nano-banana-pro-product-detail-page

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 because the user-provided --prompt (and optional --system) is sent directly as messages[].content in client.chat_text() to the AI Hive POST /openapi/v1/chat/text endpoint during generate (via cmd_image/generate_image with args.prompt as well).