conversation-memory

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

• Third-party content exposure detected (medium risk: 0.30). In the required runtime workflow, the agent stores and reuses free-form message.content from the conversation buffer and then injects it back into prompts via buildContext/promptWithMemory (e.g., Current Query/User Context/Recent Conversation and entity extraction), so outsider-authored text can flow into the LLM.