pr-review

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and reads PR diffs, files, reviews, and issue comments from GitHub using commands like gh pr diff <number>, gh pr view, and gh api "repos/$REPO/.../issues/<number>/comments", and those user-generated PR descriptions, code, and comments are required inputs that the agent must interpret and act on when composing reviews, so untrusted third-party content can influence tool use and decisions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill issues GitHub API calls at runtime (e.g., gh api "repos/$REPO/pulls//reviews" and gh pr diff , i.e. https://api.github.com/repos/:owner/:repo/pulls//reviews) to fetch PR diffs and file contents that are injected into the agent's context and directly drive its review output, so the external content controls prompts at runtime.