## What changed
Require the `openai` provider ID as well as `GuardianNativeMultiAgentHistory` to enable native agent history and review of multi-agent calls. Apply the same check to Guardian prompt construction, collaboration tool registration, and Guardian v2 configuration.
## Testing
Add coverage for fresh and resumed sessions with the feature enabled and disabled on OpenAI, and enabled on Bedrock and LiteLLM. Verify that only OpenAI sessions with the feature enabled request native review and include the agent call as evidence.
GitOrigin-RevId: bfbe547b0544e66763dca7d1e1376f805994bcb2
## Why
Guardian scoring and cached approvals need to use the requesting turn's captured policy even when the live thread configuration changes. Newly tightened mandatory review requirements must still prevent cached approvals.
## What changed
- Pass captured turn configuration through tool lifecycle and approval review inputs, and use it to resolve Guardian policy and classifier instructions.
- Apply current reviewer and model requirements before scoring and cached approval, including rechecking restrictions after collecting score authorization.
- Count computer-use JavaScript execution attempts even when background scoring is disabled, so a synchronous call consumes the initial-call allowance.
- Retain runtime configuration snapshots when applying agent role overrides, and share Guardian policy text resolution between configuration loading and runtime publication.
## Testing
Add and extend regression coverage for captured policy instructions, reviewer restrictions, stale cached evidence, and a synchronous computer-use call followed by an adaptive turn.
GitOrigin-RevId: 24ed4a1b6ce1733ad775790f30e440da614f9cd6
## 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
## 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
## What changed
Map provider `budget_reached` errors to `SessionBudgetExceeded` for HTTP 400 responses and streaming failures. End SSE streams immediately on this error, ignoring retry headers and later completion events, and display `budget reached` in the TUI.
## Testing
Add coverage for HTTP error classification, immediate SSE termination without retries, WebSocket error delivery and subsequent request recovery, and TUI error rendering and task-state cleanup.
GitOrigin-RevId: 6a62d699cdb9cb446214cacfacd0c9ed68bf21a6
Update the reviewer configuration test to call
`ManagedFeatures::from_configured_with_warnings` and pass
`parent_config.startup_warnings`, avoiding the crate-private
`from_configured` constructor.
GitOrigin-RevId: 2ba41ba6918e1007b3af013b7146940d919e4560
## What changed
- Share a lazily registered runtime across threads, model discovery, and extension clients when `runtime_agent_instances` is enabled, while honoring each thread's authentication mode and credential owner.
- Bind model catalogs to the effective provider and authentication context while preserving cache policy and live API-key discovery settings.
- Resolve credentials, routing, and transport policy together for each request. Preserve proxy and TLS settings, validate runtime account selectors, and prevent pooled connections from crossing runtime identities.
- Accept supported ChatGPT subdomains when selecting the registration environment.
## Testing
Add integration coverage for shared runtime registration, dual-header and assertion-only modes, disabled and third-party providers, credential replacement, subagents, and delegated review. Add coverage for catalog cache policy and runtime-aware connection reuse.
GitOrigin-RevId: 99dd9d987c29c93597a648f6586f98ac6a932c3c
## What changed
- Move approval-request and prompt-composition tests from `core` to `guardian-reviewer`.
- Move transcript collection, budgeting, and approval-reason truncation tests to `guardian-context`, removing the test-only rendering helper from `core`.
- Move the managed-feature reviewer configuration test to `guardian-v2`.
GitOrigin-RevId: 1886a10cd6aba07b7ca2bcce28d45e0a698111a6
## What changed
Map audio size, data, format, and sample-rate validation failures to `InvalidAudioInput` for HTTP 400 responses and streamed errors. Preserve the provider's message, falling back to `Invalid audio input.` when it is absent or blank.
Classify these errors as non-retryable and expose them as `BadRequest`. Redact their messages in doctor diagnostics and group them as invalid requests in telemetry.
## Testing
Add coverage for all six audio validation codes across HTTP errors, streamed `error` events, and `response.failed` events, checking error classification, message preservation, and the absence of retries. Extend doctor diagnostic redaction coverage for `InvalidAudioInput`.
GitOrigin-RevId: 0a59964bcabd83997bdb32fc9318a5482dac33b0
## 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
## What changed
Return `CatalogModel` from model lookups, keeping metadata together with its `Bundled`, `Provider`, `ExplicitConfig`, or `Synthetic` origin through config overrides and runtime snapshots. Replace `used_fallback_model_metadata` checks with explicit `Synthetic` origin checks.
Share the underlying `Arc<ModelInfo>` with extension stores and use copy-on-write for metadata changes. Keep origin tracking outside serialized model metadata.
## Testing
Extend coverage for origin retention across cache reloads, namespaced model lookups, catalog fallback, and config overrides. Verify that step snapshots and extension stores share the same model metadata allocation.
GitOrigin-RevId: bb6130c94a7a795aa168ee24664a1dd453579fba
## What changed
- Match final serving assessments to their response and output item, then retain them with tool invocations and code cells across waits.
- Use matching assessments instead of asynchronous scoring for adaptive review: low-risk actions can skip synchronous review, while elevated-risk actions require it.
- Revalidate captured authorization and policy before approval. Require fresh review for stale or incomplete authorization, preserve explicit review requirements, and discard assessments when hooks rewrite tool input.
- Keep asynchronous scoring as the fallback when an assessment is missing or does not match.
## Testing
Add integration coverage for direct and nested tool calls, hook rewrites, explicit escalation, interrupted responses, catalog-driven policy, incomplete authorization, and assessment retention across code-cell yields and waits.
GitOrigin-RevId: 4346a227204137f018c03c34b302d6f1e9592c96
## What changed
- Retain captured root conversation sources until synchronous preparation applies request limits. Keep asynchronous snapshots limited to the bounded messages.
- Move approval-reason truncation and tool-description escaping and truncation from composition into preparation.
- Extract synchronous image preparation and admission into `guardian-context`, preserving the prompt token cap and model context-window checks.
GitOrigin-RevId: d3a864c77d8f16bd3e77df436d80c4d2738c43cc
## What changed
Add `GuardianScope::for_multi_agent_tool` and use it for both tool review requirements and asynchronous scoring. Classify `spawn_agent`, `send_message`, `followup_task`, and `interrupt_agent` as native tools when they match the configured namespace, accepting default namespace spellings when none is configured. Keep review enablement checks at each call site.
GitOrigin-RevId: 9180ccdc629532f6c171d899dc5881e15b447141
## What changed
- Recognize `policy_capacity_unavailable` in HTTP and streaming API errors and expose it as `policyCapacityUnavailable` through the app-server protocol and generated bindings.
- Display “Policy-restricted capacity is unavailable. Please try again later.” while preserving overload retry behavior based on `Retry-After`.
- Handle the new error in TUI overload rendering, reviewer retries, and error telemetry.
## Testing
Add coverage for HTTP and streaming error classification, retry advice, WebSocket terminal error attribution, protocol serialization, and TUI rendering with and without an upstream message.
GitOrigin-RevId: 1b9536fba1e11cc891dcf689aee35410d82f5412
## What changed
When `GuardianNativeMultiAgentHistory` is enabled, require Guardian review for wrapped Multi-Agent V2 calls to `spawn_agent`, `send_message`, `followup_task`, and `interrupt_agent`.
Include these calls in Guardian's asynchronous native-tool scoring scope, honoring the configured multi-agent namespace and preserving that configuration across overrides.
## Testing
Extend the incremental-review scenario with an encrypted, namespaced `send_message` call. Verify that the encrypted parent reply remains present exactly once in both Guardian reviews and the subsequent agent request, and update the snapshot to cover native tool-call review evidence.
GitOrigin-RevId: 2af9561b0d69af4a8f878fa32365ce845368e02d
## Why
Native multi-agent calls must stay intact for backend decryption during Guardian review.
## What changed
When `guardian_native_multi_agent_history` is enabled for multi-agent v2, preserve `spawn_agent`, `send_message`, `followup_task`, and `interrupt_agent` function calls in the configured tool namespace as native transcript items. Use `functions` when no namespace is configured.
Include the paired evidence-only output in token accounting, and omit native evidence whole when limits require it.
## Testing
Extend transcript tests to cover native call ordering, paired-output budgeting, and whole-item omission. Add a native-history snapshot alongside the existing text-history delegation handoff scenario.
GitOrigin-RevId: e7d9b4f5ea772890bf14b79fdb53b0f9d0a842e4
## Why
Custom OpenAI base URLs caused Codex backend authentication to lose Responses headers and Guardian metadata when the URL did not match a recognized backend route.
## What changed
Apply model-scoped Responses headers, Guardian reviewer and classifier markers, and eligible ChatGPT parent credit requests based on the OpenAI provider and authentication configuration, regardless of the base URL. Keep API-key requests and other providers excluded from this behavior.
## Testing
Extend coverage for custom `/v1` URLs and direct `/api/codex` endpoints with ChatGPT, header-based, and API-key authentication, including HTTP fallback and WebSocket requests.
GitOrigin-RevId: cc4bb5b4a1a6596904ecf8965237c7a4f09e112b
## What changed
- Add `CoreToolRuntime::requires_guardian_review` to review opted-in function calls before dispatch and stop execution when review rejects the call.
- Deliver proposed calls as native review evidence, preserving encrypted argument metadata. When hooks rewrite arguments, review the updated arguments and discard the original encrypted replay content.
- Deduplicate native call evidence against reviewer history and include paired evidence outputs in input budgets.
- Keep review accounting and metrics for native calls without emitting action-specific client assessment events, and omit approval payloads from debug output.
## Testing
Extend regression coverage for rewritten native calls across review retries, denied calls that never execute their handlers, oversized native evidence, and review metrics without client events.
GitOrigin-RevId: c3b22c0220344581c2b38f233001abaed49cba34
## What changed
Collect original retained instructions, assistant context, and verified answers before applying synchronous request limits in `guardian-context` preparation. Carry original record sizes through collection, preserving whole-record omission notices, source ordering, and host assistant-context framing.
Apply retained transcript source size checks during synchronous transcript preparation as well.
## Testing
Extend regression assertions for synchronous/asynchronous retained-context parity and equivalent verified-answer omissions when oversized records are filtered during preparation.
GitOrigin-RevId: 9042ee3fe60ebd455f6656b4008efb2acc44ac3d
## What changed
Remove snapshot instruction deduplication, separate retained-assistant framing, and the associated omission-metadata special cases. Use `retain_new_instructions` in transcript rendering tests and replace the test-only Decisions snapshot adapter with the conversation adapter.
## Testing
Extend retained-source delivery and incomplete-metadata coverage to both synchronous and asynchronous context profiles. Keep Decisions HTTP, eviction, and cancellation tests exercising the conversation adapter.
GitOrigin-RevId: 3906f7d7349f73931e570ffa649377c3a235317a
## Why
The transport tests used a separate, test-only snapshot sampler instead of the production retained sampler, leaving them exercising different response ownership and cancellation paths.
## What changed
Route the transport test helper through `sample_retained` and remove snapshot execution, its active-request supersession bookkeeping, and test-only compaction compatibility fields. Simplify request-token metrics to reflect the remaining admitted-request path.
## Testing
Update tests to verify early score delivery while the retained response remains pending, WebSocket cancellation and connection reuse, and HTTP cancellation before headers or during response draining. Assert that HTTP cancellation releases classification permits.
GitOrigin-RevId: 33250103ff376af6bcd0f81ef27cdcf057f2bc5f
## What changed
- Always use the retained conversation backend for asynchronous classification; keep snapshot sampling test-only.
- Accept legacy `async_classifier_mode` values in configuration and model metadata but ignore them. Emit a startup warning when `[features.guardianv2].async_classifier_mode` is configured.
- Ignore legacy classifier mode differences when checking model-switch compatibility, and fail closed if the conversation backend is missing.
## Testing
Update coverage for legacy configuration warnings, model-switch compatibility, retained approval-review context, and conversation-based scoring.
GitOrigin-RevId: 5e8bb7e4e6e7d1589572f38bfe2498c9fa6825a0
## Why
Provider errors with code `model_not_found` should end the turn immediately and preserve the provider's message instead of triggering retries.
## What changed
- Recognize `model_not_found` in HTTP 400, 403, and 404 responses and in streamed `error` and `response.failed` events.
- Preserve the provider message and any available HTTP status, and expose `modelNotFound` through the app-server protocol and generated schemas and SDK types.
## Testing
Add coverage for HTTP classification, statusless stream errors, and terminal failures without retries across HTTP, SSE, and WebSocket responses. Verify that `thread/read` preserves the error reported by the live notification.
GitOrigin-RevId: 88f824ac03c8c9669915b1e40c67b76c08326fbe
## Why
A delayed conversation classification can reach execution after a newer request has already published a score and committed history, leaving it with no useful work to contribute.
## What changed
Pass the classification's sampling timestamp into conversation requests and check for newer scores and retained history before starting inference. Return `Superseded` only when both have advanced; keep requests that can still contribute either a score or history.
## Testing
Add regression tests covering supersession of scores and history separately and together, including a classification delayed by permissions. Verify skipped requests send no inference request and preserve cached scores, approvals, and score progress.
GitOrigin-RevId: ef78a59e2010a82ac4f1fb93568a562447187099
## Why
Decisions snapshot comparisons reject requests containing reasoning or compaction, while conversation comparisons already omit those items. A parent compaction also causes snapshot comparisons to be rejected.
## What changed
- Share filtering of reasoning, compaction, and context compaction across both comparison modes, and ignore parent compaction when building the Decisions request.
- Distinguish malformed request setup from unsupported input items. Record setup errors as failures and unsupported input items as skipped comparisons.
## Testing
Extend the HTTP contract test to cover both comparison modes with reasoning, compaction, context compaction, and parent compaction while preserving message content. Update assertions for the distinct setup and unsupported-input errors.
GitOrigin-RevId: e3ba0cce3b6457b3e3eab3589352b70288cbb1ef
## Why
Guardian V2 previously skipped optional Decisions comparisons in conversation mode as unsupported evidence.
## What changed
- Start comparisons from the prepared conversation input, including retained history or rebuilt evidence after a history reset.
- Extract the classifier rubric and preserve evidence message roles, content order, and boundaries. Omit retained reasoning and compaction items while continuing to reject other unsupported evidence.
- Measure baseline sampling through the usable score, before history draining and commit, and report skipped comparisons when preparation fails or the request is superseded.
## Testing
Extend coverage to both classifier modes for comparison disagreements, failures, and disabled or uninitialized comparisons without changing baseline scores. Verify conversation evidence across history reuse and resets, malformed input rejection, and budget failures before either backend sends a request.
GitOrigin-RevId: ce32cf876d15d3b82e0b83ac02c33447e7724a51
## What changed
- Rename the snapshot comparison entry point to `start_snapshot_comparison` and extract request spawning into `spawn_request`.
- Accept a `ContentItem` rubric and an iterator of evidence in the request adapter.
- Read the Decisions comparison feature flag once per classification and simplify the budget rejection test setup.
GitOrigin-RevId: b4a294eeecb6069db0f80ff6e4f13f558cbd7815
## 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
## What changed
Split retained conversation scoring into `Reservation::prepare` and `PreparedConversation::run`, exposing the complete admitted input before model execution. Keep the history fork and capacity permit owned by the prepared conversation, and recheck authorization, score expiry, and backend generation before execution.
## Testing
Extend conversation tests to verify that preparation sends no model request, execution uses the prepared input, dropping a prepared conversation releases capacity, and stale prepared reviews cannot start model execution.
GitOrigin-RevId: c96cc6fac69179a78ecd7bcdb8e5a4b3087aa6e6
## Why
Full-transcript fallback can select encrypted agent messages already present in the reviewer's history. These duplicates should not consume the input budget or displace new evidence.
## What changed
- Reconcile native agent messages with the effective reviewer history and earlier messages in the pending input before budgeting, including after compaction.
- Deduplicate messages with nonempty IDs while preserving messages without IDs. Allow delivery again when the prior copy is absent from the current history.
- Reject conflicting authors, recipients, or content under the same ID without modifying the pending context.
## Testing
Add coverage for duplicate admission, conflicting payloads, redelivery after history removal, and retained reviewer requests. Extend the history eviction regression to verify that full-transcript fallback reuses the reviewer and preserves a single encrypted agent message.
GitOrigin-RevId: 97e90f6359a7cf5a75d1c5967fc98cd885afb4e4
## What changed
- Count adaptive reviews and record synchronous review routes as `async_fallback` or `required_sync`, with the review reason in snake case.
- Record cached score readiness and score age in milliseconds, measured from input sampling time and including inference time.
## Testing
Extend approval tests to assert fallback routing, adaptive review counts, score readiness, and sampled score age.
GitOrigin-RevId: 1200b3297add6c4467e460de87b4a6d8d31e1fc4
## What changed
- Add sampling retry counters that distinguish transport retries from authentication recovery, with failure reasons.
- Record attempt counts and final outcomes for successful, failed, cancelled, and superseded sampling requests.
- Carry retained-score telemetry until the consumer accepts or discards the score, so queued scores that expire are recorded as superseded and accepted early scores remain successful even if response draining fails.
## Testing
Add metric assertions for transient recovery, retry exhaustion, authentication recovery, cancellation, supersession, and retained-score acceptance or discard, including expiry while a score is queued.
GitOrigin-RevId: 36babd6f431f22cf15b73355e6e026a256a81004
## Why
Guardian request counts grouped by input size should reflect admitted sampling inputs without counting synchronous transport retries more than once.
## What changed
- Add a shared `token_bucket` label to admitted total-input observations in `codex.guardian.context.request_tokens` for synchronous and asynchronous reviews. Use inclusive upper bounds in 20k-token steps through 240k, followed by 256k and overflow buckets.
- Record synchronous request tokens after budget admission, once per sampling step, including the complete input. A fresh step counts again even when its input is identical.
- Preserve asynchronous component histograms for oversized classifications, but omit `token_bucket` so rejected inputs do not contribute to counts grouped by that label.
## Testing
Add coverage for bucket boundaries, synchronous retry deduplication and budget rejection, asynchronous retained-history reuse and reset, and admitted versus oversized classification metrics.
GitOrigin-RevId: fac2f102a4c48c5fd21d56bfa701b95a1ff852d0
## Why
Cloning `ResponseItem::AdditionalTools` deep-copies its tool declarations for history and request copies.
## What changed
Store `tools` in an `Arc<[serde_json::Value]>` so clones share immutable declarations. Update constructors and existing test fixtures to use the shared representation.
GitOrigin-RevId: eb8831774afb45fe8431a796f34ae9d1243e884a
## Why
Contextual fragments and extension-owned world-state sections currently return strings, limiting them to text content.
## What changed
- Return `ContentItem` from fragment bodies and rendering, preserving classification and attribution. Apply text markers only to `InputText`; pass other content through unchanged.
- Allow extension prompt and world-state contributions to carry typed content, and preserve media in context-window hints.
- Hash non-text world-state fragments by their serialized content and role while retaining the existing text hashing behavior.
- Add shared content-to-tool-output conversion that preserves image references, image detail, and audio.
## Testing
Add coverage for typed world-state hashing and preservation of image content and attribution. Expand code-mode output tests to cover input text, output text, file-backed and inline images, and audio.
GitOrigin-RevId: 2ed00d466a3f169b151202f3eef71e667ea64be1
## Why
The optional Guardian Decisions comparison rejected non-user messages and trusted skill evidence, and folded trusted tool evidence into the classification rubric.
## What changed
Send evidence as separate messages with their original roles, content order, boundaries, and optional `phase`. Support both `input_text` and `output_text`, keeping trusted tool and skill evidence separate from the rubric. Continue to accept inline images only in user messages and reject unsupported evidence.
## Testing
Update adapter and HTTP contract tests to cover mixed message roles and assistant output. Extend lifecycle coverage to verify that trusted skill evidence reaches Decisions while comparison results leave the baseline risk unchanged.
GitOrigin-RevId: 1b6814fe6db5891fb61604d6846a008c9e456054
## What changed
- Add `ComposedContext::retain_new_images` and use it in Guardian review sessions and the async scorer. Compare host-prepared image references against admitted history, ignoring image detail, and skip scanning history when no images need checking.
- Restore `ConversationState` directly from a checkpoint, retaining committed history for rollback or transferring it with `take_history` for ephemeral forks.
## Testing
Extend tests for prepared-image matching across differing detail levels, image-free continuations that avoid scanning history, and checkpoint restoration and history transfer.
GitOrigin-RevId: bafeb6de9babc721182be43d0eb1a653955a0ebb
## Why
Unsupported evidence currently shares a single Decisions comparison reason, making it difficult to distinguish why a comparison was skipped.
## What changed
Report separate reasons for unsupported developer messages, other non-user messages, trusted skills, images, and audio. Keep these outcomes classified as `skipped` and retain the generic reason for other unsupported evidence. Expose `TrustedSkills::KIND` for identifying trusted skill fragments.
## Testing
Expand adapter tests for the distinct rejection reasons and add a lifecycle case verifying that trusted skill evidence skips the Decisions request, reports `trusted_skills`, and preserves the baseline score.
GitOrigin-RevId: 98f076d1aa6771d0209e7f5910eed0dcc40c6263
## Why
Message roles and content classifications do not identify who supplied each context part. Source attribution needs to survive prompt assembly and history truncation, including messages that combine text from different producers.
## What changed
- Add per-part `ContentItemMetadata` for user input, developer instructions, `AGENTS.md`, skills, harness context, command hooks, and tool context with optional callable namespaces.
- Carry attribution alongside content and classifications through contextual fragments, extension prompts, hook output, and agent messages. Keep tool-provided context-window hints separate from harness text.
- Preserve the source of retained text slices during compaction and attribute generated truncation notices to the harness. Default missing historical attribution to unknown.
- Attribute host MCP calls using cached tool namespaces only when catalog revisions remain stable, without starting servers or refreshing catalogs for attribution.
- Send the new metadata to Responses Lite and strip it from requests to endpoints that do not support it.
## Testing
Extend compaction tests to cover mixed sources, missing or extra metadata entries, and discarded media. Update Responses Lite request assertions to verify endpoint-specific attribution handling.
GitOrigin-RevId: b0ebc929b7c60f3ec49d1affbf897312d63edcc8
## What changed
Use `conversation` when neither a local `async_classifier_mode` override nor a model default specifies a mode. Keep `snapshot` available as an explicit override and update the configuration documentation and schema to describe the new fallback.
## Testing
Expand configuration tests to cover all local override and model default combinations. Update app-server approval tests to exercise conversation mode without an explicit override, and explicitly select snapshot mode in tests that depend on it.
GitOrigin-RevId: a62479a3f6454bf28d589373a236d8b0ab5023b9
## Why
Pending classifications can outlive the `max_tool_call_lag` window in which their scores are reusable, keeping metadata waits or classifier transports active after the result becomes stale.
## What changed
- Expire pending classifications when tool-call progress exceeds the configured lag allowance, preserving the wrapper discount while nested code-mode actions remain active.
- Cancel metadata waits and sampling in snapshot and conversation modes, including snapshot response drains after an early score.
- Treat expiry as supersession, preserving cached scores and the last completed conversation checkpoint and transcript cursor.
## Testing
Add regression coverage for lag boundaries, active wrapper discounts, transport cancellation before and after early scores, expired requests skipping preparation, and conversation history and transcript deltas surviving cancellation.
GitOrigin-RevId: 892ff13a05fd2b1573cef9842b934d1fc865d951
## Why
Canceled classifier requests need bounded cleanup that preserves received token usage and returns only safe WebSocket connections to the pool.
## What changed
- Own classifier response streams and connection leases together, interrupting dropped responses with a one-second cleanup deadline.
- Require a matching interruption acknowledgement and response completion before reusing a canceled response's socket. Retire sockets on rejection, invalid acknowledgements, or timeout while preserving any received completion and token usage.
- Skip requests canceled before dispatch and return their unused sockets to the pool.
- Distinguish cancellation from ordinary steering, which continues draining without requiring an interruption acknowledgement.
## Testing
Add regression coverage for cancellation, acknowledgement ordering and failures, cleanup timeouts, socket reuse, and token accounting under backpressure. Extend steering tests to cover rejected and missing acknowledgements.
GitOrigin-RevId: c4e7aac49d36cea2c2d0e9f2ba5f6147f67471dd
## Why
Encrypted compaction payloads can include sampled-token replay caches that increase their size roughly threefold. Budgeting them as ordinary text can reject checkpoints that fit the model context budget.
## What changed
Use a shared estimate of one token per twelve ciphertext bytes, rounded up, for encrypted `Compaction` and `ContextCompaction` items in history and Guardian context budgeting. Guardian continues to count framing and metadata separately.
Check parent compactions against the configured token budget and a separate hard wire-size limit. Preserve the complete encrypted checkpoint when it fits. The token estimate remains a heuristic and can underestimate legacy checkpoints without replay caches.
## Testing
Add coverage for token-estimate rounding, parent compaction token and wire-size limits, and reviewer continuation with a cached compaction preserved intact.
GitOrigin-RevId: cc18ca29802491865167eef992830efcf93d97da
## What changed
Allow Guardian Decisions to use `OPENAI_API_KEY` when the default `openai` provider has no custom base URL and neither `CODEX_GUARDIAN_DECISIONS_API_KEY` nor a cached API key is available. Preserve cached API key precedence over the environment fallback.
Report missing credentials as `credentials_unavailable` instead of `provider_error` in Decisions failure metrics.
## Testing
Add tests for cached API key precedence and the environment fallback with ChatGPT authentication. Run each test in a child process to isolate environment changes.
GitOrigin-RevId: fbb772dcb9287d9e1be1b34292ad9bdffeb3d6b3
## Why
Incomplete responses such as `max_output_tokens` were classified as stream errors, allowing retries and compaction model fallback after the response had hit its output limit.
## What changed
Introduce `ResponseIncomplete` errors that preserve the server-provided reason and stop stream processing immediately. Make these errors non-retryable and stop compaction without falling back to the selected model. Record incomplete responses separately in sampler failure metrics.
## Testing
Add regression tests that verify incomplete responses surface a terminal error without retries over SSE and WebSocket, and that compaction stops without retrying or switching models for both local and remote compaction.
GitOrigin-RevId: 6e107d441b9bf03fa47ea47b88eeb94b71a9f729
## Why
The generic `parent_compaction_error` classification metric does not distinguish serialization failures, oversized compactions, and unusable compactions.
## What changed
Record separate `failure_reason` values for each case: `parent_compaction_serialization`, `parent_compaction_oversized`, and `parent_compaction_unusable`.
## Testing
Extend parent-compaction reuse tests to cover unusable checkpoints in parent-context mode and assert specific failure reasons. Verify rejected compactions receive the maximum risk score, provide no cached approval, and do not reuse stale context.
GitOrigin-RevId: 103c4173b2ebbdc33eb71bce7f8ecb56f73055f6
## Why
Reading `OPENAI_API_KEY` directly bypasses login credentials reloaded by `AuthManager`.
## What changed
Use the API key from `AuthManager`'s cached authentication for Guardian decisions when the provider is `openai` and has no custom base URL. Keep `CODEX_GUARDIAN_DECISIONS_API_KEY` as the highest-priority override.
## Testing
Add a regression test covering a reloaded API-key login, override precedence, and rejection of the login-key fallback for a custom base URL. Assertions avoid printing credentials on failure.
GitOrigin-RevId: 10b2e6c28d3268c3808c25b2be94d9b6c31f597a
## 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
## What changed
Pass the parent turn's `cyber_access_program` to Guardian reviewers and asynchronous classifiers. Use `daybreak_blue` for either Daybreak selection, preserve `standard`, and omit `access_programs` when no program is selected.
Capture the selection with the owning turn so asynchronous observations retain it across later turns. Resolve classifier programs using current authentication and the same provider and API-key feature policy as ordinary model requests.
## Testing
Add regression tests for program changes across reused reviewer sessions with API-key and ChatGPT authentication, snapshot and conversation classifiers at low and high risk, and enabled, disabled, and unsupported-provider API-key policies. Verify parent requests retain their selected program.
GitOrigin-RevId: 447f306df3a6bd6805464b819950690c751b04b4
jif