nano-banana-2

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s required runtime workflow is the user-provided --input (including free-text prompt) passed to the inference.sh CLI/app google/gemini-3-1-flash-image-preview, so outsider-authored text is directly ingested by the LLM at runtime.