The Agent Skills Directory

[SAFE]: The skill explicitly restricts all sub-agents to a read-only role, preventing any modifications to the file system or git state.
[COMMAND_EXECUTION]: The skill uses standard git diff commands to determine the scope of the review. These commands are used for information gathering only and do not pose a security risk.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from git diffs and local documentation (Ingestion points: git diff output, local repository files like AGENTS.md). While no explicit boundary markers or sanitization steps are defined for the input data, the risk is mitigated by the lack of any write or execution capabilities (Capability inventory: git diff and file reading only; no apply_patch, shell execution, or network access).

review-swarm