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 explicitly shows workflows that ingest arbitrary public URLs (e.g., the "Full Workflow: Dub Video in Another Language" and many examples using "image_url", "audio_url", "video_url" like https://video.mp4, https://speech.mp3, https://portrait.jpg) so the agent is expected to fetch and transcribe/translate/process untrusted third‑party media whose contents could embed instructions that influence downstream actions.