fnl-review-gate

Pass

Audited by Gen Agent Trust Hub on Nov 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The scripts/merge_check.py script executes the GitHub CLI tool to perform repository state checks.\n
  • Evidence: Uses subprocess.run(["gh", ...]) to interact with GitHub APIs for permission and status verification.\n
  • Mitigation: Inputs like the PR number and repository name are strictly validated via regular expressions and digit-only checks prior to execution.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to interpret data from pull request comments and code diffs which are untrusted external inputs.\n
  • Ingestion points: Found in scripts/merge_check.py, which reads PR comments and file content through the gh tool.\n
  • Boundary markers: Uses specific hidden HTML markers (<!-- fnl-review ... -->) to identify authorized instructions.\n
  • Capability inventory: The skill is limited to reading metadata and providing a binary merge decision; it cannot write to the filesystem or initiate arbitrary network connections.\n
  • Sanitization: Employs high-fidelity checks to ensure verdict comments are authored by users with write permissions and have not been modified since creation.\n- [EXTERNAL_DOWNLOADS]: The skill references external security auditing extensions.\n
  • Evidence: Links to the official cloudflare/security-audit-skill on GitHub.\n
  • Note: This reference targets a well-known service and is documented as an optional companion for security audits.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 3, 2026, 12:00 PM