用 Amazon Bedrock AgentCore 构建事件驱动、人在回路的环境智能体
原标题：Building ambient agents with Amazon Bedrock AgentCore: From event-driven signals to human-in-the-loop workflows
AI 摘要
AWS Machine Learning Blog 发布技术文章，介绍如何用 Amazon Bedrock AgentCore 构建「环境智能体」（ambient agents）。这类智能体不再依赖用户主动发起对话，而是监听 S3 事件通知、EventBridge、Lambda 触发器等事件流并自动执行任务，在需要时通过单一的 ask_human 工具暂停等待人工审批或澄清，再从中断处恢复。文章给出端到端参考实现，结合 AgentCore Runtime、Lambda 与 DynamoDB 构建全无服务器平台，并列出部署所需的 AWS 账户、CDK、Docker 及 Claude Sonnet 4.5 模型访问等前置条件。
正文节选
Building ambient agents with Amazon Bedrock AgentCore: From event-driven signals to human-in-the-loop workflows Teams that process documents at scale know the routine: files land in storage, someone notices, opens each one, decides what it needs, and routes it for review. Monitoring alerts queue up the same way, waiting for a person to act on them. The hours lost to manual triage are the operational problem ambient agents solve. Imagine a document lands in your Amazon Simple Storage Service (Ama