gpt-image-2-marketing-image

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

• Third-party content exposure detected (medium risk: 0.30). The runtime path scripts/imagegen.py takes outsider-supplied --prompt (and optional --image) and forwards it as JSON prompt in client.generate_image(...) → POST .../generation/image, so the agent ingests arbitrary free text directly provided by the user at execution time.