- Add example payload filter rules in `config.example.yaml` for stripping tools from both flat and nested `additional_tools` Codex request payloads.
Closes: #5792
- Define a managed Claude beta set to distinguish proxy-governed betas from caller extensions.
- Forward unmanaged caller betas on direct Anthropic endpoints to support newer client features.
Closes: #5738
- Avoid acquiring write mutex in ping handlers so keepalive pongs reply immediately during active writes.
- Stream payload messages larger than 32KB in chunks using writer streams.
- Support ephemeral websocket sessions for sessionless execution and enrich disconnect logs with session kind and terminal event context.
Closes: #5734
- Add kimi-k2.8 and kimi-k2.8-code with 1M context, 64k completion tokens, low/high/max thinking, and zero_allowed support to models.json.
- Remap K2.8 aliases (kimi-k2.8, k2.8, kimi-k2.8-code, k2.8-code, and -preview variants) to upstream canonical kimi-for-coding.
- Normalize temperature for Kimi upstream to prevent 400 errors (strict 0.6 for disabled thinking, 1.0 for enabled).
- Align zero_allowed: true across K2.8 and K3 models based on live upstream verification of thinking.type=disabled.
- Add test coverage for model normalization, thinking replay family, temperature stripping, and Claude effort=max preservation.
- Derive billing fingerprint message text from the first user message to prevent prompt cache invalidation across turns.
- Centralize continuity tag resolution to maintain stable prompt IDs and request references across requests.
Closes: #5730
- Inspect `input` items for the latest `configuration_update` reasoning effort before falling back to top-level configuration.
- Route `codex`, `xai`, and `openai-response` providers through usage-specific thinking config extraction.
- Support `none`, `auto`, and explicit reasoning level modes from in-turn updates.
Closes: #5714
- Add `BaseURL` field to usage records and host auth file entries.
- Extract `base_url` from auth attributes and metadata during usage reporting.
- Propagate `base_url` through plugin usage adapters and runtime auth callbacks.
Closes: #5693
- Synthesize function response parts for interrupted or missing OpenAI Responses tool calls to maintain strict Gemini call-response pairing.
- Preserve function response ordering matching pending call IDs across parallel and partial tool execution turns.
- Degrade gracefully to the original request payload when Antigravity reasoning replay breaks Gemini function call pairing.
Closes: #5682
- Map upstream `model_not_found` errors to HTTP 404 before evaluating generic invalid request types in Codex terminal error handling.
- Prevent treating structured model not found responses as client request faults to preserve credential rotation.
- Recognize model access denial errors to apply model-level cooldown and failover.
- Respect `disable_cooling` configuration during model-level cooldown processing.
Closes: #5635
- Broaden pattern matching for Codex model capacity errors.
- Classify model capacity rejections as overload bootstrap failures to enable failover.
Closes: #5634
- Format streaming error payloads with nested error objects matching official OpenAI Responses SSE specifications.
- Extract and propagate sequence numbers from upstream terminal events and framer states.
- Use `json.Number` to prevent precision loss for large integers and token metrics.
- Sanitize sensitive keys recursively across nested error objects without dropping custom fields.
- Retain `cache_control` blocks with 1h TTL and `extended-cache-ttl` beta header when explicitly requested by subagents.
- Detect 1h TTL configuration from request payloads and incoming Anthropic-Beta headers.
- Ensure `extended-cache-ttl` beta header is preserved or injected when 1h TTL is present.
Closes: #5629
- Register builtin model definitions for `gpt-image-2.5`, `gpt-image-2.5-flare`, and `gpt-image-2.5-sunburst`.
- Update OpenAI image handlers and request routing to recognize GPT Image 2.5 models.
- Support direct image generation and edit execution for GPT Image 2.5 variants in the Codex executor.
- Apply client visibility overrides to hide new builtin image models where appropriate.
- Initialize `util.SessionIDResolver` to resolve session IDs from request contexts, metadata, and headers.
- Ensure canonical session metadata is injected into execution options across execution flows.
- Propagate session context in executors to support `$CPA-SESSION-ID` expansion in custom headers.
- Sync cleared or updated session identities back to execution contexts during conductor execution.
Closes: #5690
- Check for \u in fast-path check to prevent JSON Unicode escapes from bypassing inspection.
- Inspect regex keys under patternProperties and drop keys with unsupported Unicode property escapes.
- Add tests covering Unicode escape representations (\u005c, \u0070, \u0050) and patternProperties keys.
- Add HasUnsupportedUnicodePropertyEscape in internal/util to detect \p{...} / \P{...} escapes that fail Python re compilation.
- Strip incompatible pattern attributes during tool parameter normalization in codex/claude and openai/claude translators.
- Provide schema-aware fallback stripping in codex executor helps to protect downstream Codex requests without mutating non-schema user data.
- Export unified schema keyword lists in internal/util to eliminate duplication.
- Add comprehensive unit tests covering Artifact fixtures, lookaheads, and user data preservation.
Closes: #5644
- Add `model-level-cooling` configuration option to Codex settings.
- Scope `usage_limit_reached` quota cooldowns to the requested model instead of the entire credential when enabled.
- Propagate model-level cooling checks across HTTP, SSE, and WebSocket execution paths.
Closes: #5619
- Emit an OpenAI-compatible trailing usage chunk with an empty choices array on `message_stop`.
- Include `cache_write_tokens` in prompt token details for OpenAI response translations.
- Parse usage from `message.usage` and buffer Claude stream usage across chunks to merge input and output token counts.
- Ensure observed streaming usage details are published on completion or stream failure.
Closes: #5617
Repair placeholder functionResponse names with part-local sjson and one
Index splice of the request body, matching the sibling content-edit path
instead of one full-body sjson.SetBytes per repaired name.
- Add shortMode, idleConnTimeout, and maxIdleConnsPerHost to antigravityTransportKey.
- Prevent stale transport reuse during hot-reload race windows under load.
- Add TestAntigravityTransportKeySeparatesPoolSettingsAcrossReload regression test.
- Route `openai-response` source format requests to dedicated streaming and non-streaming responses handlers.
- Add upstream URL resolution for Kimi Responses API endpoints based on auth attributes.
- Normalize `openai-response` provider format to codex during thinking parameter resolution.
Closes: #5400#5571
2.1.258 attaches x-client-request-id only when the base URL is
api.anthropic.com. Since the helper check made this header optional, a
helper that arrived without one gained a freshly generated UUID on a
custom upstream. Attach it only for a first-party upstream or when the
caller actually sent one, and otherwise keep the header absent as the
real client would.
advanced-tool-use-2025-11-20 is sent only while tool search or another
advanced tool-use feature is on the wire, or when the caller asks for it;
2.1.258 no longer attaches it to plain tool declarations (measured: 158
inline tools, no beta). A caller-supplied afk-mode-2026-01-31 is forwarded
between fast-mode and extended-cache-ttl and is an insertion boundary for
the advisor beta.
2.1.258 Haiku helpers offer the same full compression set as the main
thread and never send X-Stainless-Async. x-client-request-id is attached
only when the client's base URL is api.anthropic.com, so the helper
transport check accepts an empty value or a valid UUID and rejects only a
malformed one.
- Check `error.message` and `message` for retry advice on upstream `server_error` responses.
- Treat server errors indicating the request can be retried as eligible overload bootstrap failures.
- Normalize complex constant `oneOf` and `anyOf` tool parameter schemas into equivalent enums to prevent upstream aborts.
- Escape property keys containing dots and colons during schema updates to prevent invalid path splitting.
- Identify terminal `response.incomplete` events with zero output tokens and no content as upstream failures.
Closes: #5551
- Configure upstream connection pool under antigravity.connection-pool with enabled: false by default.
- In short connection mode, set MaxIdleConnsPerHost = -1 with DisableKeepAlives = false, ensuring immediate TCP termination after response body completion without leaking Connection: close request headers.
- When pooling is explicitly enabled (enabled: true), cap idle-conn-timeout at 210s (leaving a 30s safety buffer below Google Frontend's 240s Keep-Alive cutoff) and default max-idle-conns-per-host to 2 (bounded at 100).
- Refactor TransportCache to execute CloseIdleConnections outside the mutex lock during LRU eviction and matching closes.
- Proactively evict and close idle connections on 429 quota exhaustion across Execute, ExecuteStream, and CountTokens.
- Wire hot-reload diff detection and server reload purge hooks for graceful transport pool updates.
- Detect `collaboration-optimize.` tool conflicts during multi-agent v2 request optimization.
- Restore dot-prefixed tool calls back to the collaboration namespace and base tool name in responses.
Closes: #5524
- Ensure Gemini, AI Studio, Vertex, and Antigravity requests end with a user turn to prevent upstream errors when trailing with model content.
- Preserve trailing turns that contain a function response and skip trailing adjustments for `countTokens` actions or Claude models.
- Introduce boundary normalization helpers to validate both leading and trailing turn roles.
Closes: #5358
- Fall back to request payload input items or `previous_response_id` during websocket compaction when transcript snapshot is empty.
- Preserve `previous_response_id` in prepared request body for xAI execution.
Closes: #5205
- Add payload detection and preparation helpers for compaction triggers and summary generation.
- Implement AES-GCM sealing and unsealing to store and restore opaque compaction capsules across turns.
- Support expanding compaction capsules into developer context for downstream turns.
- Provide summary text extraction across Gemini, Claude, and OpenAI response formats.
- Add builders for compaction non-stream JSON responses and SSE stream frames.
Closes: #5094
- Reuse coresession.ExtractSessionInfo across HTTP headers and request payloads to unify canonical session prefix namespaces with the scheduler.
- Extract hierarchical session identities in two phases: initial extraction from request headers on entry, and authoritative deep extraction once request payloads and metadata are available.
- Support Claude Code multi-level subagents (X-Claude-Code-Agent-Id, metadata.agent_id) and Codex thread fork lineages.
- Propagate SessionID and ParentSessionID across ClientRequestMetadata, UsageReporter, and coreusage.Record without root_session_id.
- Include session_id and parent_session_id in queuedUsageDetail for Home LPushUsage forwarding and Redis consumption with self-loop guards.
- Add comprehensive test coverage for canonical headers, body extraction, ghost parent elimination, and self-referential loop guards.
- Mark Codex usage limit errors as credential-scoped across HTTP and WebSocket executors.
- Support both top-level and nested error structures with case-insensitive matching when parsing retry-after resets.
- Propagate prevalidated candidate context to session affinity and built-in selectors during auth selection.
Closes: #5529
- Detect advisor tool calls and advisor results in conversation history.
- Preserve forwarded system prompt blocks in top-level system instead of splicing them into message turns or prepending reminders.
- Prevent message index shifts that break layout bindings for encrypted advisor results and trigger upstream 400 errors.
Closes: #5470
- Require exactly one non-reminder text block matching quota/test/probe/./Hi for probe request matching.
- Identify title helper requests via expanded session title patterns.
- Implement post-payload bidirectional probe reclassification: strip CPA diagnostics and billing tags on probes, while restoring continuity if declassified.
- Preserve caller-owned and payload-supplied diagnostics on probe requests.
- Gate late sensitive-word obfuscation strictly on cloaked requests in Execute and ExecuteStream.
- Add comprehensive end-to-end tests for probe classification, diagnostics isolation, and late cloaking.
- Inject '# Reporting outcomes' system block for Fable 5.1 / Mythos 5.1 models matching Claude Code 2.1.258 tengu_dapper_lagoon behavior.
- Support both string-format and block-array system prompts for Fable reporting injection.
- Add reconcileClaudeCodeFableModelAfterPayload to reconcile model additions when payload rules override model or thinking.
- Reliably re-add reporting block if omitted after payload rules modify system blocks.
- Drop injected thinking.display on Fable models when payload changes thinking type from adaptive to disabled.
- Enhance payload rule path tracking with ancestor target matching in ApplyPayloadConfigWithTrackedPaths.
