{"version":"1.0","generated_at":"2026-10-15T13:57:18.482366","id":3425,"slug":"langchain-deep-agents","title":"LangChain 详解 Deep Agents 的上下文工程：如何记住信息而不塞满上下文窗口","summary":"LangChain 发布博客，说明其开源 Agent 框架 Deep Agents 如何通过 middleware 做上下文工程，以应对长任务中出现的 context rot 问题，核心在于让 Agent 记住必要信息的同时不把上下文窗口塞满。","abstract":"LangChain 详解 Deep Agents 的上下文工程：如何记住信息而不塞满上下文窗口 LangChain 发布博客，说明其开源 Agent 框架 Deep Agents 如何通过 middleware 做上下文工程，以应对长任务中出现的 context rot 问题，核心在于让 Agent 记住必要信息的同时不把上下文窗口塞满。 LangChain 博客介绍 Deep Agents 用 middleware 做上下文工程 目的是应对长任务中的 context rot，即上下文堆积导致质量衰减 上下文治理被提升为框架中间层能力，而非业务代码细节 对跑长链路任务的 Agent 团队，稳定性比单轮效果更关键 落地前需用真实长任务验证压缩策略的信息损耗 LangChain 发布博客，介绍其开源 Agent 框架 Deep Agents 如何用 middleware 做上下文工程，应对长任务中的 context rot。 这是什么信号 LangChain 官方博客把话题聚焦到一个具体的工程切口：开源 Agent 框架 Deep Agents 如何用 middleware 做上下文工程。关键词有两个——middleware 和 context rot。前者说明上下文处理被放在了框架的中间层，而不是散落在每个工具的调用逻辑里；后者说明问题被明确定义为：任务一旦拉长，上下文会随对话与工具结果堆积而逐步劣化。 这不是又发一个 Agent 产品，而是把「上下文怎么管」当作框架级问题来回答。它释放的信号是：Agent 的竞争点正在从「能不能调用工具」转向「长任务里还能不能保持判断质量」…","access_level":"member","access_label":"免费阅读","access_mode":"full","is_preview":false,"canonical_url":"https://opc.beizhux.com/content/3425/langchain-deep-agents","html_url":"https://opc.beizhux.com/content/3425/langchain-deep-agents","json_url":"https://opc.beizhux.com/content/3425/langchain-deep-agents.json","published_at":"2026-10-15T01:00:00","updated_at":"2026-10-15T13:16:12","category":{"slug":"ai-skills","name":"AI技能杠杆"},"source":{"site":"langchain.com","author":"LangChain：Blog（RSS）","url":"https://www.langchain.com/blog/context-engineering-deep-agents-remember"},"tags":["Agent 框架","context rot","Deep Agents","LangChain","middleware","上下文工程"],"topics":[{"slug":"agent-workflow","name":"Agent工作流","url":"https://opc.beizhux.com/topics/agent-workflow","reason":"这篇内容命中「Agent、工作流」等主题信号。"},{"slug":"ai-super-individual","name":"AI超级个体","url":"https://opc.beizhux.com/topics/ai-super-individual","reason":"这篇内容命中「技能」等主题信号。"},{"slug":"ai-tools","name":"AI工具","url":"https://opc.beizhux.com/topics/ai-tools","reason":"这篇内容来自该专题长期覆盖的栏目。"}],"keywords":["AI技能杠杆","Agent工作流","AI超级个体","AI工具","工具","自动化","模型","Cursor","Claude","Agent","智能体","工作流"],"questions":[{"question":"LangChain 详解 Deep Agents 的上下文工程：如何记住信息而不塞满上下文窗口主要讲什么？","answer":"LangChain 发布博客，说明其开源 Agent 框架 Deep Agents 如何通过 middleware 做上下文工程，以应对长任务中出现的 context rot 问题，核心在于让 Agent 记住必要信息的同时不把上下文窗口塞满。"},{"question":"这篇文章最值得关注的要点是什么？","answer":"LangChain 发布博客，说明其开源 Agent 框架 Deep Agents 如何通过 middleware 做上下文工程，以应对长任务中出现的 context rot 问题，核心在于让 Agent 记住必要信息的同时不把上下文窗口…；LangChain 博客介绍 Deep Agents 用 middleware 做上下文工程；目的是应对长任务中的 context rot，即上下文堆积导致质量衰减；上下文治理被提升为框架中间层能力，而非业务代码细节"},{"question":"这篇文章和哪些AI专题相关？","answer":"它适合放在Agent工作流、AI超级个体、AI工具专题里阅读。 关联原因：这篇内容命中「Agent、工作流」等主题信号。；这篇内容命中「技能」等主题信号。；这篇内容来自该专题长期覆盖的栏目。"},{"question":"阅读这篇文章建议先理解哪些关键词？","answer":"建议先理解AI工具、工具、自动化、模型、Cursor这些关键词，再结合正文判断工具、机会或风险是否值得进入自己的工作流。"}],"terms":[{"slug":"ai-tools-term","name":"AI工具","definition":"在AI觉醒星球里，「AI工具」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/ai-tools-term","path":"/glossary/ai-tools-term","json_url":"https://opc.beizhux.com/glossary/ai-tools-term.json"},{"slug":"tools","name":"工具","definition":"在AI觉醒星球里，「工具」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/tools","path":"/glossary/tools","json_url":"https://opc.beizhux.com/glossary/tools.json"},{"slug":"automation","name":"自动化","definition":"在AI觉醒星球里，「自动化」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/automation","path":"/glossary/automation","json_url":"https://opc.beizhux.com/glossary/automation.json"},{"slug":"ai-model","name":"模型","definition":"在AI觉醒星球里，「模型」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/ai-model","path":"/glossary/ai-model","json_url":"https://opc.beizhux.com/glossary/ai-model.json"},{"slug":"cursor","name":"Cursor","definition":"在AI觉醒星球里，「Cursor」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/cursor","path":"/glossary/cursor","json_url":"https://opc.beizhux.com/glossary/cursor.json"},{"slug":"claude","name":"Claude","definition":"在AI觉醒星球里，「Claude」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/claude","path":"/glossary/claude","json_url":"https://opc.beizhux.com/glossary/claude.json"},{"slug":"agent","name":"Agent","definition":"在AI觉醒星球里，「Agent」属于「Agent工作流」方向。聚合Agent、AI编程、Claude Code、Cursor、工作流自动化等内容，帮助你把智能体能力落到真实任务链路。 把智能体当成任务系统，而不是聊天窗口。","topic_slug":"agent-workflow","topic_name":"Agent工作流","topic_title":"Agent工作流：智能体、AI编程与自动化协作","topic_url":"https://opc.beizhux.com/topics/agent-workflow","topic_path":"/topics/agent-workflow","url":"https://opc.beizhux.com/glossary/agent","path":"/glossary/agent","json_url":"https://opc.beizhux.com/glossary/agent.json"},{"slug":"ai-agent","name":"智能体","definition":"在AI觉醒星球里，「智能体」属于「Agent工作流」方向。聚合Agent、AI编程、Claude Code、Cursor、工作流自动化等内容，帮助你把智能体能力落到真实任务链路。 把智能体当成任务系统，而不是聊天窗口。","topic_slug":"agent-workflow","topic_name":"Agent工作流","topic_title":"Agent工作流：智能体、AI编程与自动化协作","topic_url":"https://opc.beizhux.com/topics/agent-workflow","topic_path":"/topics/agent-workflow","url":"https://opc.beizhux.com/glossary/ai-agent","path":"/glossary/ai-agent","json_url":"https://opc.beizhux.com/glossary/ai-agent.json"}],"preview_hidden_sections":[],"related_items":[{"id":3416,"title":"Claude Code 空闲压缩从摘要继续，可用 idleCompaction 关闭并自定义阈值","url":"https://opc.beizhux.com/content/3416/claude-code-idlecompaction","summary":"Claude Code 的空闲压缩机制会让会话在空闲后从摘要恢复而非完整上下文。该信号给出两个具体控制手段：在 settings 中设置 \"idleCompaction\": false 可关闭该行为，以及通过环境变量 CLAUDE_CODE_IDLE_COMPACT_MIN_TOKENS 自定义触发阈值，默认值为 2…","access_level":"member"},{"id":3411,"title":"我们的工程博客新文：构建自研源代码控制","url":"https://opc.beizhux.com/content/3411/new-on-our-engineering-blog-creating-in-house-source-control-in-may","summary":"Anthropic 工程博客发布“构建自研源代码控制”一文。其代码合并量从 5 月每天约 2,500 次增至 9 月每天超过 30,000 次，原有面向人类规模设计的开发者基础设施成为瓶颈。","access_level":"member"},{"id":3391,"title":"OpenAI Devs：标注搭配 WebMCP，让智能体获得更多可执行上下文","url":"https://opc.beizhux.com/content/3391/annotations-paired-with-webmcp-give-your-agents-more-context-to-act-o","summary":"OpenAI Devs 展示：标注与 WebMCP 搭配可为智能体补充可执行上下文；@kagigz 用自定义标注控件和建议提示词，让用户在 ChatGPT 浏览器中编辑场景与更新费用。","access_level":"member"}],"usage_policy":{"summarizable":true,"preferred_url":"https://opc.beizhux.com/content/3425/langchain-deep-agents","private_fields_excluded":true}}