The Agent Skills Directory

[SAFE]: No security threats detected. The skill follows its stated purpose of issue triage and PR management.\n- [COMMAND_EXECUTION]: Uses legitimate GitHub CLI (gh) commands (issue view, pr view, pr list, issue edit, issue comment) to interact with repository metadata. The commands target a specific repository (PrefectHQ/fastmcp) associated with the author's ecosystem.\n- [DATA_EXPOSURE]: Accesses public repository data using standard GitHub API calls. No sensitive local files, environment variables, or hardcoded credentials are involved.\n- [PROMPT_INJECTION]: The skill processes untrusted input from external contributors via issue and PR descriptions. It mitigates potential injection risks by instructing the agent to treat external comments as leads rather than conclusions and by mandating the use of --body-file for comments to avoid shell-level variable or command interpolation.

review-issue