review-pr

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill methodology relies entirely on standard, read-only version control commands (git log, git diff, git show) and the GitHub CLI (gh pr view) to gather context for a review.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data, including pull request descriptions, commit messages, and repository source code, which are potential vectors for indirect prompt injection.
  • Ingestion points: Pull request metadata via gh pr view, commit history via git log, and file contents via git show (SKILL.md).
  • Boundary markers: The instructions do not define specific delimiters for separating untrusted content from the agent's instructions, though they emphasize independent judgment.
  • Capability inventory: The skill is limited to shell execution of standard Git and GitHub CLI commands. It does not contain file-write, arbitrary network requests, or privilege escalation capabilities.
  • Sanitization: No explicit sanitization of command output is performed; the skill relies on the agent's analytical capabilities to distinguish between code content and instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 07:46 PM
Security Audit — agent-trust-hub — review-pr