ai-avatar-video

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md examples explicitly run the belt CLI against arbitrary public image/audio/video URLs (e.g., "image": "https://portrait.jpg", "audio_url"/"video_url": "https://video.mp4") and include a transcription/dubbing workflow that fetches and processes those external files, so untrusted third‑party content is ingested and can influence subsequent processing and actions.