043-planning-github-issues
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows security best practices by implementing an interactive gate that verifies the presence of the GitHub CLI before proceeding, ensuring the agent does not operate on missing or fabricated data.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The instructions explicitly forbid printing tokens or secrets and direct the user to handle authentication locally, preventing accidental exposure of credentials in the chat context.
- [INDIRECT_PROMPT_INJECTION]: The skill implements a defense-in-depth strategy by prohibiting the ingestion of GitHub issue prose (titles, bodies, comments, summaries). By restricting the agent to issue numbers for traceability and repository-owned planning artifacts for requirements, it effectively removes the attack surface for indirect prompt injection from untrusted external data.
- [EXTERNAL_DOWNLOADS]: All external references target the official GitHub CLI domain (cli.github.com), which is a well-known and trusted service. Recommendations for installation use standard, safe package managers like Homebrew and winget.
Audit Metadata