Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Description from the maker (original English)
Effect 官方文档的中文译文检索(234 页,v3 + v4)。每条结果带可核验引用:能点回原文小节、能看到该译文对着哪次上游提交译的。零依赖、离线自包含。
These three are never merged into one score: opinions, measurements and popularity are different things.
We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.
Pick the AI app you use, then copy the settings.
You need
Which app do you use?
Pick your app above. We remember it, so next time the button says “Install in …”.
Open Claude Desktop → Settings → Developer → Edit Config, and paste this into claude_desktop_config.json
{
"mcpServers": {
"effect-ts-cn": {
"command": "npx",
"args": [
"-y",
"effect-ts-cn-mcp"
]
}
}
}Run this in your terminal
claude mcp add effect-ts-cn -- npx -y effect-ts-cn-mcpPaste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"effect-ts-cn": {
"command": "npx",
"args": [
"-y",
"effect-ts-cn-mcp"
]
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
Paste into .vscode/mcp.json in your project
{
"servers": {
"effect-ts-cn": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"effect-ts-cn-mcp"
]
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
This one runs on your computer, and ChatGPT can only connect to servers on the internet. Pick another app.
Your choice stays on this device in a cookie; no account needed.
These settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.
Runs on your computer; the tool list appears in your AI app after you install it.
We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 7, 2026
The registry name was verified with the GitHub account aaronlou, which owns the repo
We have not read this repo from GitHub yet
We have not read this repo from GitHub yet
This one runs on your computer; there is no internet endpoint
Runs on your computer and uses local settings instead
Runs on your computer; we can only read tool lists from internet endpoints
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.
The registry does not list a required key, though some servers ask you to sign in on first use.
It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).
The GitHub repo last changed –, and the latest version (0.1.0) reached the registry 2 months ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need Node.js on your computer.
In the same category: Spec Driven Development, Remote Desktop Commander, Dialyne
Servers in the same category from other makers, sorted by stars.
Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Real phone calls in France for AI agents: dial +33 numbers, menus, hold, AI disclosure.
Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.
站点:[https://effect-ts.cn](https://effect-ts.cn) —— Effect(TypeScript 的 effect system)官方文档全量中文翻译
(v4 + v3 共 234 页,每页标注上游基线)+ 可溯源的问答(每条引用可点回原文,无依据即拒答)
+ 报错百科(真实报错的检索索引)+ 26 个用 Effect 写的开源 AI 项目推荐。
给 AI / Agent 的入口:
/llms.txt·/llms-full.txt·/docs/ .md·/cite/ .json(引用核验记录)
· MCP server(仓库内
pnpm mcp)。
中文世界 Effect 的第一入口:官方文档的高质量中文译站 + 内容社区 + 由 Effect 与 DDD 自建的开源标本。
reviewers 一律为 ecn-review,含义是机器可复核 (代码块与上游逐字节一致、结构对齐、术语 0 命中、锚点可达),非人类精读,尚无人类精读署名; 展示口径见 apps/site/src/data/provenance.ts,维护者精读抽查后请追加自己的名字。pnpm install
pnpm dev
cp .env.example .env # 填 DEEPSEEK_API_KEY=sk-...(或 LLM_BASE_URL + LLM_API_KEY) pnpm --filter @ecn/api llm:check # 一条命令验证模型真的接上了(打印提供方/模型/超时)
pnpm dev 会 watch 这几个文件:改完 Key 自动重启,不用手动重启pnpm db:up # docker 起 Postgres;或用本机 postgres 亦可
/api/* 已由 dev 代理到后端)GET /api/health 健康检查GET /api/questions 问题列表 / POST /api/questions 提问 / GET /api/questions/:id 详情GET /openapi.json 自动生成的 OpenAPI(来自 @effect/schema 契约)pnpm typecheck # 全仓类型检查(含 astro check)
pnpm test # 各包测试(api: domain/application;content: 内容门禁) pnpm build # contracts 编译 + api typecheck + astro 构建 pnpm content:check # 内容门禁:frontmatter / 路径镜像 / 术语 / 元数据残留(PR 必过) pnpm cite:check # 引用协议门禁:摘要 / 内容指纹 / 构建产物一致(保证引用可解引用) pnpm proposals:check # Agent 提案队列校验:治理不变量 + 复用内容门禁 pnpm content:status # 译文同步状态扫描 pnpm corpus:build # 生成 AI 知识层语料(内容改动后必跑;CI 有新鲜度门禁) pnpm mcp # 启动 MCP Server(stdio),把中文知识接进编码 Agent pnpm report:weekly # 周报:AI 用量 / 可验证答率 / 缓存命中 / Agent 侧下载量(见 docs/metrics.md) pnpm traffic # 访问报表(宿主 Caddy 日志;pnpm traffic:prod 从生产拉取) pnpm --filter @ecn/api llm:check # 用真实模型跑一次问答+报错诊断(验证 DeepSeek/OpenAI 兼容配置)
pnpm --filter @ecn/content exec tsx src/cli.ts snapshot --dir <上游docs> -o snap.json pnpm --filter @ecn/content exec tsx src/cli.ts diff --snapshot /abs/snap.json --docs /abs/apps/site/src/content/docs pnpm --filter @ecn/content exec tsx src/cli.ts nav --dir <上游docs> -o apps/site/src/data/docs-nav.json
| 能力 | 说明 |
|---|---|
| 文档译站 | 目录镜像官方结构,v3 + v4 234/234 页全部有中文译文(v4 110 · v3 124);未翻译页面仍会自动生成占位页(读英文原文 + 认领翻译),站内无死链 |
| 可追溯同步 | 每篇译文标注 upstreamPath + upstreamCommit;每日流水线比对上游,落后/导航漂移自动开 issue |
| 站内搜索 | ⌘/Ctrl + K,构建期索引(已译文字全、未译页面标题),零后端依赖 |
| 订阅与 AI 友好 | /rss.xml、/llms.txt、sitemap-index.xml、robots.txt |
| 阅读体验 | 侧边栏(镜像官方)、页内 TOC、上下页、版本切换、代码块「复制 / Playground」、官方 Aside/Steps/Tabs 组件 |
| 内容门禁 | PR 阶段拦截:frontmatter 必填、路径镜像、术语黑名单、twoslash/框架 import 残留、页内锚点失效 |
| 社区协作 | 行为准则、Issue 模板(翻译认领 / 站点问题)、PR 自查清单、术语表页面 |
| AI 知识层 | 「问这一页 / 问文档」(⌘I)与「报错诊断」(/debug):答案逐句带引用(页面+小节+基线),无依据直接拒答,并区分"文档没有"与"中文尚未翻译";术语门禁同样约束 AI 输出。配模型后是一条会话:追问会说人话(「它呢?」被改写成完整查询并回显 resolvedQuestion),白话问题先被改写成术语再检索(「怎么让两件事同时跑?」→ Fiber / 并发),候选重排只换顺序不动引用 |
| 引用可核验 | 每条引用都带 citationId 与 /cite/ .json:可独立核对「引用是否是原文的逐字子串」、译文基线是否已漂移、以及该基线下的官方原文 —— 引用不是修辞,是可取证的事实 |
| 生态项目榜 | /ecosystem/:用 Effect 写的 AI / Agent / LLM 开源项目精选。收录判据不看 README 看 `package.json`(必须运行时依赖 effect/@effect/*,且全仓至少一个文件真的 import 它 —— "声明了没人用"会被剔除),并标注Effect 渗透度与「该读哪一块」(每条建议都指向真实存在的文件,由采集器核对)。数据是快照 + checkedAt,页面上如实标注"截至某日" |
| 生态观测台 | /observatory/:可复现的 GitHub 生态调查 —— 4,861 个候选、856 个 TypeScript Agent、33 个真正在用 Effect(3.9%)。数据 CSV 可下载,口径与方法写在 docs/observatory/methodology.md;每个数字都从 dataset/ 现算,不手抄 |
| 选中即讲 | 选中正文里的一段,像素风吉祥物「小效」跑到选区旁问一句"要我讲讲这段吗"(并显示这段的 slug#anchor):「讲讲」直接问、「换个问法」只预填、绝不自动提交;同一段每会话只问一次、每页最多主动问 3 次、可全局关掉 |
| Agent 起草 → 人审 | .proposals/ 提案队列:Agent 起草译文与落后页更新,内容自动过与人工投稿完全相同的门禁;且不得自称已发布(status 只能是 reviewing、reviewers 必须为空) |
| 隐私与统计 | /privacy/ 如实说明记录什么(访问日志、AI 提问内容)、留多久、给了谁;不用 Google Analytics(大陆不可达,会系统性低估真正的受众),统计走服务端结构化日志(pnpm traffic 出报表)+ 可选的自建 Umami;广告默认关闭,开启时只在正文末尾与列表页底部、预留高度、明示"广告" |
| Agent 接入 | HTTP /api/knowledge/ask、MCP Server(pnpm mcp,6 个工具 + resources + prompts)、/llms.txt、/llms-full.txt、/docs/ .md、/cite/index.json |
apps/site Astro 前台(内容集合/MDX,React islands,SEO 优先;AskPanel 问这一页)
apps/api Effect 后端(@effect/platform HTTP,DDD 洋葱分层;含 Knowledge/Assistant 上下文) apps/mcp 中文知识层的 MCP Server(stdio,离线自包含) packages/knowledge 检索与答案合成(BM25F-lite、中文分词、话题归属、引用不变量) packages/contracts 前后端共享 Effect Schema DTO + 错误码(Schema-first) packages/content 内容管线 CLI(门禁校验 / 上游快照与 stale 比对 / 导航生成) infra/ docker-compose(本地 Postgres) docs/ 译者指南、部署指南、术语黑名单 PLAN.md 产品与技术规划(含 DDD 设计与路线图)
| 层 | 选型 | 一句话理由 |
|---|---|---|
| 前台 | Astro + React islands | 与官方 effect.website 同思路:内容/SEO 最优、默认零 JS |
| 后端 | Effect + @effect/platform | 类型化错误、显式依赖、结构化并发 —— dogfood |
| 契约 | @effect/schema | 一份 Schema → DTO + 类型 + OpenAPI,前后端零重复 |
| 数据 | PostgreSQL(@effect/sql-pg) | 关系模型 + 未来中文全文检索(pg_jieba) |
| 测试 | Vitest + Effect Test 思想 | 金字塔:domain → application → 集成 |
后端 DDD 要点(详见 PLAN §5):
Identity / Publishing / QnA / Curation / Notification(Moderation 先内嵌)domain / application / infrastructure / interfacesData.TaggedError,wire 错误用 @effect/schema 的 Schema.TaggedError(在 interfaces 层完成映射)Layer 可移植性的活例子:同一个
QuestionRepository端口,bootstrap/main.ts按
DATABASE_URL是否存在,在 InMemory 仓储与 Postgres 仓储之间切换 —— 这就是依赖倒置 + DI 容器。
测试里也用 Layer 注入确定性 ID 与内存仓储。
.github/workflows/ci.yml:内容门禁 → typecheck → test → build。.github/workflows/upstream-sync.yml:每日定时(可手动触发)克隆官方内容仓库 `Effect-TS/website` 的 `apps/web/src/content/docs`(官方文档真正的源 —— Effect-TS/effect 仓库里并没有 docs/)→ 生成上游快照 → 比对译文落后 → 检查导航漂移 → 自动开/更新 upstream-sync 标签的 issue。本地可复用同一套能力:packages/content 的 snapshot / diff / nav / check 子命令。
status: published 并记录译者 / 审校 / 上游基线(@bf46254);审校为 ecn-review(机器可复核,见上「审校口径」)llms.txt、robots、sitemap、404、canonical/OG 交互切换;人类精读审校署名;社区功能(Phase 2)