code-review-ai-ai-review

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill embeds full code diffs and static-analysis outputs into model prompts and asks the model to return concrete code examples / modified code in its outputs, which can force the LLM to reproduce any secrets present in those inputs verbatim.

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow for AI review (CI/CD GitHub Actions running python scripts/ai_review.py and building an LLM prompt with diff and static_results) will ingest PR-authored free text/code diff and issue content from outside the operating user (e.g., pull request author), which can contain prompt-injection payloads.