The Agent Skills Directory

[SAFE]: The skill implements a structured debugging protocol (Phase 1-8) that emphasizes reasoning and planning before code modification. This approach aligns with safe development practices.
[COMMAND_EXECUTION]: The skill instructs the agent to use standard tools like git log, git diff, and grep to gather context from the local repository. These are appropriate for a developer tool and do not present a security risk in this context.
[SAFE]: The skill includes meta-cognitive checks and confidence gates (Phase 5-6), which serve as internal validation steps to ensure the agent's reasoning is sound before suggesting changes.

holistic-analysis