address-pr-comments

Pass

Audited by Gen Agent Trust Hub on Mar 25, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection from malicious pull request comments.
  • Ingestion points: PR comment bodies and diff hunks fetched via the GitHub API in SKILL.md (Step 2) and references/gh-api-reference.md.
  • Boundary markers: Absent; instructions do not wrap external content in delimiters or specify ignoring instructions within comments.
  • Capability inventory: The skill can read and write local files and execute gh api commands to perform network operations.
  • Sanitization: No sanitization or validation of the comment body is performed before processing.
  • [COMMAND_EXECUTION]: The skill uses the GitHub CLI to interact with repositories.
  • Evidence: Commands like gh api, gh pr view, and gh auth status are used in SKILL.md.
  • Mitigation: The risk of unintended actions is significantly mitigated by a required user approval step (Step 7) before any replies are posted or code changes are finalized.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 25, 2026, 03:34 AM
Security Audit — agent-trust-hub — address-pr-comments