Fixes abound!
- https://github.com/openai/codex/pull/2200 send prompt cache key for better caching (should improve token efficiency!)
- https://github.com/openai/codex/pull/2202 split multiline commands to improve readability
- https://github.com/openai/codex/pull/2095 Parse exec commands and format them more nicely in the UI
- https://github.com/openai/codex/pull/2186 recognize
applypatch
command asapply_patch
command string - https://github.com/openai/codex/pull/2183 Include output truncation message in tool call results to tell the model when it did not get the full output from a command that it might have expected
- https://github.com/openai/codex/pull/2162 show feedback message after
/compact
command, fixing an issue when no deltas were sent - https://github.com/openai/codex/pull/2050 show diff preview for
apply_patch
- https://github.com/openai/codex/pull/1975 add JSON schema sanitization for MCP tools to ensure compatibility with internal JsonSchema enum external contributor: Thanks @yaroslavyaroslav!