model-routing-preset-builder

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow ingests outsider-authored free text at runtime by reading user-owned task/thread content via channels like create_thread, read_thread, and send_message_to_thread (e.g., routes.field_coordinator, routes.field_fleet, routes.delegated_subtask, routes.parallel_task), so an outsider can poison what the agent later reads.