alicloud-ai-video-digital-human

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly requires and transmits arbitrary public HTTP/HTTPS image and audio URLs (see SKILL.md Quick start and scripts/prepare_digital_human_request.py) so the agent ingests untrusted third‑party content that directly affects detection, generation decisions, and resulting tool behavior.