broadway-data-pipelines

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow ingests message bodies from an outsider-authored producer (e.g., SQS/Kafka/RabbitMQ messages) via the Broadway producer modules, and those message data/status.reason values are then logged/handled in handle_failed/2 and passed through Broadway.Message.failed/2, making them readable LLM context inputs.