anysite-trend-analysis

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md "Quick Start" and "Common Workflows" explicitly instruct the agent to call functions like search_twitter_posts, search_reddit_posts, search_youtube_videos, search_instagram_posts, and related get_* endpoints to fetch and analyze user-generated content from public platforms (Twitter/X, Reddit, YouTube, LinkedIn, Instagram), which the agent reads and uses to drive decisions and actions, creating clear exposure to untrusted third-party content that could carry indirect prompt injection.