nano-banana-image-edit

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

• Third-party content exposure detected (medium risk: 0.30). Outsider-authored free text via the CLI --prompt (and optional --system) is sent at runtime to the provider in cmd_image() → client.generate_image() → POST /openapi/v1/generation/image where the prompt is ingested by the model.