last30days-v3-spec

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

• Third-party content exposure detected (high risk: 1.00). This skill explicitly retrieves public, user-generated content (e.g., Reddit/TikTok/Instagram via ScrapeCreators, X via xAI/Bird, web via Brave/Serper, YouTube via yt-dlp) as part of its required runtime and synthesis workflow (see "Runtime expectations" and "Security and permissions"), and that third-party content is directly read and used to drive planning, ranking, and recommendations, creating a clear risk of indirect prompt injection.