- 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
Replace Codex-specific response model APIs with shared model observation and remove redundant substitution warning throttling. Consolidate repeated Antigravity context-hash rejection messages into a single request-level log while preserving rejection details across index rebuilds.
- Add `PayloadFinalizer` to apply configuration rules at the final serialization barrier
- Support decoding, modifying, and re-encoding Devin protobuf business fields
- Add multipart form-data payload normalization to JSON for media payload rules
Closes: #5067
- Add ASCII pre-filter to skip expensive gjson traversals for title helper and 1h TTL checks
- Check raw bytes for target keywords before initiating full JSON walks
- Fall back to full inspection when printable ASCII Unicode escapes are detected
Closes: #6291
- Pin local date to continuity context on the first request of a session
- Reuse pinned session date during cloaked date injection across execute and stream flows
- Prevent midnight calendar date flips from invalidating conversation prompt cache prefixes
Closes: #6282
- Fall back to top-level token usage when Codex response usage contains only service tier metadata
- Propagate parsed service tier while retaining top-level token counts
- Avoid early return during stream usage observation for chunks with zero token counts
Closes: #6261
- Handle apply_patch custom tool call streaming deltas and snapshots for Responses API translation
- Track and resolve step indexes, function call identities, and error states across stream events
- Propagate apply_patch tool input errors and upstream failure states in stream and non-stream responses
- Pass target executor identity during Codex token counting and xAI request translation
- Skip tool integer normalization when target executor is Codex during compatibility translation
- Preserve number types in tool parameter schemas during token counting requests
- Skip normalizing tool parameter number types to integer when the target executor is Codex or Codex WebSocket
- Remove tool integer type normalization from schema cleanup in native Codex execution paths to preserve reserved tool schemas
- Apply tool integer type normalization only for non-Codex upstream targets with Codex client User-Agents
Closes: #6244
- Select HTTP/2 or HTTP/1.1 transport dynamically based on TLS ALPN negotiation
- Route empty ALPN and HTTP/1.1 requests through a dedicated non-pooled transport reusing the handshaken connection
- Ensure uTLS connections are cleanly closed upon response completion or request failure
Closes: #6223
- Defer intervening non-tool messages until all batch tool outputs are emitted to keep outputs contiguous
- Apply responses input normalization across standard and streaming executor execution paths
Closes: #6231
- Detect Codex client requests via User-Agent and convert target numeric parameters to integer types
- Support schemas across OpenAI functions, Claude input schema, Gemini declarations, and namespace tools
- Apply normalization during executor payload processing, multi-agent request translation, and WebSocket streaming
Closes: #6237
Use Home credential options as authoritative model capability
configuration for compat handling and prompt cache behavior.
Expose quota provider operations through the public plugin host API.
- Add `SanitizeMetaWebSearchTools` to remove `search_content_types` from `web_search` tool definitions
- Support sanitizing nested tools within `namespace` tool definitions
- Apply web search tool sanitization during Meta request preparation
Closes: #6156
- Use `ExtractSummaryConfig` to parse summary configuration from OpenAI response payloads
- Default to enabled auto summaries when reasoning effort is set without explicit summary visibility
- Apply the resolved summary configuration directly to Antigravity translations
Closes: #6143
- Implement `splitDevinUID` for consistent UID splitting and effort extraction.
- Update model aggregation to handle thinking levels and merge variants.
- Refine display name cleaning logic to remove redundant suffixes.
- Adjust Devin catalog integration to exclude exposed thinking variants.
- Normalize and sort thinking levels with an extended rank scheme.
Closes: #6141
- Attach resolved execution model info during execution and token counting even when restoring execution models
- Propagate configuration update support to resolved Home dispatch model info
- Track modifications to configuration updates across plugin normalizers in the translation pipeline
- Mark configuration update support for eligible Codex models in the registry
- Remove direct Messages passthrough bypass for unconfirmed OAuth credentials
- Ensure standard cache control handling and CLI fingerprinting apply upstream
- Support detection for Claude Code 2.1.280 title helper payload shape
Closes: #6120
* fix(claude): align Opus 5.5 cloak and preserve native CLI hints
* fix(claude): recognize the 2.1.280 Haiku title helper
The native title request now appends server-side fallback, fallback credit,
and cache diagnosis after the structured-output beta. Keep the helper
allowlist exact so this sequence passes through without cloaking.
* fix(claude): align thinking visibility and Messages passthrough
* fix(claude): preserve thinking visibility through compatibility paths
* fix(claude): harden OAuth beta, session alignment, and 2.1.280 helper headers
- Add `RequestID` (unique execution UUID) and `TraceID` (inbound request ID) fields to usage records and plugin API
- Generate unique UUIDs per model attempt in `UsageReporter` while linking to parent trace context
- Expose `execution_id` and `trace_id` in Redis queue usage payloads while preserving legacy `request_id`
- Update default Claude CLI version to 2.1.280 in cloaking and device profile
- Add `mid-conversation-tool-changes-2026-07-01` beta header support for supported models
Closes: #6054
Skip the second compatibility translation for Claude, Gemini, and Gemini
interactions when the baseline and working payloads share a backing array
and no plugin hooks are installed. Keep independent buffers, the existing
hook order, and native interactions copies.
- Add `use-max-completion-tokens` setting to OpenAI compatibility model configuration
- Provide token normalization between `max_tokens` and `max_completion_tokens` based on model preference
Closes: #5939
- Set expected upstream model for Devin in both streaming and non-streaming paths to avoid false positive substitution warnings on intentional mappings
- Account for line overhead and enforce max lines per stream event in StreamResponseModelObserver, dropping overflowed events until the event boundary
- Add unit tests for Devin intentional mappings and stream observer bounded memory behavior
- observe response model from terminal sourceEvent in Meta non-stream multi-event SSE responses
- record expected upstream model in UsageReporter to prevent false substitution warnings on Kimi canonical mappings
- use bounded stream observer to extract response models across image stream chunk boundaries
- add regression tests for Meta SSE non-stream, Kimi model mappings, and chunked image streams
- Devin: extract authentic upstream model name from parsed Usage.ModelName
instead of synthesized interactions JSON or binary Connect frames. Keep
response model empty when upstream does not report one.
- Gemini: support interaction.model and event_type terminal semantics
in response model extractors for Gemini Interactions streaming.
- Add unit tests for Devin and Gemini Interactions response model observability.
- Add `AlignOpenAIToolCallMessages` to reorder tool results immediately after the matching assistant tool calls while preserving original content and numeric precision.
- Prevent deferred message reordering and call ID guessing when tool outputs are incomplete, duplicate, or missing IDs.
- Normalize translated requests after applying summary configuration in Codex multi-agent execution.
Closes: #5925
Codex upstreams can silently serve a different model than the one requested
(HTTP 200, with response.model naming the substitute). The proxy kept no record
of it: nothing logged, nothing reported, only the pass-through response body.
- Add Record.ResponseModel to sdk/cliproxy/usage, aligned with the existing
ResponseServiceTier field, and emit it from the redis usage queue as the
optional response_model payload field alongside response_service_tier. Only
the record for the requested model carries it: additional-model records
(image generation tool usage) describe a side model the upstream response
never refers to, and would otherwise look like a substitution downstream.
- Add internal/runtime/executor/helps/response_model.go with
extractCodexResponseModelEvent (SSE frames and raw JSON, restricted to the events
that embed the authoritative response object, rejecting non-string and
oversized upstream model names) and IsCodexModelSubstituted (both sides
trimmed, lower-cased and stripped of thinking suffixes, dated aliases such as
gpt-5.6-terra-2026-05-13 accepted in either direction).
- UsageReporter records the served model on the event path and emits the WARN
when the attempt publishes its usage record, so no logging work happens
before the first event is forwarded. Repeats are throttled per
(auth id, requested model, served model) with a 10 minute window, because on
an affected credential every request is substituted and an unthrottled
warning would mirror the whole request volume into the logs. The credential
is labelled auth_index=<index> only: codex credential file names embed the
account e-mail, which must not be written to the logs at request rate.
Coverage, by entry point. The served model is observed on the HTTP streaming
path (both the bootstrap-buffered handshake and the streaming goroutine), the
HTTP non-streaming Execute loop, the websocket streaming and non-streaming
paths, and the two /responses-shaped image entry points. The remaining codex
entry points cannot report it and are therefore left alone: executeCompact
(/responses/compact answers with a compaction object that has no event type and
no response.model), the two direct image endpoints (/images/generations and
/images/edits answer in the Images API shape and stream image_generation.*
events), and CountTokens (counts locally with tiktoken, never reaching an
upstream).
TokenAccountingSchemaVersion is not bumped: it versions the token accounting
contract (token breakdown semantics), and this change only adds an optional
non-token field that leaves existing consumers and all token math untouched.
Note: response_model ships with the usage record and is the counting source;
the WARN is a throttled alerting signal and must not be used to count
substitutions.
Tests: table-driven unit tests for both helpers over real model ids, reporter
tests covering the published record, the single throttled warning, the absence
of account identifiers in it, concurrent observation and publishing under
-race, the throttle window and its entry bound, an executor-level guard for the
observeCodexTokenEvent wiring and the per-model records, plus a redisqueue
payload assertion for response_model. gofmt, go vet, go test -race on the
touched packages and go test ./... are clean.
- Skip parsing the `Retry-After` header when a rate limit rejection is overage-only to prevent global credential cooldowns.
- Allow exponential backoff to handle model recovery while keeping shared subscription windows available for other models.
Closes: #5920
- Add `claude.model-level-cooling` configuration to scope rate limit cooldowns to the requested model.
- Treat overage-only and spend cap rejections as model-scoped when shared subscription windows remain healthy.
- Propagate model-level cooling settings into streaming, token counting, and direct execution error classifiers.
Closes: #5915
- Match tool results against pending tool calls and downgrade unmatched results to user messages.
- Prevent downgraded orphaned tool results from consuming images intended for user turns.
- Unwrap protocol wrapper envelopes and extract structured text parts while preserving arbitrary business JSON.
- Provide a placeholder for empty or whitespace-only tool results.
Closes: #5911
- Track and aggregate tool calls by call ID instead of slot index in streaming and buffered execution.
- Support raw arguments from invalid JSON fields for custom tool calls.
- Parse usage field 4 as cache write tokens instead of adding to prompt tokens.
- Align client metadata with the default client name and drop deprecated tag 28.
Closes: #5910
