comprehensive-code-review

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow is code-reviewing user-specified code (e.g., “Review this pull request / path/to/file”), so the LLM ingests outsider-authored free text from PR/issue bodies or repository files that the operating user did not author, via the code-reading step in Phase 1.