New Features
- Added a configurable grace period for discovering tools from optional MCP servers. (openai#41199)
- Extensions can now inspect or replace MCP tool results before they reach the model. (openai#41202)
- Plugin catalogs now combine per-repository configuration and report invalid project marketplaces without hiding valid plugins. (openai#41208)
Bug Fixes
- Preserved restored permission profiles across TUI turns and prevented
/cdfrom weakening sandbox restrictions. (openai#41192) - Kept tool availability and reasoning effort correct when switching models or falling back to another model. (openai#41195, openai#41206)
- Improved remote sandbox enforcement using the executor’s actual home directory, operating system, and path conventions. (openai#41196, openai#41204, openai#41207, openai#41209)
- Preserved structured MCP tool and resource errors in app-server responses. (openai#41196)
- Counted nested subagent token usage toward root goal budgets. (openai#41183)
- Prevented stale Guardian classifications from authorizing actions after permission state changes. (openai#41196)
Chores
- Added telemetry for escalated stdin reviews and remote executor MCP discovery. (openai#41189, openai#41205)
- Stabilized Guardian WebSocket and core fixture tests under slow or highly concurrent CI. (openai#41191, openai#41194)
Changelog
Full Changelog: openai/codex@rust-v0.150.0...rust-v0.151.0