building-streaming-pipelines

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via runtime message bodies consumed from external systems (e.g., Kafka topic messages, MQTT payloads, or NATS JetStream messages) that are decoded as JSON/prose and then typically summarized/processed by the agent; this is outsider-authored content from other producers.