image-edit

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow posts the user-supplied prompt (free text) and user-supplied image_urls/images/mask_image URLs into runcomfy run <model_id> (image edit model invocation), so outsider-authored text is directly ingested by the LLM without selecting a specific preexisting item.