve the Python SDK's `InputImageContentItem` and `InputImageFunctionCallOutputContentItem` class names during generation.
## Testing
Add a regression test for stable Python image class names and adapt existing image tests to the shared representation.
GitOrigin-RevId: c38a780ac3314c2ac2deb3afc1b93b94b6f93fec
## Why
Linux proxy-routed sandboxing denied standalone Unix sockets even when the effective network policy enabled `dangerously_allow_all_unix_sockets`.
## What changed
- Carry Unix socket permissions in `ManagedNetworkSandboxContext` and pass the prepared context through Linux sandbox launches with `--managed-network`.
- Allow `AF_UNIX` socket creation in proxy-routed mode when `dangerously_allow_all_unix_sockets` is enabled, while preserving network namespace isolation and restrictions on other socket families.
- Keep standalone Unix sockets denied by default and for path-only grants. Default missing fields in older serialized contexts to restrictive values.
## Testing
Add coverage for policy preparation and transport, legacy deserialization, and malformed policy rejection. Add a Linux integration test covering default denial, path-only denial, and explicit allow-all access, while checking that direct TCP access and `AF_NETLINK`/`AF_VSOCK` sockets remain blocked.
GitOrigin-RevId: 2695b945ad3e59fcb3faf7662d852a26650af16c
## Why
Building MCP search entries eagerly cloned tool specs and normalized schemas even for tools that were never selected.
## What changed
Store search specs in `Arc<ToolSpec>` and let MCP search entries share the handler's spec. Materialize and normalize loadable specs only for selected results, preserving existing result formatting and dynamic-tool cache equality behavior.
## Testing
Add coverage for function, freeform, and namespace specs that verifies shared specs produce equivalent results, retain the source while needed, and release it when the search entry is dropped.
GitOrigin-RevId: 4260f2e2527834d8a856b1528654c6951aab8a7d
## Why
Command execution and plugin measurement events lack model and reasoning-effort labels. Attribution needs to reflect the step that invoked the command, even when model settings change before a background process finishes.
## What changed
- Add `model_slug` and `reasoning_effort` to command execution and plugin measurement analytics.
- Capture model context from resolved step settings and carry it through execution, approval, Guardian review, and plugin metrics collection.
- Preserve the first command-start model context when subsequent start notifications arrive.
- Keep the carried context out of serialized protocol items and generated schemas.
## Testing
Extend analytics tests to cover model switches before invocation and during background execution, default reasoning effort, Guardian-denied commands, and repeated start notifications retaining the original model context.
GitOrigin-RevId: af90e1c0d39bab625f2e89786085b61a9b96c0ce
## Why
MCP tool parsing eagerly cloned structured output schemas and built full call-result envelopes. Cloning tool definitions also copied their output-schema JSON, even before a consumer needed it.
## What changed
- Introduce `ToolOutputSchema` with immutable `Arc` storage so tool definitions share output schemas when cloned.
- Retain MCP structured output schemas and materialize the call-result envelope only when JSON is requested.
- Update code-mode consumers and schema mutation sites to materialize JSON explicitly, reusing uniquely owned storage when possible.
- Move structured content into the MCP envelope without an extra clone, preserving property order.
## Testing
Add tests for JSON preservation, mutation isolation, equality between lazy and materialized schemas, reuse of uniquely owned storage, and equivalent code-mode definitions.
GitOrigin-RevId: e98ba4c2f0efedc99f7cbc7bba206cc63a3bd8f4
## What changed
Include `sessionId` and the originating `windowId` alongside `threadId` and optional `itemId` in MCP request metadata. Retain the originating item and window for code-mode cells across waits and compaction, including the window when no matching history item is found.
## Testing
Extend metadata assertions for direct and nested MCP calls. Add a regression test verifying that a code-mode cell retains its original item and window IDs when it resumes after compaction.
GitOrigin-RevId: 5e3cf761f2bed4c16565334c94654496367325fd
## Why
Model and reasoning effort updates during a turn can leave request metadata and tool hooks reporting the turn's initial settings. Metadata should describe the step that issued the request or tool call.
## What changed
- Share captured execution metadata across Responses, MCP, and extension tool calls, including model, reasoning effort, and automatic review and Node REPL flags.
- Build Responses tool inventory metadata from the issuing step's finalized tool router. Attach the finalized inventory separately for remote compaction.
- Use captured step settings for pre- and post-tool hooks, and captured review settings for permission-request hooks.
## Testing
Add and extend regression tests for model and effort changes during a turn, MCP metadata, pre-tool hook model attribution, captured review flags, and tool inventory matching the issuing request.
GitOrigin-RevId: 9dee46a8b4839de8b434cdad4ed441cc0bb6ff9b
## Why
Direct tool-call records need to stay associated with the invocation that produced each output, including when call IDs are reused. Completeness must describe the recorded call inventory, independently of tool success.
## What changed
- Attach direct-call records to outputs before they enter history, and set `tool_calls_complete` when the invocation's arguments are fully recorded.
- Bound pending recordings and retained metadata, release reservations on completion or cancellation, and invalidate pending records when capture is disabled.
- Apply request budgets to direct metadata and strip it from inference and compaction inputs when capture is disabled.
- Remove executed-call metadata from app-server raw response notifications and exclude its size from Guardian history retention budgets.
- Track call IDs that bypass dispatch so their reuse cannot incorrectly establish Code Mode completeness.
## Testing
Add regression coverage for direct-call attribution, malformed calls, metadata budgets, cancellation, configuration changes, compaction, notification filtering, and Guardian context isolation.
GitOrigin-RevId: 2ebd39c7f141d04788736491495109841656b4c0
## What changed
- Add explicit MXC backend selection and carry its identity through exec-server process reporting and sandbox violation classification.
- Launch MXC through the Codex executable with the effective permission profile and command environment.
- Reject exec-server MXC requests when native MXC is unavailable or when they request a TTY, an `arg0` override, or managed networking. Reject private desktop isolation during MXC preparation.
- Allow an explicitly empty child environment and avoid exposing request payload values in launcher decode errors.
## Testing
Add coverage for sandbox selection and unsupported-request rejection, plus Windows RPC tests for stdin writes and temporary-directory permissions derived from the command environment. Native MXC tests skip when MXC is unavailable.
GitOrigin-RevId: 3626ff0f9ad7f9b812ce09b68c31ea9a5a9c72b1
## What changed
Add the disabled-by-default `send_message_to_user_async` feature flag so root agents can use the tool without model catalog support. Preserve catalog-based opt-in and keep the tool unavailable to subagents. Register the flag as under development and expose it in the configuration schema.
## Testing
Extend integration tests to cover feature and catalog opt-ins, deduplication when both are enabled, subagent exclusion, and the retired `send_async_message` flag. Check that either opt-in allows messages to be emitted without ending the turn.
GitOrigin-RevId: 3f53181ab074b0432c795550de73a9b1da7caf15
## What changed
Carry optional `available_access_programs` metadata through model information, presets, caches, and the TUI, and expose it as `availableAccessPrograms` in app-server `model/list` responses. Update the generated JSON schemas, TypeScript types, and Python models.
Preserve the distinction between missing metadata and an empty `cyber` list. Ignore unknown cyber program names when reading the catalog so new server programs do not prevent older clients from loading it. Discovery metadata does not grant access; inference still enforces authorization.
## Testing
Add coverage for absent, null, empty, and populated metadata, unknown program names, and app-server serialization. Extend cache tests and verify that online refreshes persist changed access metadata even when the catalog ETag stays unchanged.
GitOrigin-RevId: b3ef5805c1c61b8d64b9b896c9c0a79120143667
## Why
A yielded code mode cell can emit a notification after the session switches models. Its output should retain the originating model's truncation budget instead of using the receiving model's budget.
## What changed
Carry the cell's output token limit, including the serialization allowance, through notification dispatch and attach it as `history_truncation_token_limit` metadata. Allow `inject_if_running` to accept response item envelopes so queued notifications preserve this metadata in history and rollouts.
## Testing
Add regression coverage for delayed notifications across a model switch, verifying per-cell truncation budgets, full notification text in raw events, and persisted rollout metadata. Add a task-completion test that checks notification budgets are preserved alongside queued user input.
GitOrigin-RevId: 28380a039a6d41ea324401fcf1f611d81a279477
## Why
A code-mode cell can outlive its turn. Nested tool calls dispatched during a later turn need the original step's tools and context, while approval handling must still work across that turn boundary.
## What changed
- Retain each cell's originating `StepContext` for nested tool dispatch and cancel queued calls whose context has been released.
- Keep MCP approval metadata available for the lifetime of the running invocation, keyed by server and call ID.
- Apply Guardian denial accounting and interruption to the active turn, including reviews from an earlier cell.
- Carry `turn_id` in legacy MCP begin/end events so history reconstruction and rollout migration assign late completions to their original turn. Preserve the current-turn fallback for older records without it.
## Testing
Add regression coverage for originating tool selection across turns, callback and metadata cleanup, delayed MCP elicitation review, Guardian interruption of the active turn, and late MCP completion attribution with legacy compatibility.
GitOrigin-RevId: 4533379aae6e4ecb4aabe5a25e4aca5d65385d76
## Why
A delegate bound at session creation cannot provide different callbacks for cells sharing that session. Each execution needs to retain its own delegate across yields and release it when the cell is cleaned up.
## What changed
- Pass `CodeModeSessionDelegate` to `execute` instead of session creation.
- Route tool calls, notifications, and cell closure callbacks through the execution's delegate in the in-process runtime and the gRPC and stdio transports.
- Retain delegates with pending executions and live cells, releasing them through closure and cancellation cleanup.
## Testing
Add coverage for distinct delegates across yielded cells, gRPC callbacks before cell admission, and delegate release after completion or abandoned execution cleanup. Update transport tests to verify callbacks reach the owning cell's delegate.
GitOrigin-RevId: 7469f52104b993e790a40c65fb800ad02b7fd606
## Why
Tool results can be ready before the sampling loop collects them, and code-mode cells can outlive a turn. Separate trace milestones make these timings visible and link nested calls to the turn that dispatches them.
## What changed
- Emit `codex.tool_call_received` and `codex.tool_result_ready` events for direct and code-mode calls, using identifiers and tool names without arguments or output. Record readiness before result encoding or collection, including ordinary errors and aborted responses, but excluding fatal errors.
- Assign nested call IDs at broker receipt and carry them through dispatch. Trace dispatch and interruptions caused by cancellation, cell closure, or abandonment.
- Add code-mode execute and wait handler spans with conversation, turn, call, and cell identifiers and `completed`, `failed`, or `interrupted` outcomes.
GitOrigin-RevId: b098a29d8ee030c823665e4b1031a0d67cc09502
## Why
Delegated usage should retain the trigger of the turn that initiated it, including composer input and scheduled automation.
## What changed
Propagate `turn_trigger` through agent spawning and follow-up messages in both multi-agent tool implementations. Apply the supplied trigger when pending mailbox work starts a new turn.
## Testing
Extend tests to verify trigger propagation to spawned and nested agents, updated triggers on reused agents and peer follow-ups, and preservation of the active turn's trigger when steering.
GitOrigin-RevId: e8fea208587c633fec7cb3f1c0815ff91d1a44e5
## Why
Background commands can outlive their launching turn. Their network reviews need the original execution environment even when current turn settings change, and pending reviews must stop when the execution is cancelled.
## What changed
- Capture the execution's environment snapshot for Guardian network reviews while continuing to use current review settings.
- Deny requests from already cancelled executions before checking session host grants, and resolve pending approvals as denied when execution cancellation interrupts review.
- Preserve accepted review results and session grants when cancellation happens after acceptance.
## Testing
Add regression coverage for same-turn and later-turn settings changes, the original review working directory, cancellation before and after review acceptance, late callbacks, background process termination, and reuse of accepted session grants by new executions.
GitOrigin-RevId: 29487488da9ab43700ba6684c61eea86f30a03a8
## Why
Reused call IDs, ambiguous exec/wait associations, and missing history can make tool-call completeness claims unreliable. Truncating one cell's recorded calls should also preserve completeness for unrelated cells.
## What changed
- Track observed call and runtime cell IDs with bounded memory, including call IDs from initial history, and withhold completeness when evidence is missing or ambiguous.
- Validate exec/wait input-output associations and keep invalidated completeness from returning on retries. Withhold wait completeness after resume or fork, where runtime cell IDs can be reused.
- Keep truncation markers and omission counts local to each output, clearing completeness across the affected cell while preserving unrelated cells within the metadata budget.
- Refresh recording enablement from runtime configuration without changing execution features, and preserve the live recorder during legacy configuration reloads.
## Testing
Add regression coverage for reused IDs, compaction, resumed waits, mismatched wrappers, runtime recording changes, and metadata budgeting that preserves unrelated complete calls.
GitOrigin-RevId: 0bc85964bee079879620ec0664eea530a13b61fb
## What changed
- Add host-recorded `tool_result_metadata` snapshots with size limits, redacted debug output, and protection against deserialization from untrusted input. MCP capture remains disabled.
- Preserve accepted metadata across retries and delayed Code Mode results. Shed raw metadata before source evidence, calls, or completion markers when the request budget is exceeded.
- Send raw metadata only to allowed OpenAI and ChatGPT HTTPS endpoints, filtering by the resolved destination for HTTP and WebSocket requests.
- Encode source parse failures as an array entry with type `parse_failed` and an empty ID.
## Testing
Add coverage for metadata bounds and redaction, retained-call updates, budget shedding, disabled MCP capture, and HTTP/WebSocket destination filtering.
GitOrigin-RevId: b1a396323ad0623167cdd9e20ff9453f0d01bbf7
## Why
Failed environments were omitted from environment context, and `wait_for_environment` returned a generic failure or treated an already-failed environment as unselected. The model needs the failure reason to explain problems such as an empty repository.
## What changed
- Preserve the selected environment and error when startup or configuration fails.
- Include failed status and XML-escaped error details in environment context, and return the reason from `wait_for_environment` for failures detected before or during a wait.
- Limit error details to 256 bytes per environment and 512 bytes total in context, preserving UTF-8 boundaries. Limit wait failure details to 256 bytes.
## Testing
Add coverage for failures before a turn and while waiting, verifying that the model receives the reason and the turn completes. Add unit tests for XML escaping, incremental context updates, clearing errors on recovery, and UTF-8-safe truncation.
GitOrigin-RevId: 093ab9b447c7b3243fbb53b665180db95479f2a7
## Why
Saved model selections can outlive their bundled catalog entries and still need migration guidance.
## What changed
- Remove `gpt-5.2` and `gpt-5.4-mini` from the bundled model catalog.
- Preserve migration from saved `gpt-5.4-mini` selections to `gpt-5.6-luna` when the old model is absent, while respecting catalog-provided migration metadata.
- Update tests and snapshots to use current models or explicit test fixtures, and remove the obsolete prompt fixture.
## Testing
Extend migration tests to cover catalog metadata precedence, saved selections, replacement model and reasoning-effort persistence, acknowledged prompts, and missing or hidden migration targets.
GitOrigin-RevId: 82beea504cc4a15699677c28a61ec0939af3069c
## Why
Replaying tool outputs under a different model can expand or shrink the history shown to the model if truncation uses the new model's budget.
## What changed
Save the originating history truncation budget on function and custom tool outputs and reuse it during replay, preserving existing tool-specific overrides. Include the existing 20% serialization allowance once, before converting byte budgets to tokens.
Rename the metadata field to `history_truncation_token_limit` while retaining `fallback_token_limit_override` as its serialized name for compatibility.
## Testing
Add regression coverage for resume and fork with different model budgets, including custom tool outputs. Extend unit coverage for existing overrides and byte-budget conversion.
GitOrigin-RevId: a62c8fd3e43198e2252b150fba72f022939c1070
## Why
History recording used the model captured at turn start, so switching models within a turn could apply stale image preparation and tool-output truncation rules.
## What changed
Pass explicit model metadata through conversation recording. Record model responses and tool results using their originating step's settings, and capture the current model for pending input recorded during startup interruption or task completion.
## Testing
Add regression coverage for tool results across a model switch, explicit model selection for plain and annotated history, and image input retained during interruption and completion. Extend model-switching tests to verify that prepared media and item identity survive switching back to a multimodal model in live, resumed, and forked threads.
GitOrigin-RevId: 5bc750c30c334acb9b802cb80fdea01daffa84a1
## Why
After a model change within a turn, tools could still use the turn's original model capabilities, output limits, and telemetry attribution.
## What changed
- Build each step's tool plan from its captured model settings.
- Use those settings for image support and detail, command and MCP output truncation, extension metadata, and tool telemetry.
- Preserve the issuing step's model for delayed command completion events while allowing later `write_stdin` calls to use their own step's output limits.
## Testing
Add regression coverage for model changes within a turn, including tool replanning, Code Mode execution, image handling, MCP tool and resource output limits, extension settings, and delayed command completion.
GitOrigin-RevId: ef991598e89a846889baf770a0c9e12e56886d5d
## Why
Credential brokerage must retain trusted destination hints even when shell environment policy hides them from child processes. Snapshot replay also needs to preserve credential aliases when tokens are short or captured dummy values change.
## What changed
- Add `features.network_proxy.credentials` and preserve provider settings when toggling the proxy. Prevent project configuration from overriding providers or their credential and destination environment variables.
- Use provider metadata throughout snapshot capture and replay, and pass hidden destination context to the broker without exposing it in the child environment.
- Rewrite aliases containing short credentials or previously captured dummies, respecting environment filters and explicit overrides.
- Preserve case-distinct `shell_environment_policy.set` keys so Windows credential ambiguity checks survive configuration merging and permission profile changes.
- Restore independent MITM settings when credential brokerage is disabled.
## Testing
Add regression coverage for custom providers, hidden destination hints, Bash and Zsh alias replay, configuration edits and merging, Windows ambiguity handling, and credential protection during approved execution.
GitOrigin-RevId: b2bd935d2d8555be56baf8feb1a8e5b3a0c666bc
## Why
Shell quoting can hide credentials from raw-text checks, and startup files can restore real credentials after the broker replaces them with dummy values.
## What changed
- Decode shell literals without evaluating them and reject snapshots containing credentials in executable source, including aliases, functions, and heredocs.
- Preserve credential policy overrides, explicit unsets, and aliases whose source variables were removed. Support credential aliases in Zsh tied arrays while rejecting credentials that span array elements.
- Guard snapshot replay against credential restoration through shell startup files and preserve unrelated `ENV` settings.
- Apply Windows environment-key casing rules to credential overrides, suppress unredacted sandbox diagnostics during snapshot capture, and clear inherited environment variables before launching escalated commands.
## Testing
Add regression coverage for shell quoting and escaped credentials, Zsh tied arrays, startup-file replay, readonly credentials, policy overrides, and sensitive capture timeout and cancellation handling.
GitOrigin-RevId: 58274c07c715423241d26ce6dd2c2b4230cf0f64
## What changed
Introduce `ExecutedToolCalls` to own optional shared recorder state and centralize feature checks, call recording, and accepted result source lookup. Update session, tool runtime, MCP, and Code Mode callers to use this interface.
Move prompt attachment, retry handling, and metadata budgeting into the `request_metadata` module, preserving existing recording and request behavior.
## Testing
Add coverage for all session and turn feature combinations, lazy result source lookup, and unchanged prompts when recording is disabled. Retain the existing metadata budget and cell completeness tests in the new module.
GitOrigin-RevId: bc19f34317232b6599c7ec71c15db4a80123ff6f
## What changed
Parse optional `generation_id` values from image API responses and carry the selected image's ID through the image generation tool into analytics events. Keep the ID out of serialized extension items, JSON schemas, and TypeScript types. Responses without an ID remain supported.
## Testing
Add coverage for distinct IDs in multi-image responses and responses without IDs. Extend analytics and app-server tests to verify that the selected image's ID reaches analytics, and item tests to verify that it is omitted from serialization and TypeScript types.
GitOrigin-RevId: 70a600856990140b76fdbda51a0d73b3414338b1
## Why
Shell startup files can introduce credentials or overwrite brokered dummy values. Snapshot capture and replay need to respect the command's sandbox and environment policy without persisting real credentials.
## What changed
- Capture and validate brokered snapshots per command using its requested shell, working directory, login mode, sandbox, and environment policy.
- Discover shell-initialized credentials and prepare snapshot exports for brokered replay. Preserve shell functions and aliases while preventing startup files from overwriting dummy credentials during replay.
- Rebuild snapshots as credential broker state changes and retain each command's snapshot for the lifetime of its process.
- Reject brokered POSIX shell execution when a protected snapshot cannot be created, and reject brokerage with `shell_zsh_fork`. Restore real credentials and remove brokerage markers for unsupported PowerShell and Cmd execution; escalated commands bypass protected snapshots.
## Testing
Add regression coverage for credential discovery and redaction, environment filtering, readonly exports, shell function and heredoc replay, command working directories and login modes, sandbox enforcement, and missing or disabled snapshots.
GitOrigin-RevId: 685356cb9e80f11cf69d053e4d239ee6ccf43a9e
## Why
Code-mode callbacks can resume outside the turn's local span ancestry, leaving tool dispatch traces without the thread identity inherited from a turn span.
## What changed
Record `thread.id` directly on the `dispatch_tool_call_with_code_mode_result` span using the session's thread ID.
## Testing
Add a regression test that dispatches a code-mode tool callback from a fresh task without a turn span and verifies that the dispatch span records the thread ID, tool name, and an `exec-` call ID with a valid UUID.
GitOrigin-RevId: a25a0d8fcaf1b6501473151c486fc9940065df91
## Why
A code-mode `exec` wrapper and its nested tool calls each advance Guardian's score lag. Counting the unscored wrapper against its own nested approval can trigger a synchronous review sooner than needed.
## What changed
Track originating response items and approval tool-call IDs so cached approvals discount their own wrapper when it is newer than the latest score. Preserve wrapper identity across yields and waits, and retain the full lag when provenance is missing or evicted. Other wrappers and scoring failures still count.
Use the current `write_stdin` call's identity for approval freshness, and omit tool-call identity for `execve` intercepts that only retain the launch ID.
## Testing
Add coverage for wrapper-specific discounts, already-scored wrappers, missing and evicted provenance, identity across code-mode resumes, and approval identities for stdin writes and `execve` intercepts.
GitOrigin-RevId: 4002a06d64c4e1e0a13c072303503d6ddfacad09
## What changed
Resolve `GuardianContextMode` once at session construction and share it across history retention, replay, evidence capture, compaction, and review consumers.
Extract synchronous reviewer context policy and consolidate Guardian V2 parent checkpoint selection so sampling and fast approval use the same eligibility checks. Preserve legacy checkpoint omission and strict compatibility requirements for thread-owned context.
## Testing
Adapt existing review and history tests to the explicit context mode. Move checkpoint tests alongside the extracted selector, retaining coverage for latest-item identity, unusable checkpoints, and complete-item size limits.
GitOrigin-RevId: c73fb6fea556921acfffe758a5f709d7494019dd
## What changed
- Let approval extensions choose between cached approval, synchronous review, and a user prompt for tool and permission requests, while core enforces mandatory Guardian and fresh-review requirements.
- Carry an optional `review_reason` on Guardian assessment events and use it to trigger strict-review notifications, retaining the fallback for older events.
- Require synchronous review when cached evidence has an unusable encrypted parent compaction.
## Testing
Add regression coverage for manual prompts and cached approvals with non-UTF-8 working directories. Update approval contributors in existing tests to use the decision API.
GitOrigin-RevId: a4e66416070bdaa881daa41429c49021629f44c1
## Why
Experimental context activation previously checked the provider and account eligibility without checking model support. Child sessions also inherited token-budget activation from their parent, even when starting fresh with a different model.
## What changed
- Add `ModelInfo.supports_experimental_context`, defaulting to `false`, and enable it for the bundled `gpt-6-astra` model. Require this capability when activating experimental context.
- Snapshot configured token-budget preferences before startup activation. Restore them for fresh child sessions before applying their starting model's defaults, while history forks retain their parent's activation.
- Pass unresolved token-budget preferences to child sessions so they can use their own model's prompts.
## Testing
Extend coverage for unsupported models, model-switch guidance under explicit and experimental activation, and child configuration from both active and inactive parents. Verify that omitted capability metadata defaults to `false`.
GitOrigin-RevId: 02df9e171682267232fa923d5ea3f7af36527808
## Why
Guardian reviews need the original user instructions even when compaction or
transcript selection removes them from the active conversation window.
## What changed
- Capture complete user messages in host-owned retained context and include them
as ordered authorization evidence alongside verified answers.
- Omit oversized instructions atomically and emit an explicit incomplete-evidence
notice instead of exposing a potentially misleading partial permission.
- Mark user messages inherited by forked agents so they remain model-visible
without becoming child-local authorization.
## Testing
- Cover retained instructions across compaction, resume, rollback, and agent
forks, including oversized evidence and legacy context behavior.
GitOrigin-RevId: 97122509d818c90d2f87c6a1b29c99c05ad0316e