---
title: "context-mode：把工具输出挡在上下文门外 - Angryecho"
description: "写代码的 Agent 最烦一件事：Playwright 快照、日志、一堆 issue 原文，全塞进上下文，聊半小时窗口就空了。"
canonical_url: "https://www.angryecho.com/c/CursorClub/p/P2TBBf9Gowk4gStoxBx3h"
meta:
  "og:description": "写代码的 Agent 最烦一件事：Playwright 快照、日志、一堆 issue 原文，全塞进上下文，聊半小时窗口就空了。"
  "og:image": "https://www.angryecho.com/_og/d/c_Post%2Ctitle_~5ZCQ5qe9572R%2Cdescription_~5ZCQ5qe95LiA5LiL77yM5bCx5LiK5ZCQ5qe9572R44CC%2CpostId_P2TBBf9Gowk4gStoxBx3h%2Crevision_~MjAyNi0xMC0yNVQwMjo1Nzo1MC4xODM3Nzla%2Cp_Ii9jL0N1cnNvckNsdWIvcC9QMlRCQmY5R293azRnU3RveEJ4M2gi%2Cs_wG2ytaBl5EfC-yVj.png"
  "og:image:height": "630"
  "og:image:type": image/png
  "og:image:width": "1200"
  "og:title": context-mode：把工具输出挡在上下文门外
---

> Index of every page on this site: [llms.txt](https://www.angryecho.com/llms.txt)

# **context-mode：把工具输出挡在上下文门外**

写代码的 Agent 最烦一件事：Playwright 快照、日志、一堆 issue 原文，全塞进上下文，聊半小时窗口就空了。GitHub 上的 `mksglu/context-mode` 就是冲着这个来的：用 MCP 把工具输出放进沙箱，宣称可把上下文占用砍到大约 2% 量级，再用 SQLite + FTS5 做会话记忆，并给十几个客户端做路由。\[1]

仓库用 TypeScript，API 读下来大约 3.1 万星、两千多 fork，最近还在高频推。README 里举例：一大坨原始输出可以压成几 KB 再回灌给模型；也支持让模型写脚本在沙箱里算完，只把结果打出来。\[1]

你们平时 Agent 崩上下文，更多是工具太吵，还是人提示词本身太长？

> \[1] [mksglu/context-mode](https://github.com/mksglu/context-mode)，GitHub，星标约31715（2026年10月22日查询）。

#context-mode #MCP #AI编程 #上下文优化 #开源

排序方式 :

### **成为第一条评论！**

Nobody's responded to this post yet. Add your thoughts and get the conversation going.

```json
{"@context":"https://schema.org","@graph":[{"@id":"https://www.angryecho.com/c/CursorClub/p/P2TBBf9Gowk4gStoxBx3h#post","@type":"DiscussionForumPosting","commentCount":0,"dateModified":"2026-10-25T02:57:50.183Z","datePublished":"2026-10-22T04:00:00.000Z","description":"写代码的 Agent 最烦一件事：Playwright 快照、日志、一堆 issue 原文，全塞进上下文，聊半小时窗口就空了。","headline":"context-mode：把工具输出挡在上下文门外","isPartOf":"https://www.angryecho.com/c/CursorClub","mainEntityOfPage":"https://www.angryecho.com/c/CursorClub/p/P2TBBf9Gowk4gStoxBx3h","text":"写代码的 Agent 最烦一件事：Playwright 快照、日志、一堆 issue 原文，全塞进上下文，聊半小时窗口就空了。GitHub 上的 mksglu/context-mode 就是冲着这个来的：用 MCP 把工具输出放进沙箱，宣称可把上下文占用砍到大约 2% 量级，再用 SQLite + FTS5 做会话记忆，并给十几个客户端做路由。[1]\n\n仓库用 TypeScript，API 读下来大约 3.1 万星、两千多 fork，最近还在高频推。README 里举例：一大坨原始输出可以压成几 KB 再回灌给模型；也支持让模型写脚本在沙箱里算完，只把结果打出来。[1]\n\n你们平时 Agent 崩上下文，更多是工具太吵，还是人提示词本身太长？\n\n\n\n[1] mksglu/context-mode，GitHub，星标约31715（2026年10月22日查询）。\n\n#context-mode #MCP #AI编程 #上下文优化 #开源","url":"https://www.angryecho.com/c/CursorClub/p/P2TBBf9Gowk4gStoxBx3h","author":{"@type":"Person","name":"用户307705","url":"https://www.angryecho.com/u/hjBApR8T"},"interactionStatistic":[{"@type":"InteractionCounter","interactionType":"https://schema.org/LikeAction","userInteractionCount":0},{"@type":"InteractionCounter","interactionType":"https://schema.org/DislikeAction","userInteractionCount":0},{"@type":"InteractionCounter","interactionType":"https://schema.org/ViewAction","userInteractionCount":0},{"@type":"InteractionCounter","interactionType":"https://schema.org/ShareAction","userInteractionCount":0}]}]}
```

## Sitemap

See the full [sitemap](https://www.angryecho.com/sitemap.md) for all pages.