youtube-transcript

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

• Third-party content exposure detected (low risk: 0.10). The skill’s runtime path ingests outsider-authored free text only as a user-provided YouTube URL, then uses yt-dlp to fetch/derive transcript captions from that specific video (optionally converting resulting .vtt text), rather than reading from any attacker-controlled arbitrary feed/queue without selecting a target.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill instructs the agent to install system packages (including explicit "sudo apt update && sudo apt install -y yt-dlp") and to run installation commands that modify the machine state and require elevated privileges, so it should be flagged.