summarize

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow explicitly summarizes arbitrary user-supplied URLs (and can extract YouTube transcripts), so it will fetch outsider-authored web/video text and feed that readable content into the LLM context via the URL/video extraction path (public web content / third-party video transcripts).