nano-banana-2-livestream-image

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow ingests outsider-authored free text from the user-provided --prompt (and optional --system in cmd_chat) and sends it directly to the AI Hive API via client.generate_image(... prompt ...), so an outsider can submit arbitrary prompt content.