review-prs

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data originating from external GitHub pull request contributors.
  • Ingestion points: Step 1 fetches metadata including title and body from open pull requests via gh pr list.
  • Boundary markers: No explicit delimiter markers are used when passing PR content to the audit-the-list sub-skill, though the context is structured by the workflow.
  • Capability inventory: The skill possesses the ability to merge code (gh pr merge, git merge), close pull requests (gh pr close), and run local build scripts (make test).
  • Sanitization: The skill mitigates risks by requiring explicit maintainer adjudication in Step 4. All actions are preceded by an AskUserQuestion call where the maintainer reviews the draft comments and verdicts before execution.
  • [COMMAND_EXECUTION]: The skill uses shell commands to interact with the repository and pull request queue.
  • Evidence: The workflow uses gh pr list, gh pr diff, gh pr close, gh pr merge, git fetch, git merge, and make test.
  • Analysis: These operations are restricted to the author's own repository (vinta/awesome-python) and are consistent with the primary purpose of a pull request management skill. The use of make test is a standard practice for verifying PR quality before merging.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 04:51 PM
Security Audit — agent-trust-hub — review-prs