video-edit

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

• Third-party content exposure detected (high risk: 0.90). The skill's youtube_clip.py explicitly downloads and transcribes arbitrary YouTube URLs via yt-dlp and faster-whisper (and reads YouTube chapters), and simple_video_edit.py transcribes videos and sends the transcript to an LLM (Claude) to generate metadata/chapters and drive uploads—i.e., it ingests untrusted, user-generated content from public websites (YouTube) and acts on that content to make tool-use decisions.

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

• Potentially malicious external URL detected (high risk: 1.00). The jump_cut_vad_singlepass.py script calls torch.hub.load(repo_or_dir='snakers4/silero-vad'), which at runtime fetches and executes code from the GitHub repository https://github.com/snakers4/silero-vad and is a required dependency for the Silero VAD processing.