review-pr

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [Command Execution]: The skill interacts with the GitHub CLI (gh) and executes a local Python script (review-pr.py) to manage the review process. The Python script follows security best practices by using subprocess.run with a list of arguments and shell=False, which prevents potential command injection vulnerabilities.
  • [Indirect Prompt Injection]: The skill ingests and analyzes external code changes from pull requests. Content within a PR diff could potentially contain instructions designed to influence the agent's analysis. This risk is effectively addressed by the skill's design, which generates a YAML file for user review and manual editing before any comments are submitted to GitHub.
  • [Data Handling]: Generated review comments are temporarily stored in the system's local temporary directory (e.g., AppData\Local\Temp on Windows). This is standard behavior for temporary task data, and the skill provides the file path to the user for oversight.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 10:40 AM
Security Audit — agent-trust-hub — review-pr