refactor codex websocket request normalization

feat(codex): default agent mode for non-stream prompts

feat(codex): add planner-reviewer agent mode

refactor(stream): harden forwarding and hook composition

feat(codex): add Redis-backed prompt cache fallback

refactor codex request normalization and stream event parsing

ein

fix(executor): remove unused import and vertex variable

fix(executor): restore retryAfter and count-token payload deletes

fix(executor): restore shared statusErr type

fix(executor): restore usageDetail alias for provider parse usage

refactor(executor): split providers and preserve legacy entrypoints

feat: add cache-user-id toggle for Claude cloaking

fix: ensure connection-scoped headers are filtered in upstream requests

Merge PR #1626 into codex/pr-1626

Merge pull request #1628 from thebtf/fix/masquerading-headers

Merge pull request #1625 from thebtf/feat/tool-prefix-config

Merge branch 'pr-1624' into dev

Merge pull request #1608 from thebtf/fix/tool-reference-proxy-prefix-mainline

chore(executor): remove unused Openai-Beta header from Codex executor

feat(logging, executor): add request logging tests and WebSocket-based Codex executor

fix: update Claude masquerading headers and make them configurable

feat: passthrough upstream response headers to clients

feat: add per-auth tool_prefix_disabled option

fix: skip proxy_ prefix for built-in tools in message history

fix: skip built-in tools in tool_reference prefix + refactor to switch

fix: handle tool_reference nested inside tool_result.content[]

fix: add proxy_ prefix handling for tool_reference content blocks (#1)

chore(runtime): update Qwen executor user agent and headers for compatibility with new runtime standards

refactor(all): replace manual pointer assignments with `new` to enhance code readability and maintainability

chore(runtime, registry): update Codex client version and GPT-5.3 model creation date

fix(schema): remove Gemini-incompatible tool metadata fields

feat(executor): add session ID and HMAC-SHA256 signature generation for iFlow API requests

Merge pull request #1467 from dusty-du/fix/kimi-toolcall-reasoning-content

fix(kimi): reduce redundant payload cloning and simplify translation calls

fix(kimi): update base URL and integrate ClaudeExecutor fallback

Fix Kimi tool-call reasoning_content normalization

feat: add Kimi authentication support and streamline device ID handling

Add Kimi (Moonshot AI) provider support

refactor(executor): centralize Codex client version and user agent constants

refactor(translator, executor): remove redundant `bytes.Clone` calls for improved performance

refactor(executor): optimize payload cloning and streamline SDK translator usage

refactor(util, executor): optimize payload handling and schema processing

fix(executor): handle "global" location in AI platform URL generation

refactor(codex): remove codex instructions injection support

feat(config): add payload filter rules to remove JSON paths

Fixed: #1372 #1366

Merge pull request #1248 from shekohex/feat/responses-compact

fix(caching): ensure prompt-caching beta is always appended and add multi-turn cache control tests

feat(caching): implement Claude prompt caching with multi-turn support