voice-lab

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

• Third-party content exposure detected (high risk: 0.80). The skill's runtime workflow clearly ingests and displays public, user-generated media — SKILL.md lists "create-stems → extract vocals from movie/youtube clip" and follow-along flows, app.py accepts and embeds arbitrary YouTube URLs for playback, and content_provider.py issues HTTP requests to a Jellyfin server to fetch/display remote media metadata, so untrusted third‑party content (videos, SRTs, thumbnails) is consumed and used in creating lesson/training packages.