luma-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 accepts arbitrary public start_image_url/end_image_url and video_url (see Parameters and Gotchas) so the skill will fetch and consume untrusted, user-provided web content to guide generation, which can materially influence the agent's outputs and behavior.