social-fetch

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

• Third-party content exposure detected (medium risk: 0.30). /social-fetch takes a user-supplied post URL, infers the platform, and then fetches/reads the post’s free-form text from that external source (e.g., Reddit .json and HN Algolia comments, and agent-browser previews/scrapes for other platforms), meaning outsider-authored content is ingested at runtime.