runpod-comfyui-models

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

• Third-party content exposure detected (medium risk: 0.30). Source-user-supplied content is ingested as ComfyUI workflow JSON/PNG at runtime (the agent extracts embedded workflow JSON from the user-provided PNG and then inventories/resolves it), so attacker-authored free text can directly enter the LLM processing pipeline via those workflow fields and missing-model identifiers.