supadata

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly fetches and ingests public user-generated content from open web sources (e.g., YouTube/TikTok/Instagram/X/Facebook URLs via /youtube/transcript, /youtube/search, /extract, and /web/scrape in SKILL.md) and the Content Pipeline ("Feed to agent → classify → store in laniameda-kb") shows that those transcripts/extracted page contents are read and used to drive agent decisions, so untrusted third‑party content can materially influence tool use and behavior.