trending-content-scout

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

• Third-party content exposure detected (high risk: 1.00). This skill explicitly uses web_search and web_fetch fallback to scan public social platforms (YouTube, TikTok, X, Reddit) and ingests titles, posts, view/like/comment counts from user-generated pages to drive engagement scoring, pattern analysis, gap finding, and downstream actions, so untrusted third-party content can influence agent behavior.