youtube-watcher-hardened

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches transcripts from public YouTube URLs (SKILL.md examples and the scripts/get_transcript.py that runs yt-dlp to download .vtt captions), then reads and summarizes that user-generated, untrusted content as part of its workflow, so embedded instructions in transcripts could influence the agent's decisions or actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill invokes yt-dlp at runtime to fetch subtitles from YouTube URLs (e.g., https://www.youtube.com/watch?v=VIDEO_ID), and that externally-fetched transcript content is a required dependency that can contain prompt-injection directives which directly control agent behavior.