watch-asc-pr

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow explicitly fetches “checks, reviews, top-level comments, and GraphQL review threads” from the target PR, which includes outsider-authored free-form comment/review text from other contributors, and then feeds that content into the agent’s LLM context for re-evaluation.