youtube-video-analyst

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md and the included scripts/fetch_transcript.py explicitly auto-fetch arbitrary YouTube video transcripts (public, user-generated content) and require the agent to read and interpret those transcripts as the basis for generating templates and actions, allowing third-party content to directly influence decisions and outputs.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime fetch_transcript.py script explicitly fetches transcripts from YouTube at runtime (e.g., "https://youtube.com/watch?v=VIDEO_ID") and those fetched transcripts are injected into the agent's analysis prompts, meaning external content directly controls the model input.