audit-pr

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes gh (GitHub CLI) and git commands to retrieve pull request metadata, comments, and branch history. These tools are used for read-only operations during the audit, with the exception of posting status comments to the PR.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes content from pull request bodies, comments, and local project documentation which could contain malicious instructions.
  • Ingestion points: Fetches PR data via gh pr view and reads files such as PLAN.md, TASKS.md, and SPEC.md.
  • Boundary markers: The skill relies on specific SHA-bound metadata tags (<!-- review-change:pass ... -->) to verify the integrity and currency of previous review steps before proceeding.
  • Capability inventory: The skill is capable of posting comments via gh pr comment and analyzing repository history via git log.
  • Sanitization: It explicitly mandates writing comment bodies to temporary files (--body-file) before execution to mitigate shell injection risks and ensures verdict formats are strictly followed to prevent instructions from being misinterpreted by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 11:08 PM
Security Audit — agent-trust-hub — audit-pr