fix-coderabbit-review

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

• Third-party content exposure detected (medium risk: 0.30). In fix-coderabbit-review, Step 1 runs scripts/pr_review.py which fetches outsider-authored PR review/comment bodies and thread text from GitHub REST/GraphQL (/repos/{owner}/{repo}/pulls/{pr_number}/comments, /issues/{pr_number}/comments, and reviewThreads), then reads and processes the generated markdown files (ai-docs/reviews-pr-<PR_NUMBER>/_summary.md and issues/*.md) at runtime.