seedream-5-lite

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow ingests the outsider-authored --prompt string (and optional --image reference content) directly into client.generate_image(..., prompt=args.prompt, ...) and sends it as JSON to the provider’s POST /openapi/v1/generation/image endpoint.