- Tag locally synthesized missing-finishReason failures as stream-truncated in the Gemini and Vertex executors, and carry that classification through the auth manager so it skips credential cooldown and stops cross-credential replay while the public 502 and partial usage are preserved.
- Carry the truncation classification through usage reporting and the Home queue payload so Home-side schedulers also skip the spurious cooldown, and add regression coverage for both the manager and usage paths.
Closes: #7005
- Fall back to schema prompts for Antigravity Claude 5.5 while preserving unrelated request settings.
- Normalize unsupported forced tool choices before thinking preparation in direct Claude execution and token-count paths.
- Add regression coverage for schema integrity, model boundaries, retries, and final payload rule precedence.
Closes: #7000
- Compress finalized Codex OAuth request bodies with zstd on the verified ChatGPT backend path, with size/benefit gating, header-body alignment, and one uncompressed retry on 400/415 rejection.
- Negotiate downstream HTTP response compression under server.response-compression (zstd/gzip with q-values, SSE-safe flushing, commit-time Vary merging) and enable negotiated permessage-deflate on Codex and downstream WebSockets.
- Advertise Accept-Encoding to Codex upstreams and decode gzip/deflate/br/zstd upstream responses through a shared decoder before parsing or relaying.
Closes: #6905
- generate a unique selection_id for each handled Scheduler.Pick decision and freeze it into the execution context and UsageReporter
- propagate SelectionID through core usage records and pluginapi.UsageRecord while clearing stale IDs on unhandled, rejected, Home, and credits fallback paths
Closes: #6988
- Aggregate -{level}-lightning compound suffixes into a single swe-2-lightning catalog entry and preserve the Lightning display name suffix across registry and fetch_devin_models.
- Resolve -lightning base models into stem-{effort}-lightning UIDs while passing through complete upstream lightning UIDs and regenerating the embedded Devin catalog.
Closes: #6975
- Apply canonical thinking normalization, suffix stripping, and plugin model metadata fallback in prepareExecutorCall across Execute, ExecuteStream, and CountTokens.
- Resolve plugin-registered ThinkingApplier by provider key before format fallback and validate unknown thinking suffixes on non-native-Responses models.
- Run thinking normalization before NewPayloadFinalizer so user payload rules act as the final host-side mutation.
Closes: #6940
- Populate Detail.ResponseServiceTier from usage.speed in parseClaudeUsageNode so non-streaming and streaming Claude usage records carry the billed response speed tier.
- Add regression tests verifying non-streaming usage.speed extraction and retention across message_start and message_delta stream events.
Closes: #6884
- Capture the post-translation, post-payload-rule service tier on UsageReporter and propagate EffectiveServiceTier through SDK usage records, plugin usage adapters, and Redis usage queue payloads.
- Ensure executors record translated request metadata after final payload configuration rules so server-side service_tier overrides are accurately reported alongside client and response tiers.
Closes: #6873
- Route multi-region locations us and eu to https://aiplatform.{us,eu}.rep.googleapis.com in both native Gemini and OpenAPI Vertex host builders.
- Recognize official Vertex REP domains in IsGoogleVertexHost so location rewriting and credential checks handle REP base URLs accurately.
Closes: #6867
- Pass the active configuration to ApplyClaudeLegacyDeviceHeaders under cloak_mode never while allowing plausible newer SDK and runtime versions for confirmed clients.
- Document that cloak_mode never preserves confirmed caller headers ahead of stabilize-device-profile pinning and add regression coverage.
Closes: #6863
- Treat cancellation after an emitted terminal event as successful completion while retaining observed token usage.
- Preserve buffered usage on genuine stream failures and share terminal detection with Vertex.
- Add regression coverage for terminal disconnects, EOF, usage tails, and read errors.
Closes: #6821
- Add per-model and per-credential openapi and openapi-location settings across Vertex API-key configs, OAuth model aliases/settings, and service-account auth files with precedence over automatic publisher detection.
- Preserve alias-specific OpenAPI overrides across credential prefixes, thinking suffixes, and mixed model arrays while synchronizing explicit OpenAPI location overrides in full and short base URLs.
- Populate publisher ownership and default reasoning levels for Vertex OpenAPI models registered via vertex-api-key configurations.
Closes: #6763
- Allowlist the sdk-ts Agent SDK entrypoint and recognize its measured Haiku 5.5 helper requests so genuine Claude Code traffic passes through uncloaked.
- Preserve plausible newer Stainless SDK and runtime versions for confirmed clients and retain confirmed-shape caller headers when cloak_mode is never.
- Recognize the Agent SDK CLI identity prompt block alongside existing Claude Code identity strings.
Closes: #6812
Responses reasoning emitted only in response.output_item.done was ignored
until a later token or terminal event. Count non-empty summary and content
text in the shared classifier while excluding empty and encrypted-only
items.
Add deterministic coverage for Meta Responses and Claude downstreams,
classification boundaries, and preservation of an earlier token time.
Refresh the bootstrap-buffering comment to match the shared helper.
Refs #6664
SGLang's OpenAI-compatible server emits reasoning tokens flat on the
usage object instead of nesting them in completion_tokens_details /
output_tokens_details, so the keeper recorded reasoning_tokens=0 for
affected upstreams. Add a third fallback plus bucket detection for the
flat key; regression included.
Fixes#6792
- Intercept compaction_trigger items on Kimi and Meta OpenAI Responses execution paths to synthesize sealed CPA compaction capsules and reject incomplete upstream summaries.
- Expand replayed CPA compaction capsules into developer context messages and drop foreign compaction items across Kimi and Meta execution and token-counting paths.
Closes: #6781
- Store and retrieve cached Claude web search results via Home KV with sliding TTL expiration when Home mode is enabled.
- Propagate request context through Claude web search result recording and restoration helpers in execution and token-counting paths.
Closes: #6780
- Strip incoming Claude Code billing attribution, CLI identity system blocks, metadata user IDs, and diagnostics when a non-CLI API key explicitly configures cloak mode as never or uncloak.
- Remove Claude Code CLI headers, beta flags, and user-agent strings on uncloaked API key requests while preserving capability betas and credential header overrides.
- Keep global disable-claude-cloak-mode passthrough behavior unchanged for credentials without explicit per-key cloak configuration.
Closes: #6762
- Route Vertex partner/open MaaS models and explicit openapi-enabled credentials through the OpenAI-compatible endpoints/openapi endpoint while reusing service-account OAuth token acquisition and supporting openapi_location overrides.
- Register Vertex service-account MaaS models and aliases from credential metadata and OAuth alias rules so partner models are discoverable and routable.
- Normalize Vertex OpenAPI model identifiers and reject static API keys against official Google Vertex OpenAPI hosts that require OAuth tokens.
Closes: #6763
- Implement LRU cache and helpers for Claude native web search tool results.
- Record search results from stream events and responses, and restore them in subsequent requests.
- Add configuration options and test coverage for Claude web search result caching.
Closes: #6753
- Only preserve OpenRouter dot-delimited router suffixes when the upstream baseURL is openrouter.ai / *.openrouter.ai
- Strip RouterSuffix down to the native GPT/Grok/Meta envelope when forwarding to official tokens (chatgpt.com, api.openai.com, api.x.ai) or non-OpenRouter third-party providers (while keeping native api.meta.ai fail-closed against cross-caller envelopes)
- Strip RouterSuffix before storing GPT signatures in the Codex reasoning replay cache
- Add `ParseClaudeRateLimitResetForError` to inspect response bodies and avoid inheriting hours-long unclassified unified resets for explicit per-minute rate limits.
- Update Claude standard and fast-direct executors to use the body-aware rate-limit parser when handling upstream error responses.
Closes: #6719
- Introduce Devin quota exhaustion detection for daily and weekly usage windows with reset timestamp parsing.
- Apply credential-scoped 429 status errors with reset-based Retry-After for Devin quota limits.
- Support automatic quota revalidation and early credential cooldown release upon confirmed usage recovery.
Closes: #6690
- Add `session-date-pin-days` configuration option with a 7-day default retention.
- Store session date pins independently of continuity state to prevent re-anchoring on credential switches or idle gaps.
- Strip caller-supplied current date reminders from the first user message to avoid duplication.
Closes: #6682
- Translate Gemini `MALFORMED_FUNCTION_CALL` responses into Claude API error payloads instead of refusal stop reasons.
- Capture translated stream errors in Gemini and Vertex executors to ensure proper error propagation, usage reporting, and state management.
Closes: #6606
- Introduce `sdk/httptransport` package with `Decorator`, `WithDecorator`, `WithoutDecorator`, `InheritDecorator`, and `Decorate` functions.
- Propagate transport decorators from incoming HTTP request contexts in `BaseAPIHandler.GetContextWithCancel`.
- Apply transport decoration after client transport selection across proxy-aware, Devin, uTLS, and Antigravity HTTP clients.
Closes: #5616
- Inject API key, access provider, and native key flags into the request context during authentication
- Add `IsNativeKey` and `AccessProvider` fields to usage records, plugin API types, and redis queue payloads
- Provide context helper functions and automatic fallback resolution in the usage manager and reporter
Closes: #6509
- Normalize tool use and tool result IDs in Claude message payloads to adhere to character constraints
- Encode non-portable tool call IDs in Devin executor responses and restore original IDs before request construction
Closes: #6478
- Handle compaction trigger requests by generating summaries and sealing antigravity compaction capsules
- Expand replayed compaction capsules into message history before translation and filter foreign compaction items
- Preserve tool definitions with `tool_choice: none` or flatten tool blocks when tool definitions are missing
- Support streaming compaction execution and map cache creation and read token usage
Closes: #6466
- Pass configured proxy URL to the npm registry updater client
- Validate fetched versions against strict semver and minimum server version floor (`1.0.13`)
- Update default fallback client version to `1.0.46`
- Manage background updater lifecycle and context via service lifecycle initialization
- Track dropped attachments and reject emptied user turns during OpenAI Responses to Interactions translation
- Register checked request translator for OpenAI Responses to Interactions format conversion
- Fetch latest @xai-official/grok client version from npm registry on startup
- Periodically refresh Grok CLI version every 3 hours without multiple retries
- Update hardcoded fallback version to 1.0.50 if npm registry is unreachable
- Dynamically inject active version into x-grok-client-version and User-Agent headers
An attachment a translator could not represent was dropped without a word, and
a turn whose only content was that attachment went upstream as an empty
messages, parts, or input array. The model then answered as if nothing was
attached, or the upstream rejected a request that never named the attachment.
A base64 document now reaches OpenAI as a file part. A file the target cannot
receive, such as container_upload, an OpenAI file_id, or input_audio sent to
Claude, is reported by name so the route answers 400 instead of forwarding an
empty turn. Text beside it still travels.
One rule, ErrIfNothingLeft, decides when to refuse, and RegisterCheckedRequest
carries the refusal on the request envelope. Every executor that translates
checks it before calling upstream. The exported converters keep their []byte
signature, so no caller or sibling test had to change.
- Buffer and publish observed upstream usage when apply patch transformation or translation encounters an error in Meta and xAI executors
- Add stream usage buffer helpers to report failure metrics with captured upstream token usage
Closes: #6427
- Track pending completion state in OpenAI-to-Responses translator when item output is complete and a finish reason is received
- Synthesize terminal `[DONE]` event on clean EOF when response completion is pending instead of failing the stream
Closes: #6381
