331 Commits

• feat(qwen): add rate limiting and quota error handling

  - Add 60 requests/minute rate limiting per credential using sliding window
  - Detect insufficient_quota errors and set cooldown until next day (Beijing time)
  - Map quota errors (HTTP 403/429) to 429 with retryAfter for conductor integration
  - Cache Beijing timezone at package level to avoid repeated syscalls
  - Add redactAuthID function to protect credentials in logs
  - Extract wrapQwenError helper to consolidate error handling

  piexian · 2026-02-23 00:38:46 +08:00
  3b421c8181
• feat: add cache-user-id toggle for Claude cloaking

  Default to generating a fresh random user_id per request instead of
  reusing cached IDs. Add cache-user-id config option to opt in to the
  previous caching behavior.
  
  - Add CacheUserID field to CloakConfig
  - Extract user_id cache logic to dedicated file
  - Generate fresh user_id by default, cache only when enabled
  - Add tests for both paths

  matchch · 2026-02-21 12:31:20 +08:00
  2fdf5d2793
• fix: ensure connection-scoped headers are filtered in upstream requests

  - Added `connectionScopedHeaders` utility to respect "Connection" header directives.
  - Updated `FilterUpstreamHeaders` to remove connection-scoped headers dynamically.
  - Refactored and tested upstream header filtering with additional validations.
  - Adjusted upstream header handling during retries to replace headers safely.

  Luis Pater · 2026-02-19 13:19:10 +08:00
  2789396435
• Merge PR #1626 into codex/pr-1626
  Luis Pater · 2026-02-19 04:49:14 +08:00
  61da7bd981
• Merge pull request #1628 from thebtf/fix/masquerading-headers

  fix: update Claude masquerading headers and configurable defaults

  Luis Pater · 2026-02-19 04:19:59 +08:00
  fff866424e
• Merge pull request #1625 from thebtf/feat/tool-prefix-config

  feat: add per-auth tool_prefix_disabled option

  Luis Pater · 2026-02-19 04:07:22 +08:00
  252f7e0751
• Merge branch 'pr-1624' into dev

  # Conflicts:
  #	internal/runtime/executor/claude_executor.go
  #	internal/runtime/executor/claude_executor_test.go

  Luis Pater · 2026-02-19 04:05:04 +08:00
  b2b17528cb
• Merge pull request #1608 from thebtf/fix/tool-reference-proxy-prefix-mainline

  fix: add proxy_ prefix handling for tool_reference content blocks

  Luis Pater · 2026-02-19 03:50:34 +08:00
  76294f0c59
• chore(executor): remove unused Openai-Beta header from Codex executor
  Luis Pater · 2026-02-19 02:19:48 +08:00
  e5b5dc870f
• feat(logging, executor): add request logging tests and WebSocket-based Codex executor

  - Introduced unit tests for request logging middleware to enhance coverage.
  - Added WebSocket-based Codex executor to support Responses API upgrade.
  - Updated middleware logic to selectively capture request bodies for memory efficiency.
  - Enhanced Codex configuration handling with new WebSocket attributes.

  Luis Pater · 2026-02-19 01:57:02 +08:00
  bb86a0c0c4
• fix: update Claude masquerading headers and make them configurable

  Update hardcoded X-Stainless-* and User-Agent defaults to match
  Claude Code 2.1.44 / @anthropic-ai/sdk 0.74.0 (verified via
  diagnostic proxy capture 2026-02-17).
  
  Changes:
  - X-Stainless-Os/Arch: dynamic via runtime.GOOS/GOARCH
  - X-Stainless-Package-Version: 0.55.1 → 0.74.0
  - X-Stainless-Timeout: 60 → 600
  - User-Agent: claude-cli/1.0.83 (external, cli) → claude-cli/2.1.44 (external, sdk-cli)
  
  Add claude-header-defaults config section so values can be updated
  without recompilation when Claude Code releases new versions.

  Kirill Turanskiy · 2026-02-18 03:38:51 +03:00
  73dc0b10b8
• feat: passthrough upstream response headers to clients

  CPA previously stripped ALL response headers from upstream AI provider
  APIs, preventing clients from seeing rate-limit info, request IDs,
  server-timing and other useful headers.
  
  Changes:
  - Add Headers field to Response and StreamResult structs
  - Add FilterUpstreamHeaders helper (hop-by-hop + security denylist)
  - Add WriteUpstreamHeaders helper (respects CPA-set headers)
  - ExecuteWithAuthManager/ExecuteCountWithAuthManager now return headers
  - ExecuteStreamWithAuthManager returns headers from initial connection
  - All 11 provider executors populate Response.Headers
  - All handler call sites write filtered upstream headers before response
  
  Filtered headers (not forwarded):
  - RFC 7230 hop-by-hop: Connection, Transfer-Encoding, Keep-Alive, etc.
  - Security: Set-Cookie
  - CPA-managed: Content-Length, Content-Encoding

  Kirill Turanskiy · 2026-02-18 00:16:22 +03:00
  1f8f198c45
• feat: add per-auth tool_prefix_disabled option

  Allow disabling the proxy_ tool name prefix on a per-account basis.
  Users who route their own Anthropic account through CPA can set
  "tool_prefix_disabled": true in their OAuth auth JSON to send tool
  names unchanged to Anthropic.
  
  Default behavior is fully preserved — prefix is applied unless
  explicitly disabled.
  
  Changes:
  - Add ToolPrefixDisabled() accessor to Auth (reads metadata key
    "tool_prefix_disabled" or "tool-prefix-disabled")
  - Gate all 6 prefix apply/strip points with the new flag
  - Add unit tests for the accessor

  Kirill Turanskiy · 2026-02-17 21:48:19 +03:00
  9261b0c20b
• fix: skip proxy_ prefix for built-in tools in message history

  The proxy_ prefix logic correctly skips built-in tools (those with a
  non-empty "type" field) in tools[] definitions but does not skip them
  in messages[].content[] tool_use blocks or tool_choice. This causes
  web_search in conversation history to become proxy_web_search, which
  Anthropic does not recognize.
  
  Fix: collect built-in tool names from tools[] into a set and also
  maintain a hardcoded fallback set (web_search, code_execution,
  text_editor, computer) for cases where the built-in tool appears in
  history but not in the current request's tools[] array. Skip prefixing
  in messages and tool_choice when name matches a built-in.

  Kirill Turanskiy · 2026-02-17 21:42:32 +03:00
  7cc725496e
• fix: skip built-in tools in tool_reference prefix + refactor to switch

  - Collect built-in tool names (those with a "type" field like
    web_search, code_execution) and skip prefixing tool_reference
    blocks that reference them, preventing name mismatch.
  - Refactor if-else if chains to switch statements in all three
    prefix functions for idiomatic Go style.

  Kirill Turanskiy · 2026-02-16 19:37:11 +03:00
  24c18614f0
• fix: handle tool_reference nested inside tool_result.content[]

  tool_reference blocks can appear nested inside tool_result.content[]
  arrays, not just at the top level of messages[].content[]. The prefix
  logic now iterates into tool_result blocks with array content to find
  and prefix/strip nested tool_reference.tool_name fields.

  Kirill Turanskiy · 2026-02-16 19:06:24 +03:00
  603f06a762
• fix: add proxy_ prefix handling for tool_reference content blocks (#1)

  applyClaudeToolPrefix, stripClaudeToolPrefixFromResponse, and
  stripClaudeToolPrefixFromStreamLine now handle "tool_reference" blocks
  (field "tool_name") in addition to "tool_use" blocks (field "name").
  
  Without this fix, tool_reference blocks in conversation history retain
  their original unprefixed names while tool definitions carry the proxy_
  prefix, causing Anthropic API 400 errors: "Tool reference 'X' not found
  in available tools."
  
  Co-authored-by: Kirill Turanskiy <kt@novamedia.ru>

  Kirill Turanskiy · 2026-02-16 19:06:24 +03:00
  98f0a3e3bd
• chore(runtime): update Qwen executor user agent and headers for compatibility with new runtime standards
  Luis Pater · 2026-02-16 23:29:47 +08:00
  453aaf8774
• refactor(all): replace manual pointer assignments with new to enhance code readability and maintainability
  Luis Pater · 2026-02-15 14:10:10 +08:00
  46a6782065
• chore(runtime, registry): update Codex client version and GPT-5.3 model creation date
  Luis Pater · 2026-02-13 12:47:48 +08:00
  ae1e8a5191
• fix(schema): remove Gemini-incompatible tool metadata fields

  Sanitize tool schemas by stripping prefill, enumTitles, $id, and patternProperties to prevent Gemini INVALID_ARGUMENT 400 errors, and add unit and executor-level tests to lock in the behavior.
  
  Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

  Nathan · 2026-02-11 18:29:17 +11:00
  166d2d24d9
• feat(executor): add session ID and HMAC-SHA256 signature generation for iFlow API requests
  hkfires · 2026-02-09 19:29:42 +08:00
  49c1740b47
• Merge pull request #1467 from dusty-du/fix/kimi-toolcall-reasoning-content

  Fix Kimi tool-call payload normalization for reasoning_content

  Luis Pater · 2026-02-07 09:35:04 +08:00
  7e9d0db6aa
• fix(kimi): reduce redundant payload cloning and simplify translation calls
  Luis Pater · 2026-02-07 08:51:48 +08:00
  78ef04fcf1
• fix(kimi): update base URL and integrate ClaudeExecutor fallback

  - Updated `KimiAPIBaseURL` to remove versioning from the root path.
  - Integrated `ClaudeExecutor` fallback in `KimiExecutor` methods for compatibility with Claude requests.
  - Simplified token counting by delegating to `ClaudeExecutor`.

  Luis Pater · 2026-02-07 06:42:08 +08:00
  f7d0019df7
• Fix Kimi tool-call reasoning_content normalization
  test · 2026-02-06 14:46:16 -05:00
  52364af5bf
• feat: add Kimi authentication support and streamline device ID handling

  - Introduced `RequestKimiToken` API for Kimi authentication flow.
  - Integrated device ID management throughout Kimi-related components.
  - Enhanced header management for Kimi API requests with device ID context.

  Luis Pater · 2026-02-06 20:43:30 +08:00
  68cb81a258
• Add Kimi (Moonshot AI) provider support

  - OAuth2 device authorization grant flow (RFC 8628) for authentication
  - Streaming and non-streaming chat completions via OpenAI-compatible API
  - Models: kimi-k2, kimi-k2-thinking, kimi-k2.5
  - CLI `--kimi-login` command for device flow auth
  - Token management with automatic refresh
  - Thinking/reasoning effort support for thinking-enabled models
  
  Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

  test · 2026-02-05 19:24:46 -05:00
  f5f26f0cbe
• refactor(executor): centralize Codex client version and user agent constants

  - Introduced `codexClientVersion` and `codexUserAgent` constants for better maintainability.
  - Updated `EnsureHeader` calls to use the new constants.

  Luis Pater · 2026-02-06 05:30:28 +08:00
  b4e034be1c
• refactor(translator, executor): remove redundant bytes.Clone calls for improved performance

  - Replaced all instances of `bytes.Clone` with direct references to enhance efficiency.
  - Simplified payload handling across executors and translators by eliminating unnecessary data duplication.

  Luis Pater · 2026-02-06 03:26:29 +08:00
  a5a25dec57
• refactor(executor): optimize payload cloning and streamline SDK translator usage

  - Replaced unnecessary `bytes.Clone` calls for `opts.OriginalRequest` throughout executors.
  - Introduced intermediate variable `originalPayloadSource` to simplify payload processing.
  - Ensured better clarity and structure in request translation logic.

  Luis Pater · 2026-02-06 01:44:20 +08:00
  09ecfbcaed
• refactor(util, executor): optimize payload handling and schema processing

  - Replaced repetitive string operations with a centralized `escapeGJSONPathKey` function.
  - Streamlined handling of JSON schema cleaning for Gemini and Antigravity requests.
  - Improved payload management by transitioning from byte slices to strings for processing.
  - Removed unnecessary cloning of byte slices in several places.

  Luis Pater · 2026-02-05 19:00:30 +08:00
  25c6b479c7
• fix(executor): handle "global" location in AI platform URL generation
  Luis Pater · 2026-02-03 01:39:57 +08:00
  250f212fa3
• refactor(codex): remove codex instructions injection support
  hkfires · 2026-02-01 14:33:31 +08:00
  ac802a4646
• feat(config): add payload filter rules to remove JSON paths

  Introduce `Filter` rules in the payload configuration to remove specified JSON paths from the payload. Update related helper functions and add examples to `config.example.yaml`.

  Luis Pater · 2026-02-01 05:29:41 +08:00
  6d8609e457
• Fixed: #1372 #1366

  fix(caching): ensure unique cache_control injection using count validation

  Luis Pater · 2026-01-31 23:48:50 +08:00
  d216adeffc
• Merge pull request #1248 from shekohex/feat/responses-compact

  feat(openai): add responses/compact support

  Luis Pater · 2026-01-31 03:12:55 +08:00
  f887f9985d
• fix(caching): ensure prompt-caching beta is always appended and add multi-turn cache control tests
  Luis Pater · 2026-01-31 01:42:58 +08:00
  7ff3936efe
• feat(caching): implement Claude prompt caching with multi-turn support

  - Add ensureCacheControl() to auto-inject cache breakpoints
  - Cache tools (last tool), system (last element), and messages (2nd-to-last user turn)
  - Add prompt-caching-2024-07-31 beta header
  - Return original payload on sjson error to prevent corruption
  - Include verification test for caching logic
  
  Enables up to 90% cost reduction on cached tokens.
  
  Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

  Martin Schneeweiss · 2026-01-29 22:59:33 +01:00
  3a43ecb19b
• fix(codex): avoid empty prompt_cache_key
  Shady Khalifa · 2026-01-27 19:06:42 +02:00
  04b2290927
• fix(openai): drop stream for responses/compact
  Shady Khalifa · 2026-01-27 18:27:34 +02:00
  53920b0399
• feat(openai): add responses/compact support
  Shady Khalifa · 2026-01-26 16:36:01 +02:00
  95096bc3fc
• feat(auth): add support for request_retry and disable_cooling overrides

  Implement `request_retry` and `disable_cooling` metadata overrides for authentication management. Update retry and cooling logic accordingly across `Manager`, Antigravity executor, and file synthesizer. Add tests to validate new behaviors.

  Luis Pater · 2026-01-26 21:59:08 +08:00
  70897247b2
• refactor(runtime): implement retry logic for Antigravity executor with improved error handling and capacity management
  Luis Pater · 2026-01-26 06:22:46 +08:00
  2af4a8dc12
• feat(executor): add request_id to error logs

  Extract error.message from JSON error responses when summarizing error bodies for debug logs

  hkfires · 2026-01-25 21:31:46 +08:00
  f30ffd5f5e
• Merge pull request #1132 from XYenon/fix/gemini-models-displayname-override

  fix(gemini): preserve displayName and description in models list

  Luis Pater · 2026-01-25 03:40:04 +08:00
  2e6a2b655c
• fix(claude): skip built-in tools in OAuth tool prefix
  Mauricio Allende · 2026-01-23 21:29:39 +00:00
  f16461bfe7
• feat(executor): apply payload rules using requested model
  hkfires · 2026-01-23 16:38:41 +08:00
  ecc850bfb7
• fix(executor): adjust model-specific request payload
  hkfires · 2026-01-22 20:28:08 +08:00
  7ca045d8b9
• refactor(util): reorder gemini schema cleaner helpers
  hkfires · 2026-01-22 18:38:48 +08:00
  abfca6aab2