code-reviewer

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly checks out remote GitHub pull requests and instructs the agent to "Read the PR description and any existing comments" (SKILL.md, Remote PRs → Preparation), which are untrusted, user-generated third-party contents that the agent is expected to interpret and that can materially influence review decisions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill instructs the agent to fetch remote PRs using the GitHub CLI (e.g., https://github.com///pull/<PR_NUMBER> or git@github.com:owner/repo.git via "gh pr checkout") and then run the project's verification suite ("npm run preflight"), so code fetched at runtime from that external URL can be executed on the host.