review-github-pr

Warn

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill identifies and executes validation commands (such as linting or type-checking) specified in the repository's CLAUDE.md file during Phase 1. Since the skill checkouts the Pull Request branch before running these checks, a malicious pull request could modify CLAUDE.md to execute arbitrary shell commands on the host machine.
  • [EXTERNAL_DOWNLOADS]: When invoked with a URL, the skill clones remote repositories from GitHub to the local /tmp directory using gh repo clone. While GitHub is a trusted service, this action downloads unreviewed third-party code into the local environment.
  • [PROMPT_INJECTION]: The skill processes untrusted input from pull request diffs, descriptions, and commit messages. It attempts to mitigate indirect prompt injection by wrapping this data in <pr-content> tags and providing explicit instructions to sub-agents to ignore any embedded commands or instructions within those tags.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 6, 2026, 03:56 PM
Security Audit — agent-trust-hub — review-github-pr