character-consistency

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow reads outsider-authored free text from the CLI arguments --prompt (and optionally --system) and sends it directly as messages[].content to the chat/text or generation/image endpoints (e.g., cmd_image → client.generate_image → _request JSON body).