image-to-image-edit

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly accepts and loads arbitrary public image URLs (see the required image_url parameter, the "Prepare Source Image" step, and the ComfyUI "LoadImage" backend step), so untrusted user-generated images from the open web are ingested into the img2img workflow and could indirectly influence model behavior or subsequent actions.