当前位置:首页 > 报告详情

为 Kiro 添加 MCP 支持以实现更丰富的上下文.pdf

上传人: 明**** 编号:1012591 2025-12-21 30页 352.93KB

1、 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.D V T 2 1 3Extend Kiro with MCP support for richer contextArundeep Nagaraj(he/him)Sr.SA Manager,KiroNathan Jones(he/him)Sr.Manager,Software Dev,Kiro 2025,Amazon Web

2、Services,Inc.or its affiliates.All rights reserved.Show of hands.Who here has used or is familiar with Kiro?How many of you are currently building or using AI agents in your development workflow?2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Context EngineeringRetrieval Augmented

3、 Generation(RAG)Model Context Protocol(MCP)Remote MCP with KiroUnlocking productivity for KiroDemo 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Context Engineering 2025,Amazon Web Services,Inc.or its affiliates.

4、All rights reserved.The Evolution of AI EngineeringContext EngineeringContext EngineeringPrompt EngineeringPrompt EngineeringRAGRAGAgentsAgentsAdd KnowledgeAdd Tools&AutonomyProduction at Scale 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Context is everything 2025,Amazon Web S

5、ervices,Inc.or its affiliates.All rights reserved.Long autonomous loops accumulate lots of messagesContext is everything in the messages array 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.“Context engineering is the delicate art and science of filling the context window with ju

6、st the right information for the next step.”Andrey KarpathyFounding team OpenAI 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.System ContextRole definition,behavioral instructions and sets the foundationConversational ContextPrevious messages in the conversation,and memoryRetrie

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
根据报告的内容,全文主要内容概括如下: - **Kiro扩展与MCP支持**:Kiro平台增加了对Model Context Protocol (MCP)的支持,以增强上下文工程能力。 - **上下文工程**:强调上下文在AI工程中的重要性,包括系统上下文、对话上下文和工具上下文。 - **RAG技术**:介绍RAG(Retrieval Augmented Generation)技术,它通过检索和增强上下文来提高AI模型的生成质量。 - **MCP协议**:介绍MCP作为开放标准,简化AI与数据、工具的集成。 - **MCP应用**:MCP可用于查询工具、调用API、获取格式化响应等。 - **远程MCP**:Kiro支持远程MCP,允许与外部系统集成,提高生产力。 - **反馈循环**:强调通过观察结果和调整来闭环反馈,实现真正的智能代理行为。 - **Kiro演示和活动**:提供Kiro的演示和活动信息,包括白板会议、Kiosk和Meetups。
AI生产力新篇章?" AI工具的完美融合?" Kiro如何提升开发效率?"
客服
商务合作
小程序
服务号
折叠