The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill instructions focus on read-only review and analysis.
[COMMAND_EXECUTION]: The skill utilizes standard git diff and git diff --cached commands to determine the scope of the review. These commands are used solely for reading repository state and do not involve untrusted input or destructive actions.
[PROMPT_INJECTION]: The skill processes external data from local source code and documentation which constitutes a surface for indirect prompt injection. This risk is mitigated by explicit instructions that forbid sub-agents from performing any state-mutating actions such as editing files or applying patches.

review-swarm