The Agent Skills Directory

COMMAND_EXECUTION (SAFE): The skill mentions 'pnpm run dev' for local testing of documentation. This is a standard development command and is not used in a malicious context.
DATA_EXPOSURE (SAFE): The skill references local file paths for documentation ('site/src/content/docs/') and source code ('src/') to provide context for the review. No sensitive paths or exfiltration methods were found.
PROMPT_INJECTION (LOW): As the skill processes external documentation files, it is theoretically susceptible to indirect prompt injection if those files contain malicious instructions. However, the risk is minimal given the specialized nature of the task and lack of sensitive capabilities.

review-docs