The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill provides documentation for executing standard GitHub CLI commands. No malicious or unauthorized commands were found.\n- [DATA_EXFILTRATION] (SAFE): Network access is limited to the official GitHub API. There is no evidence of hardcoded credentials or unauthorized data exfiltration.\n- [PROMPT_INJECTION] (LOW): The skill reads untrusted external data (PR comments and CI logs), creating a surface for indirect prompt injection (Category 8). Evidence: 1. Ingestion: gh api (comments) and gh run view (logs). 2. Boundary markers: Absent. 3. Capability inventory: Subprocess calls for gh commands. 4. Sanitization: Absent. This is a common characteristic of tools interacting with external repository content.

github-cli