The Agent Skills Directory

[SAFE]: The skill explicitly enforces a 'read-only' policy, prohibiting any modifications to files, branches, or system state.
[COMMAND_EXECUTION]: The skill uses local CLI tools (git and comet). These are used exclusively for status queries and diff inspections, which are consistent with the skill's stated purpose of performing a code review.
[SAFE]: No external network operations, package installations, or remote code downloads were detected. The interface configuration in agents/openai.yaml is standard and does not grant excessive permissions.

comet-review