ingest-youtube

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

• Third-party content exposure detected (high risk: 0.80). This skill explicitly fetches transcripts/summaries from public YouTube URLs by delegating to the learning/youtube-summary pipeline (via yt-dlp) and then reads/uses the produced summary.md and transcript.md as part of its workflow, so untrusted, user-generated third-party content from YouTube could indirectly inject instructions into the agent's outputs.