trigger-chat-agent-advanced

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

• Third-party content exposure detected (medium risk: 0.30). The chat.agent / sessions runtime workflow ingests outsider-authored message text that clients send into the session input channel (e.g., AgentChat.sendMessage() and the wire endpoint POST /realtime/v1/sessions/{id}/in/append with kind: "message"), which the agent then feeds into streamText as messages.