pp-youtube

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly instructs the agent to fetch and parse live YouTube data (e.g., youtube videos-transcript via the timedtext endpoint and youtube videos-comments using commentThreads.list) — public, user-generated content from YouTube — and to read/interpret those transcripts/comments (per the "Verify a candidate via transcript" and "Use to gauge" steps) to decide which videos to embed, so untrusted third-party content can materially influence agent actions.