github-code-review

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

• Third-party content exposure detected (high risk: 0.75). For PR interactions, the skill fetches PR metadata and especially the PR body text from GitHub’s API (curl ... /pulls/$PR_NUMBER then print(... pr['body'] ...)), which is outsider-authored free text, and that text can be ingested into the agent’s LLM context during review.