The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to use the Bash tool to execute standard security auditing commands, such as git diff, gh pr diff, npm audit, pip audit, and semgrep. These commands are necessary for the skill's primary purpose of differential code review and are used in a controlled, local context.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it is designed to analyze untrusted code from external sources (pull requests and git diffs). Maliciously crafted code comments could attempt to influence the agent's behavior.
Ingestion points: The skill ingests data from git diff and gh pr diff commands as specified in SKILL.md (Step 1).
Boundary markers: The instructions do not define specific boundary markers or instructions for the agent to ignore embedded commands within the diff data.
Capability inventory: The skill is granted powerful tools including Bash, Write, Edit, and Glob, which could be targeted by an injection attack.
Sanitization: No sanitization or filtering of the code content is performed prior to analysis.

differential-review