From e295751eba3460f97f09c69ec1c7ef774e6c07b9 Mon Sep 17 00:00:00 2001 From: patrick-eu Date: Thu, 9 Jul 2026 10:35:52 +0000 Subject: [PATCH] =?UTF-8?q?=E6=9D=A1=E7=9B=AE=E8=87=AA=E5=8A=A8=E5=B1=95?= =?UTF-8?q?=E7=A4=BA=20star=20=E6=95=B0=E4=B8=8E=E6=9C=80=E5=90=8E?= =?UTF-8?q?=E6=9B=B4=E6=96=B0=E6=97=A5=E6=9C=9F?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit check.py 新增 --update：把体检时拉到的数据写回 README 条目； 每周 CI 体检后自动提交刷新。 Co-Authored-By: Claude Fable 5 --- .github/workflows/check.yml | 12 ++- README.en.md | 158 ++++++++++++++++++------------------ README.md | 158 ++++++++++++++++++------------------ check.py | 35 +++++++- 4 files changed, 201 insertions(+), 162 deletions(-) diff --git a/.github/workflows/check.yml b/.github/workflows/check.yml index 726cb7d..dda1ba9 100644 --- a/.github/workflows/check.yml +++ b/.github/workflows/check.yml @@ -8,11 +8,21 @@ on: - cron: "0 6 * * 1" # 每周一 workflow_dispatch: +permissions: + contents: write + jobs: check: runs-on: ubuntu-latest steps: - uses: actions/checkout@v4 - - run: python3 check.py + - name: 体检并刷新 star 数 + run: python3 check.py --update env: GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} + - name: 提交刷新结果 + if: github.event_name != 'pull_request' + run: | + git config user.name "github-actions[bot]" + git config user.email "41898282+github-actions[bot]@users.noreply.github.com" + git diff --quiet || { git add README.md README.en.md && git commit -m "chore: 自动刷新 star 数与更新日期" && git push; } diff --git a/README.en.md b/README.en.md index a551dc3..da29f21 100644 --- a/README.en.md +++ b/README.en.md @@ -23,7 +23,7 @@ To keep out malicious and abandoned code, every listed repo must satisfy all of: - 🔄 **committed to within the last 6 months** - 📄 **a clear open-source license** -Enforced automatically by [`check.py`](check.py); GitHub Actions re-checks weekly and failing repos get removed. +Enforced automatically by [`check.py`](check.py); GitHub Actions re-checks weekly and failing repos get removed. Each entry's `⭐ stars · last-updated date` badge is also refreshed weekly by CI. --- @@ -31,113 +31,113 @@ Enforced automatically by [`check.py`](check.py); GitHub Actions re-checks weekl The agents and assistants that actually write the code. -- [anthropics/claude-code](https://github.com/anthropics/claude-code) — Terminal-based AI coding agent that understands your whole codebase and executes multi-step tasks. -- [cline/cline](https://github.com/cline/cline) — Autonomous coding agent in VS Code: creates/edits files, runs commands, uses the browser. -- [All-Hands-AI/OpenHands](https://github.com/All-Hands-AI/OpenHands) — Open-source AI software engineer platform that completes full development tasks. -- [Aider-AI/aider](https://github.com/Aider-AI/aider) — Pair programming in your terminal, editing your local git repo directly. -- [continuedev/continue](https://github.com/continuedev/continue) — Open-source IDE assistant with customizable models and context sources. -- [RooCodeInc/Roo-Code](https://github.com/RooCodeInc/Roo-Code) — Community-enhanced fork of Cline with multi-mode agent teams. -- [openai/codex](https://github.com/openai/codex) — OpenAI's terminal coding agent, written in Rust, light and fast. -- [google-gemini/gemini-cli](https://github.com/google-gemini/gemini-cli) — Google's open-source terminal AI agent, wired to Gemini. -- [anomalyco/opencode](https://github.com/anomalyco/opencode) — Open-source terminal coding agent; swap in any model you like. -- [aaif-goose/goose](https://github.com/aaif-goose/goose) — Local open-source AI agent, extensible for automating engineering tasks. -- [TabbyML/tabby](https://github.com/TabbyML/tabby) — Self-hosted AI code completion; the open-source Copilot alternative. -- [Kilo-Org/kilocode](https://github.com/Kilo-Org/kilocode) — Open-source VS Code AI agent combining the best of Cline and Roo. +- [anthropics/claude-code](https://github.com/anthropics/claude-code) `⭐137k · 2026-07-08` — Terminal-based AI coding agent that understands your whole codebase and executes multi-step tasks. +- [cline/cline](https://github.com/cline/cline) `⭐64.5k · 2026-07-09` — Autonomous coding agent in VS Code: creates/edits files, runs commands, uses the browser. +- [All-Hands-AI/OpenHands](https://github.com/All-Hands-AI/OpenHands) `⭐80.1k · 2026-07-09` — Open-source AI software engineer platform that completes full development tasks. +- [Aider-AI/aider](https://github.com/Aider-AI/aider) `⭐47.2k · 2026-05-22` — Pair programming in your terminal, editing your local git repo directly. +- [continuedev/continue](https://github.com/continuedev/continue) `⭐34.8k · 2026-07-09` — Open-source IDE assistant with customizable models and context sources. +- [RooCodeInc/Roo-Code](https://github.com/RooCodeInc/Roo-Code) `⭐24.3k · 2026-05-15` — Community-enhanced fork of Cline with multi-mode agent teams. +- [openai/codex](https://github.com/openai/codex) `⭐96.5k · 2026-07-09` — OpenAI's terminal coding agent, written in Rust, light and fast. +- [google-gemini/gemini-cli](https://github.com/google-gemini/gemini-cli) `⭐105.9k · 2026-07-09` — Google's open-source terminal AI agent, wired to Gemini. +- [anomalyco/opencode](https://github.com/anomalyco/opencode) `⭐184k · 2026-07-09` — Open-source terminal coding agent; swap in any model you like. +- [aaif-goose/goose](https://github.com/aaif-goose/goose) `⭐50.9k · 2026-07-09` — Local open-source AI agent, extensible for automating engineering tasks. +- [TabbyML/tabby](https://github.com/TabbyML/tabby) `⭐33.7k · 2026-06-30` — Self-hosted AI code completion; the open-source Copilot alternative. +- [Kilo-Org/kilocode](https://github.com/Kilo-Org/kilocode) `⭐25.9k · 2026-07-09` — Open-source VS Code AI agent combining the best of Cline and Roo. ## 📝 Prompts & Rules Resources for setting rules and stealing inspiration. -- [PatrickJS/awesome-cursorrules](https://github.com/PatrickJS/awesome-cursorrules) — Ready-to-use .cursorrules for every stack. -- [f/awesome-chatgpt-prompts](https://github.com/f/awesome-chatgpt-prompts) — The classic prompt collection; a reference for writing system prompts. -- [x1xhlol/system-prompts-and-models-of-ai-tools](https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools) — Reverse-engineered system prompts of major AI coding tools; learn how the pros steer models. -- [anthropics/skills](https://github.com/anthropics/skills) — Anthropic's official Agent Skills collection; give Claude domain expertise. -- [obra/superpowers](https://github.com/obra/superpowers) — A full skill kit for coding agents: brainstorming, TDD, systematic debugging and more. -- [dair-ai/Prompt-Engineering-Guide](https://github.com/dair-ai/Prompt-Engineering-Guide) — The systematic prompt-engineering curriculum: papers, techniques, examples. -- [anthropics/prompt-eng-interactive-tutorial](https://github.com/anthropics/prompt-eng-interactive-tutorial) — Anthropic's official interactive prompt-engineering tutorial. -- [NirDiamant/Prompt_Engineering](https://github.com/NirDiamant/Prompt_Engineering) — Hands-on prompt technique tutorials, basic to advanced, all with code. -- [ai-boost/awesome-prompts](https://github.com/ai-boost/awesome-prompts) — Curated prompts from top-rated GPTs plus prompt-engineering resources. +- [PatrickJS/awesome-cursorrules](https://github.com/PatrickJS/awesome-cursorrules) `⭐40.3k · 2026-05-30` — Ready-to-use .cursorrules for every stack. +- [f/awesome-chatgpt-prompts](https://github.com/f/awesome-chatgpt-prompts) `⭐165.1k · 2026-07-09` — The classic prompt collection; a reference for writing system prompts. +- [x1xhlol/system-prompts-and-models-of-ai-tools](https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools) `⭐141.7k · 2026-07-08` — Reverse-engineered system prompts of major AI coding tools; learn how the pros steer models. +- [anthropics/skills](https://github.com/anthropics/skills) `⭐159.7k · 2026-07-01` — Anthropic's official Agent Skills collection; give Claude domain expertise. +- [obra/superpowers](https://github.com/obra/superpowers) `⭐250.4k · 2026-07-06` — A full skill kit for coding agents: brainstorming, TDD, systematic debugging and more. +- [dair-ai/Prompt-Engineering-Guide](https://github.com/dair-ai/Prompt-Engineering-Guide) `⭐76.3k · 2026-03-11` — The systematic prompt-engineering curriculum: papers, techniques, examples. +- [anthropics/prompt-eng-interactive-tutorial](https://github.com/anthropics/prompt-eng-interactive-tutorial) `⭐36.9k · 2026-03-01` — Anthropic's official interactive prompt-engineering tutorial. +- [NirDiamant/Prompt_Engineering](https://github.com/NirDiamant/Prompt_Engineering) `⭐7.7k · 2026-07-04` — Hands-on prompt technique tutorials, basic to advanced, all with code. +- [ai-boost/awesome-prompts](https://github.com/ai-boost/awesome-prompts) `⭐8.4k · 2026-07-09` — Curated prompts from top-rated GPTs plus prompt-engineering resources. ## 📚 Making Code & Docs AI-Readable Turn repos, web pages, and docs into context an AI can digest. -- [coderamp-labs/gitingest](https://github.com/coderamp-labs/gitingest) — One-click conversion of any Git repo into an LLM-friendly text digest. -- [upstash/context7](https://github.com/upstash/context7) — MCP server giving the AI live access to up-to-date official docs and code examples. -- [AsyncFuncAI/deepwiki-open](https://github.com/AsyncFuncAI/deepwiki-open) — Auto-generates a wiki for any repo, helping both AI and humans understand a project fast. -- [yamadashy/repomix](https://github.com/yamadashy/repomix) — Packs an entire codebase into one AI-friendly file; popular gitingest sibling. -- [microsoft/markitdown](https://github.com/microsoft/markitdown) — Microsoft's converter turning Office/PDF and more into LLM-ready Markdown. -- [docling-project/docling](https://github.com/docling-project/docling) — IBM's open-source document parser: PDFs, tables, layout, structured output. -- [opendatalab/MinerU](https://github.com/opendatalab/MinerU) — High-quality PDF to Markdown/JSON, formulas and tables included. -- [datalab-to/marker](https://github.com/datalab-to/marker) — Fast, accurate PDF-to-Markdown with solid formula/table support. -- [jina-ai/reader](https://github.com/jina-ai/reader) — Prefix any URL with r.jina.ai and get LLM-friendly body text. -- [Unstructured-IO/unstructured](https://github.com/Unstructured-IO/unstructured) — ETL preprocessing for documents of every format; a RAG-pipeline staple. +- [coderamp-labs/gitingest](https://github.com/coderamp-labs/gitingest) `⭐15k · 2026-07-08` — One-click conversion of any Git repo into an LLM-friendly text digest. +- [upstash/context7](https://github.com/upstash/context7) `⭐58.8k · 2026-07-09` — MCP server giving the AI live access to up-to-date official docs and code examples. +- [AsyncFuncAI/deepwiki-open](https://github.com/AsyncFuncAI/deepwiki-open) `⭐17.2k · 2026-06-03` — Auto-generates a wiki for any repo, helping both AI and humans understand a project fast. +- [yamadashy/repomix](https://github.com/yamadashy/repomix) `⭐27k · 2026-07-05` — Packs an entire codebase into one AI-friendly file; popular gitingest sibling. +- [microsoft/markitdown](https://github.com/microsoft/markitdown) `⭐164.2k · 2026-06-24` — Microsoft's converter turning Office/PDF and more into LLM-ready Markdown. +- [docling-project/docling](https://github.com/docling-project/docling) `⭐62.9k · 2026-07-09` — IBM's open-source document parser: PDFs, tables, layout, structured output. +- [opendatalab/MinerU](https://github.com/opendatalab/MinerU) `⭐74k · 2026-07-09` — High-quality PDF to Markdown/JSON, formulas and tables included. +- [datalab-to/marker](https://github.com/datalab-to/marker) `⭐37.3k · 2026-07-07` — Fast, accurate PDF-to-Markdown with solid formula/table support. +- [jina-ai/reader](https://github.com/jina-ai/reader) `⭐11.5k · 2026-05-22` — Prefix any URL with r.jina.ai and get LLM-friendly body text. +- [Unstructured-IO/unstructured](https://github.com/Unstructured-IO/unstructured) `⭐15.1k · 2026-07-09` — ETL preprocessing for documents of every format; a RAG-pipeline staple. ## 🕷️ Scraping & Data Collection -- [scrapy/scrapy](https://github.com/scrapy/scrapy) — The veteran Python scraping framework; mature ecosystem the AI knows inside out. -- [unclecode/crawl4ai](https://github.com/unclecode/crawl4ai) — A crawler designed for LLMs; outputs clean Markdown, born to be fed to AI. -- [mendableai/firecrawl](https://github.com/mendableai/firecrawl) — Crawls entire sites into LLM-ready Markdown/structured data, with an API. -- [apify/crawlee](https://github.com/apify/crawlee) — Node.js scraping and browser automation library with solid anti-blocking. -- [microsoft/playwright](https://github.com/microsoft/playwright) — The standard for browser automation; dynamic-page scraping and E2E testing alike. -- [browser-use/browser-use](https://github.com/browser-use/browser-use) — Lets AI agents drive a real browser to scrape and operate web pages. -- [D4Vinci/Scrapling](https://github.com/D4Vinci/Scrapling) — Adaptive anti-blocking Python scraper that survives site redesigns. -- [NanmiCoder/MediaCrawler](https://github.com/NanmiCoder/MediaCrawler) — Crawler suite for Chinese social media: Xiaohongshu, Douyin, Bilibili, Weibo. -- [NaiboWang/EasySpider](https://github.com/NaiboWang/EasySpider) — Visual no-code scraping; design collection flows in a GUI. -- [getmaxun/maxun](https://github.com/getmaxun/maxun) — Open-source no-code platform: train robots to turn websites into APIs/spreadsheets. -- [puppeteer/puppeteer](https://github.com/puppeteer/puppeteer) — Chrome's official headless browser automation library. -- [SeleniumHQ/selenium](https://github.com/SeleniumHQ/selenium) — The browser-automation veteran with the best cross-browser coverage. +- [scrapy/scrapy](https://github.com/scrapy/scrapy) `⭐63k · 2026-07-08` — The veteran Python scraping framework; mature ecosystem the AI knows inside out. +- [unclecode/crawl4ai](https://github.com/unclecode/crawl4ai) `⭐71.6k · 2026-07-08` — A crawler designed for LLMs; outputs clean Markdown, born to be fed to AI. +- [mendableai/firecrawl](https://github.com/mendableai/firecrawl) `⭐148.1k · 2026-07-09` — Crawls entire sites into LLM-ready Markdown/structured data, with an API. +- [apify/crawlee](https://github.com/apify/crawlee) `⭐24.6k · 2026-07-09` — Node.js scraping and browser automation library with solid anti-blocking. +- [microsoft/playwright](https://github.com/microsoft/playwright) `⭐92.5k · 2026-07-09` — The standard for browser automation; dynamic-page scraping and E2E testing alike. +- [browser-use/browser-use](https://github.com/browser-use/browser-use) `⭐103.9k · 2026-07-08` — Lets AI agents drive a real browser to scrape and operate web pages. +- [D4Vinci/Scrapling](https://github.com/D4Vinci/Scrapling) `⭐68.8k · 2026-07-08` — Adaptive anti-blocking Python scraper that survives site redesigns. +- [NanmiCoder/MediaCrawler](https://github.com/NanmiCoder/MediaCrawler) `⭐55.9k · 2026-07-09` — Crawler suite for Chinese social media: Xiaohongshu, Douyin, Bilibili, Weibo. +- [NaiboWang/EasySpider](https://github.com/NaiboWang/EasySpider) `⭐44.2k · 2026-07-03` — Visual no-code scraping; design collection flows in a GUI. +- [getmaxun/maxun](https://github.com/getmaxun/maxun) `⭐16.4k · 2026-07-08` — Open-source no-code platform: train robots to turn websites into APIs/spreadsheets. +- [puppeteer/puppeteer](https://github.com/puppeteer/puppeteer) `⭐95.3k · 2026-07-09` — Chrome's official headless browser automation library. +- [SeleniumHQ/selenium](https://github.com/SeleniumHQ/selenium) `⭐34.3k · 2026-07-09` — The browser-automation veteran with the best cross-browser coverage. ## 🎨 Frontend & UI The key to AI-generated interfaces that don't look terrible. -- [shadcn-ui/ui](https://github.com/shadcn-ui/ui) — Copy-paste component collection; the de facto standard for AI-generated React UIs. -- [tailwindlabs/tailwindcss](https://github.com/tailwindlabs/tailwindcss) — Utility-first CSS; the AI's first choice for styling, with the most training data. -- [saadeghi/daisyui](https://github.com/saadeghi/daisyui) — Tailwind component library; polished interfaces from pure class names. -- [ant-design/ant-design](https://github.com/ant-design/ant-design) — Enterprise React component library; admin dashboards ready to copy. -- [claude-code · frontend-design skill](https://github.com/anthropics/claude-code/blob/main/plugins/frontend-design/skills/frontend-design/SKILL.md) — Claude Code's official frontend-design skill; teaches the AI intentional, non-templated visual design. -- [pbakaus/impeccable](https://github.com/pbakaus/impeccable) — A design language to feed your AI, making your AI harness genuinely better at design. -- [Leonxlnx/taste-skill](https://github.com/Leonxlnx/taste-skill) — Gives your AI good taste; stops it from generating boring, generic slop. -- [VoltAgent/awesome-design-md](https://github.com/VoltAgent/awesome-design-md) — DESIGN.md files distilled from popular brand design systems; drop one in and let the agent generate a matching UI. -- [greensock/gsap-skills](https://github.com/greensock/gsap-skills) — Official AI skills for GSAP: best practices, common animation patterns, plugin usage. -- [JCodesMore/ai-website-cloner-template](https://github.com/JCodesMore/ai-website-cloner-template) — Clone any website with one command using AI coding agents. -- [diffusionstudio/lottie](https://github.com/diffusionstudio/lottie) — Generate production-ready Lottie animations with Claude Code or Codex. +- [shadcn-ui/ui](https://github.com/shadcn-ui/ui) `⭐118.5k · 2026-07-09` — Copy-paste component collection; the de facto standard for AI-generated React UIs. +- [tailwindlabs/tailwindcss](https://github.com/tailwindlabs/tailwindcss) `⭐95.8k · 2026-07-07` — Utility-first CSS; the AI's first choice for styling, with the most training data. +- [saadeghi/daisyui](https://github.com/saadeghi/daisyui) `⭐41.5k · 2026-07-08` — Tailwind component library; polished interfaces from pure class names. +- [ant-design/ant-design](https://github.com/ant-design/ant-design) `⭐98.6k · 2026-07-09` — Enterprise React component library; admin dashboards ready to copy. +- [claude-code · frontend-design skill](https://github.com/anthropics/claude-code/blob/main/plugins/frontend-design/skills/frontend-design/SKILL.md) `⭐137k · 2026-07-08` — Claude Code's official frontend-design skill; teaches the AI intentional, non-templated visual design. +- [pbakaus/impeccable](https://github.com/pbakaus/impeccable) `⭐44.9k · 2026-07-09` — A design language to feed your AI, making your AI harness genuinely better at design. +- [Leonxlnx/taste-skill](https://github.com/Leonxlnx/taste-skill) `⭐61k · 2026-07-04` — Gives your AI good taste; stops it from generating boring, generic slop. +- [VoltAgent/awesome-design-md](https://github.com/VoltAgent/awesome-design-md) `⭐98.9k · 2026-06-16` — DESIGN.md files distilled from popular brand design systems; drop one in and let the agent generate a matching UI. +- [greensock/gsap-skills](https://github.com/greensock/gsap-skills) `⭐11.2k · 2026-04-21` — Official AI skills for GSAP: best practices, common animation patterns, plugin usage. +- [JCodesMore/ai-website-cloner-template](https://github.com/JCodesMore/ai-website-cloner-template) `⭐27k · 2026-07-04` — Clone any website with one command using AI coding agents. +- [diffusionstudio/lottie](https://github.com/diffusionstudio/lottie) `⭐4.7k · 2026-07-06` — Generate production-ready Lottie animations with Claude Code or Codex. - [Web to Figma (Figma plugin)](https://www.figma.com/community/plugin/1297530151115228662/web-to-figma-convert-any-website-or-html-code-to-design) — Convert any website or HTML into a Figma design. ⚠️ Not an open-source repo; Figma community plugin, listed as an exception. ## 🏗️ Backend & Full-Stack Scaffolds Letting the AI modify a proven skeleton beats generating from zero. -- [fastapi/fastapi](https://github.com/fastapi/fastapi) — High-performance Python API framework; type-driven, so AI-generated code comes out cleaner. -- [nestjs/nest](https://github.com/nestjs/nest) — Enterprise Node.js backend framework with a clear structure that's easy for AI to follow. -- [supabase/supabase](https://github.com/supabase/supabase) — Open-source Firebase alternative: database + auth + storage in one. -- [vercel/next.js](https://github.com/vercel/next.js) — React full-stack framework; one of the frameworks AI knows best. -- [django/django](https://github.com/django/django) — Batteries-included Python web framework with admin/ORM; massive AI training corpus. -- [expressjs/express](https://github.com/expressjs/express) — The classic Node.js web framework; simple and direct. -- [gin-gonic/gin](https://github.com/gin-gonic/gin) — High-performance Go web framework; first pick for API services. -- [spring-projects/spring-boot](https://github.com/spring-projects/spring-boot) — The Java enterprise standard, ready out of the box. -- [honojs/hono](https://github.com/honojs/hono) — Ultra-light web framework running on Node/Deno/Bun/Cloudflare Workers alike. -- [nuxt/nuxt](https://github.com/nuxt/nuxt) — The Vue full-stack framework; Vue's answer to Next.js. -- [prisma/prisma](https://github.com/prisma/prisma) — TypeScript ORM; a type-safe database access layer. -- [pocketbase/pocketbase](https://github.com/pocketbase/pocketbase) — Single-file backend: database + auth + file storage in one binary. +- [fastapi/fastapi](https://github.com/fastapi/fastapi) `⭐100.3k · 2026-07-07` — High-performance Python API framework; type-driven, so AI-generated code comes out cleaner. +- [nestjs/nest](https://github.com/nestjs/nest) `⭐76.1k · 2026-07-09` — Enterprise Node.js backend framework with a clear structure that's easy for AI to follow. +- [supabase/supabase](https://github.com/supabase/supabase) `⭐105.9k · 2026-07-09` — Open-source Firebase alternative: database + auth + storage in one. +- [vercel/next.js](https://github.com/vercel/next.js) `⭐140.4k · 2026-07-09` — React full-stack framework; one of the frameworks AI knows best. +- [django/django](https://github.com/django/django) `⭐88k · 2026-07-08` — Batteries-included Python web framework with admin/ORM; massive AI training corpus. +- [expressjs/express](https://github.com/expressjs/express) `⭐69.3k · 2026-07-06` — The classic Node.js web framework; simple and direct. +- [gin-gonic/gin](https://github.com/gin-gonic/gin) `⭐88.9k · 2026-06-26` — High-performance Go web framework; first pick for API services. +- [spring-projects/spring-boot](https://github.com/spring-projects/spring-boot) `⭐81.1k · 2026-07-07` — The Java enterprise standard, ready out of the box. +- [honojs/hono](https://github.com/honojs/hono) `⭐31.3k · 2026-07-07` — Ultra-light web framework running on Node/Deno/Bun/Cloudflare Workers alike. +- [nuxt/nuxt](https://github.com/nuxt/nuxt) `⭐60.6k · 2026-07-04` — The Vue full-stack framework; Vue's answer to Next.js. +- [prisma/prisma](https://github.com/prisma/prisma) `⭐46.7k · 2026-07-08` — TypeScript ORM; a type-safe database access layer. +- [pocketbase/pocketbase](https://github.com/pocketbase/pocketbase) `⭐59.5k · 2026-07-08` — Single-file backend: database + auth + file storage in one binary. ## 🔌 MCP & Agent Ecosystem Infrastructure for giving your AI superpowers. -- [modelcontextprotocol/servers](https://github.com/modelcontextprotocol/servers) — Official MCP server collection: filesystem, databases, search and more, plug and play. -- [punkpeye/awesome-mcp-servers](https://github.com/punkpeye/awesome-mcp-servers) — The community MCP server directory; start here when hunting for add-ons. -- [langchain-ai/langgraph](https://github.com/langchain-ai/langgraph) — Framework for building stateful multi-agent applications. -- [microsoft/autogen](https://github.com/microsoft/autogen) — Microsoft's multi-agent conversation framework. -- [microsoft/playwright-mcp](https://github.com/microsoft/playwright-mcp) — Official Playwright MCP; let the AI drive a browser directly. -- [github/github-mcp-server](https://github.com/github/github-mcp-server) — GitHub's official MCP; let the AI manage repos, issues, PRs. -- [PrefectHQ/fastmcp](https://github.com/PrefectHQ/fastmcp) — The fast Python framework for building MCP servers and clients. -- [crewAIInc/crewAI](https://github.com/crewAIInc/crewAI) — Role-based multi-agent collaboration framework. -- [mem0ai/mem0](https://github.com/mem0ai/mem0) — A long-term memory layer for AI agents. -- [browserbase/stagehand](https://github.com/browserbase/stagehand) — AI browser automation blending natural language and code control. -- [openai/openai-agents-python](https://github.com/openai/openai-agents-python) — OpenAI's official multi-agent orchestration SDK. -- [google/adk-python](https://github.com/google/adk-python) — Google's Agent Development Kit (ADK). +- [modelcontextprotocol/servers](https://github.com/modelcontextprotocol/servers) `⭐88.3k · 2026-07-06` — Official MCP server collection: filesystem, databases, search and more, plug and play. +- [punkpeye/awesome-mcp-servers](https://github.com/punkpeye/awesome-mcp-servers) `⭐90.5k · 2026-07-04` — The community MCP server directory; start here when hunting for add-ons. +- [langchain-ai/langgraph](https://github.com/langchain-ai/langgraph) `⭐36.9k · 2026-07-09` — Framework for building stateful multi-agent applications. +- [microsoft/autogen](https://github.com/microsoft/autogen) `⭐59.6k · 2026-04-15` — Microsoft's multi-agent conversation framework. +- [microsoft/playwright-mcp](https://github.com/microsoft/playwright-mcp) `⭐34.9k · 2026-06-29` — Official Playwright MCP; let the AI drive a browser directly. +- [github/github-mcp-server](https://github.com/github/github-mcp-server) `⭐31.3k · 2026-07-08` — GitHub's official MCP; let the AI manage repos, issues, PRs. +- [PrefectHQ/fastmcp](https://github.com/PrefectHQ/fastmcp) `⭐26.1k · 2026-07-09` — The fast Python framework for building MCP servers and clients. +- [crewAIInc/crewAI](https://github.com/crewAIInc/crewAI) `⭐55.2k · 2026-07-09` — Role-based multi-agent collaboration framework. +- [mem0ai/mem0](https://github.com/mem0ai/mem0) `⭐60.5k · 2026-07-09` — A long-term memory layer for AI agents. +- [browserbase/stagehand](https://github.com/browserbase/stagehand) `⭐23.4k · 2026-07-09` — AI browser automation blending natural language and code control. +- [openai/openai-agents-python](https://github.com/openai/openai-agents-python) `⭐27.8k · 2026-07-09` — OpenAI's official multi-agent orchestration SDK. +- [google/adk-python](https://github.com/google/adk-python) `⭐20.5k · 2026-07-09` — Google's Agent Development Kit (ADK). --- diff --git a/README.md b/README.md index 0c2dd74..6d1507f 100644 --- a/README.md +++ b/README.md @@ -23,7 +23,7 @@ - 🔄 **最近 6 个月内有提交** - 📄 **开源协议明确** -标准由 [`check.py`](check.py) 自动校验，GitHub Actions 每周体检一次，不达标的会被移除。 +标准由 [`check.py`](check.py) 自动校验，GitHub Actions 每周体检一次，不达标的会被移除。每个条目的 `⭐ star 数 · 最后更新日期` 也由 CI 每周自动刷新。 --- @@ -31,113 +31,113 @@ 直接帮你写代码的 Agent / 助手本体。 -- [anthropics/claude-code](https://github.com/anthropics/claude-code) — 终端里的 AI 编程 Agent，理解整个代码库，执行多步任务。 -- [cline/cline](https://github.com/cline/cline) — VS Code 里的自主编程 Agent，能创建/编辑文件、跑命令、用浏览器。 -- [All-Hands-AI/OpenHands](https://github.com/All-Hands-AI/OpenHands) — 开源 AI 软件工程师平台，能完成完整开发任务。 -- [Aider-AI/aider](https://github.com/Aider-AI/aider) — 终端结对编程工具，直接在本地 git 仓库上改代码。 -- [continuedev/continue](https://github.com/continuedev/continue) — 开源 IDE 助手，可自定义模型和上下文源。 -- [RooCodeInc/Roo-Code](https://github.com/RooCodeInc/Roo-Code) — Cline 的社区增强分支，多模式 Agent 团队。 -- [openai/codex](https://github.com/openai/codex) — OpenAI 的终端编程 Agent，Rust 实现，轻量快速。 -- [google-gemini/gemini-cli](https://github.com/google-gemini/gemini-cli) — Google 开源的终端 AI Agent，直连 Gemini。 -- [anomalyco/opencode](https://github.com/anomalyco/opencode) — 开源终端 AI 编程 Agent，模型随便换。 -- [aaif-goose/goose](https://github.com/aaif-goose/goose) — 本地运行的开源 AI Agent，可装扩展自动化工程任务。 -- [TabbyML/tabby](https://github.com/TabbyML/tabby) — 自托管 AI 代码补全，Copilot 的开源替代。 -- [Kilo-Org/kilocode](https://github.com/Kilo-Org/kilocode) — VS Code 开源 AI Agent，集 Cline/Roo 之长。 +- [anthropics/claude-code](https://github.com/anthropics/claude-code) `⭐137k · 2026-07-08` — 终端里的 AI 编程 Agent，理解整个代码库，执行多步任务。 +- [cline/cline](https://github.com/cline/cline) `⭐64.5k · 2026-07-09` — VS Code 里的自主编程 Agent，能创建/编辑文件、跑命令、用浏览器。 +- [All-Hands-AI/OpenHands](https://github.com/All-Hands-AI/OpenHands) `⭐80.1k · 2026-07-09` — 开源 AI 软件工程师平台，能完成完整开发任务。 +- [Aider-AI/aider](https://github.com/Aider-AI/aider) `⭐47.2k · 2026-05-22` — 终端结对编程工具，直接在本地 git 仓库上改代码。 +- [continuedev/continue](https://github.com/continuedev/continue) `⭐34.8k · 2026-07-09` — 开源 IDE 助手，可自定义模型和上下文源。 +- [RooCodeInc/Roo-Code](https://github.com/RooCodeInc/Roo-Code) `⭐24.3k · 2026-05-15` — Cline 的社区增强分支，多模式 Agent 团队。 +- [openai/codex](https://github.com/openai/codex) `⭐96.5k · 2026-07-09` — OpenAI 的终端编程 Agent，Rust 实现，轻量快速。 +- [google-gemini/gemini-cli](https://github.com/google-gemini/gemini-cli) `⭐105.9k · 2026-07-09` — Google 开源的终端 AI Agent，直连 Gemini。 +- [anomalyco/opencode](https://github.com/anomalyco/opencode) `⭐184k · 2026-07-09` — 开源终端 AI 编程 Agent，模型随便换。 +- [aaif-goose/goose](https://github.com/aaif-goose/goose) `⭐50.9k · 2026-07-09` — 本地运行的开源 AI Agent，可装扩展自动化工程任务。 +- [TabbyML/tabby](https://github.com/TabbyML/tabby) `⭐33.7k · 2026-06-30` — 自托管 AI 代码补全，Copilot 的开源替代。 +- [Kilo-Org/kilocode](https://github.com/Kilo-Org/kilocode) `⭐25.9k · 2026-07-09` — VS Code 开源 AI Agent，集 Cline/Roo 之长。 ## 📝 提示词与规则 给 AI 定规矩、给灵感的资源库。 -- [PatrickJS/awesome-cursorrules](https://github.com/PatrickJS/awesome-cursorrules) — 各种技术栈的 .cursorrules 规则合集，拿来即用。 -- [f/awesome-chatgpt-prompts](https://github.com/f/awesome-chatgpt-prompts) — 经典提示词大全，写系统提示词的参考。 -- [x1xhlol/system-prompts-and-models-of-ai-tools](https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools) — 各大 AI 编程工具的系统提示词逆向合集，学习工具怎么"调教"模型。 -- [anthropics/skills](https://github.com/anthropics/skills) — Anthropic 官方 Agent Skills 合集，给 Claude 装领域技能。 -- [obra/superpowers](https://github.com/obra/superpowers) — 编程 Agent 技能全家桶：头脑风暴、TDD、系统化调试等工作流。 -- [dair-ai/Prompt-Engineering-Guide](https://github.com/dair-ai/Prompt-Engineering-Guide) — 提示词工程系统教程，论文、技巧、案例俱全。 -- [anthropics/prompt-eng-interactive-tutorial](https://github.com/anthropics/prompt-eng-interactive-tutorial) — Anthropic 官方交互式提示词教程。 -- [NirDiamant/Prompt_Engineering](https://github.com/NirDiamant/Prompt_Engineering) — 提示词技术实战教程集，从基础到高级都带代码。 -- [ai-boost/awesome-prompts](https://github.com/ai-boost/awesome-prompts) — 精选高分 GPTs 提示词与提示词工程资料。 +- [PatrickJS/awesome-cursorrules](https://github.com/PatrickJS/awesome-cursorrules) `⭐40.3k · 2026-05-30` — 各种技术栈的 .cursorrules 规则合集，拿来即用。 +- [f/awesome-chatgpt-prompts](https://github.com/f/awesome-chatgpt-prompts) `⭐165.1k · 2026-07-09` — 经典提示词大全，写系统提示词的参考。 +- [x1xhlol/system-prompts-and-models-of-ai-tools](https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools) `⭐141.7k · 2026-07-08` — 各大 AI 编程工具的系统提示词逆向合集，学习工具怎么"调教"模型。 +- [anthropics/skills](https://github.com/anthropics/skills) `⭐159.7k · 2026-07-01` — Anthropic 官方 Agent Skills 合集，给 Claude 装领域技能。 +- [obra/superpowers](https://github.com/obra/superpowers) `⭐250.4k · 2026-07-06` — 编程 Agent 技能全家桶：头脑风暴、TDD、系统化调试等工作流。 +- [dair-ai/Prompt-Engineering-Guide](https://github.com/dair-ai/Prompt-Engineering-Guide) `⭐76.3k · 2026-03-11` — 提示词工程系统教程，论文、技巧、案例俱全。 +- [anthropics/prompt-eng-interactive-tutorial](https://github.com/anthropics/prompt-eng-interactive-tutorial) `⭐36.9k · 2026-03-01` — Anthropic 官方交互式提示词教程。 +- [NirDiamant/Prompt_Engineering](https://github.com/NirDiamant/Prompt_Engineering) `⭐7.7k · 2026-07-04` — 提示词技术实战教程集，从基础到高级都带代码。 +- [ai-boost/awesome-prompts](https://github.com/ai-boost/awesome-prompts) `⭐8.4k · 2026-07-09` — 精选高分 GPTs 提示词与提示词工程资料。 ## 📚 让 AI 读懂代码与文档 把仓库、网页、文档转成 AI 能吃的上下文。 -- [coderamp-labs/gitingest](https://github.com/coderamp-labs/gitingest) — 把任意 Git 仓库一键变成适合喂给 LLM 的纯文本摘要。 -- [upstash/context7](https://github.com/upstash/context7) — MCP 服务器，让 AI 实时获取库的最新官方文档和示例代码。 -- [AsyncFuncAI/deepwiki-open](https://github.com/AsyncFuncAI/deepwiki-open) — 给任意仓库自动生成 Wiki 文档，帮 AI 和人快速理解项目。 -- [yamadashy/repomix](https://github.com/yamadashy/repomix) — 把整个代码库打包成一份 AI 友好的文件，gitingest 的热门同类。 -- [microsoft/markitdown](https://github.com/microsoft/markitdown) — 微软出品，Office/PDF 等各种文件转 Markdown 喂 LLM。 -- [docling-project/docling](https://github.com/docling-project/docling) — IBM 开源文档解析器，PDF/表格/版面理解，输出结构化数据。 -- [opendatalab/MinerU](https://github.com/opendatalab/MinerU) — 高质量 PDF 转 Markdown/JSON，公式表格都能抽。 -- [datalab-to/marker](https://github.com/datalab-to/marker) — PDF 快速精准转 Markdown，公式表格支持好。 -- [jina-ai/reader](https://github.com/jina-ai/reader) — 任意 URL 前加 r.jina.ai 即得 LLM 友好的正文文本。 -- [Unstructured-IO/unstructured](https://github.com/Unstructured-IO/unstructured) — 各种格式文档的 ETL 预处理库，RAG 数据管道常客。 +- [coderamp-labs/gitingest](https://github.com/coderamp-labs/gitingest) `⭐15k · 2026-07-08` — 把任意 Git 仓库一键变成适合喂给 LLM 的纯文本摘要。 +- [upstash/context7](https://github.com/upstash/context7) `⭐58.8k · 2026-07-09` — MCP 服务器，让 AI 实时获取库的最新官方文档和示例代码。 +- [AsyncFuncAI/deepwiki-open](https://github.com/AsyncFuncAI/deepwiki-open) `⭐17.2k · 2026-06-03` — 给任意仓库自动生成 Wiki 文档，帮 AI 和人快速理解项目。 +- [yamadashy/repomix](https://github.com/yamadashy/repomix) `⭐27k · 2026-07-05` — 把整个代码库打包成一份 AI 友好的文件，gitingest 的热门同类。 +- [microsoft/markitdown](https://github.com/microsoft/markitdown) `⭐164.2k · 2026-06-24` — 微软出品，Office/PDF 等各种文件转 Markdown 喂 LLM。 +- [docling-project/docling](https://github.com/docling-project/docling) `⭐62.9k · 2026-07-09` — IBM 开源文档解析器，PDF/表格/版面理解，输出结构化数据。 +- [opendatalab/MinerU](https://github.com/opendatalab/MinerU) `⭐74k · 2026-07-09` — 高质量 PDF 转 Markdown/JSON，公式表格都能抽。 +- [datalab-to/marker](https://github.com/datalab-to/marker) `⭐37.3k · 2026-07-07` — PDF 快速精准转 Markdown，公式表格支持好。 +- [jina-ai/reader](https://github.com/jina-ai/reader) `⭐11.5k · 2026-05-22` — 任意 URL 前加 r.jina.ai 即得 LLM 友好的正文文本。 +- [Unstructured-IO/unstructured](https://github.com/Unstructured-IO/unstructured) `⭐15.1k · 2026-07-09` — 各种格式文档的 ETL 预处理库，RAG 数据管道常客。 ## 🕷️ 爬虫与数据采集 -- [scrapy/scrapy](https://github.com/scrapy/scrapy) — Python 爬虫框架老大哥，生态成熟，AI 对它的用法非常熟。 -- [unclecode/crawl4ai](https://github.com/unclecode/crawl4ai) — 为 LLM 设计的爬虫，输出干净的 Markdown，天生适合喂 AI。 -- [mendableai/firecrawl](https://github.com/mendableai/firecrawl) — 把整站抓成 LLM-ready 的 Markdown/结构化数据，带 API。 -- [apify/crawlee](https://github.com/apify/crawlee) — Node.js 爬虫与浏览器自动化库，反爬处理完善。 -- [microsoft/playwright](https://github.com/microsoft/playwright) — 浏览器自动化标准件，动态页面抓取和 E2E 测试都靠它。 -- [browser-use/browser-use](https://github.com/browser-use/browser-use) — 让 AI Agent 直接操控浏览器抓取和操作网页。 -- [D4Vinci/Scrapling](https://github.com/D4Vinci/Scrapling) — 自适应反爬的 Python 爬虫库，网站改版也不易失效。 -- [NanmiCoder/MediaCrawler](https://github.com/NanmiCoder/MediaCrawler) — 小红书/抖音/B站/微博等中文社媒爬虫合集。 -- [NaiboWang/EasySpider](https://github.com/NaiboWang/EasySpider) — 可视化无代码爬虫，图形界面设计采集流程。 -- [getmaxun/maxun](https://github.com/getmaxun/maxun) — 开源无代码平台，训练机器人把网页变成 API/表格。 -- [puppeteer/puppeteer](https://github.com/puppeteer/puppeteer) — Chrome 官方无头浏览器自动化库。 -- [SeleniumHQ/selenium](https://github.com/SeleniumHQ/selenium) — 浏览器自动化元老，跨浏览器兼容性最好。 +- [scrapy/scrapy](https://github.com/scrapy/scrapy) `⭐63k · 2026-07-08` — Python 爬虫框架老大哥，生态成熟，AI 对它的用法非常熟。 +- [unclecode/crawl4ai](https://github.com/unclecode/crawl4ai) `⭐71.6k · 2026-07-08` — 为 LLM 设计的爬虫，输出干净的 Markdown，天生适合喂 AI。 +- [mendableai/firecrawl](https://github.com/mendableai/firecrawl) `⭐148.1k · 2026-07-09` — 把整站抓成 LLM-ready 的 Markdown/结构化数据，带 API。 +- [apify/crawlee](https://github.com/apify/crawlee) `⭐24.6k · 2026-07-09` — Node.js 爬虫与浏览器自动化库，反爬处理完善。 +- [microsoft/playwright](https://github.com/microsoft/playwright) `⭐92.5k · 2026-07-09` — 浏览器自动化标准件，动态页面抓取和 E2E 测试都靠它。 +- [browser-use/browser-use](https://github.com/browser-use/browser-use) `⭐103.9k · 2026-07-08` — 让 AI Agent 直接操控浏览器抓取和操作网页。 +- [D4Vinci/Scrapling](https://github.com/D4Vinci/Scrapling) `⭐68.8k · 2026-07-08` — 自适应反爬的 Python 爬虫库，网站改版也不易失效。 +- [NanmiCoder/MediaCrawler](https://github.com/NanmiCoder/MediaCrawler) `⭐55.9k · 2026-07-09` — 小红书/抖音/B站/微博等中文社媒爬虫合集。 +- [NaiboWang/EasySpider](https://github.com/NaiboWang/EasySpider) `⭐44.2k · 2026-07-03` — 可视化无代码爬虫，图形界面设计采集流程。 +- [getmaxun/maxun](https://github.com/getmaxun/maxun) `⭐16.4k · 2026-07-08` — 开源无代码平台，训练机器人把网页变成 API/表格。 +- [puppeteer/puppeteer](https://github.com/puppeteer/puppeteer) `⭐95.3k · 2026-07-09` — Chrome 官方无头浏览器自动化库。 +- [SeleniumHQ/selenium](https://github.com/SeleniumHQ/selenium) `⭐34.3k · 2026-07-09` — 浏览器自动化元老，跨浏览器兼容性最好。 ## 🎨 前端与 UI 让 AI 生成的界面不丑的关键。 -- [shadcn-ui/ui](https://github.com/shadcn-ui/ui) — 复制即用的组件集，AI 生成 React 界面的事实标准。 -- [tailwindlabs/tailwindcss](https://github.com/tailwindlabs/tailwindcss) — 原子化 CSS，AI 写样式的首选，训练语料最多。 -- [saadeghi/daisyui](https://github.com/saadeghi/daisyui) — Tailwind 组件库，纯 class 就能出成品感界面。 -- [ant-design/ant-design](https://github.com/ant-design/ant-design) — 企业级 React 组件库，中后台界面直接抄。 -- [claude-code · frontend-design 技能](https://github.com/anthropics/claude-code/blob/main/plugins/frontend-design/skills/frontend-design/SKILL.md) — Claude Code 官方前端设计技能，教 AI 做有设计感、不落模板俗套的界面。 -- [pbakaus/impeccable](https://github.com/pbakaus/impeccable) — 一套喂给 AI 的设计语言，让 AI 助手真正懂设计。 -- [Leonxlnx/taste-skill](https://github.com/Leonxlnx/taste-skill) — 给 AI 装上"审美品味"，不再生成千篇一律的平庸界面。 -- [VoltAgent/awesome-design-md](https://github.com/VoltAgent/awesome-design-md) — 各大品牌设计系统提炼成的 DESIGN.md 合集，丢进项目让 AI 生成风格统一的 UI。 -- [greensock/gsap-skills](https://github.com/greensock/gsap-skills) — GSAP 官方 AI 技能，教 AI 正确写动画：最佳实践、常用模式、插件用法。 -- [JCodesMore/ai-website-cloner-template](https://github.com/JCodesMore/ai-website-cloner-template) — 一条命令让 AI 克隆任意网站的模板。 -- [diffusionstudio/lottie](https://github.com/diffusionstudio/lottie) — 用 Claude Code / Codex 生成可直接上线的 Lottie 动画。 +- [shadcn-ui/ui](https://github.com/shadcn-ui/ui) `⭐118.5k · 2026-07-09` — 复制即用的组件集，AI 生成 React 界面的事实标准。 +- [tailwindlabs/tailwindcss](https://github.com/tailwindlabs/tailwindcss) `⭐95.8k · 2026-07-07` — 原子化 CSS，AI 写样式的首选，训练语料最多。 +- [saadeghi/daisyui](https://github.com/saadeghi/daisyui) `⭐41.5k · 2026-07-08` — Tailwind 组件库，纯 class 就能出成品感界面。 +- [ant-design/ant-design](https://github.com/ant-design/ant-design) `⭐98.6k · 2026-07-09` — 企业级 React 组件库，中后台界面直接抄。 +- [claude-code · frontend-design 技能](https://github.com/anthropics/claude-code/blob/main/plugins/frontend-design/skills/frontend-design/SKILL.md) `⭐137k · 2026-07-08` — Claude Code 官方前端设计技能，教 AI 做有设计感、不落模板俗套的界面。 +- [pbakaus/impeccable](https://github.com/pbakaus/impeccable) `⭐44.9k · 2026-07-09` — 一套喂给 AI 的设计语言，让 AI 助手真正懂设计。 +- [Leonxlnx/taste-skill](https://github.com/Leonxlnx/taste-skill) `⭐61k · 2026-07-04` — 给 AI 装上"审美品味"，不再生成千篇一律的平庸界面。 +- [VoltAgent/awesome-design-md](https://github.com/VoltAgent/awesome-design-md) `⭐98.9k · 2026-06-16` — 各大品牌设计系统提炼成的 DESIGN.md 合集，丢进项目让 AI 生成风格统一的 UI。 +- [greensock/gsap-skills](https://github.com/greensock/gsap-skills) `⭐11.2k · 2026-04-21` — GSAP 官方 AI 技能，教 AI 正确写动画：最佳实践、常用模式、插件用法。 +- [JCodesMore/ai-website-cloner-template](https://github.com/JCodesMore/ai-website-cloner-template) `⭐27k · 2026-07-04` — 一条命令让 AI 克隆任意网站的模板。 +- [diffusionstudio/lottie](https://github.com/diffusionstudio/lottie) `⭐4.7k · 2026-07-06` — 用 Claude Code / Codex 生成可直接上线的 Lottie 动画。 - [Web to Figma（Figma 插件）](https://www.figma.com/community/plugin/1297530151115228662/web-to-figma-convert-any-website-or-html-code-to-design) — 把任意网站或 HTML 代码转成 Figma 设计稿。⚠️ 非开源仓库，Figma 社区插件，破例收录。 ## 🏗️ 后端与全栈脚手架 在成熟骨架上让 AI 改，比从零生成靠谱得多。 -- [fastapi/fastapi](https://github.com/fastapi/fastapi) — Python 高性能 API 框架，类型驱动，AI 生成的代码质量高。 -- [nestjs/nest](https://github.com/nestjs/nest) — Node.js 企业级后端框架，结构清晰适合 AI 遵循。 -- [supabase/supabase](https://github.com/supabase/supabase) — 开源 Firebase 替代品，数据库+认证+存储一条龙。 -- [vercel/next.js](https://github.com/vercel/next.js) — React 全栈框架，AI 最熟悉的前端框架之一。 -- [django/django](https://github.com/django/django) — Python 全家桶式 Web 框架，自带 Admin/ORM，AI 语料极多。 -- [expressjs/express](https://github.com/expressjs/express) — Node.js 最经典的 Web 框架，简单直接。 -- [gin-gonic/gin](https://github.com/gin-gonic/gin) — Go 高性能 Web 框架，写 API 服务的首选。 -- [spring-projects/spring-boot](https://github.com/spring-projects/spring-boot) — Java 企业级标准，开箱即用。 -- [honojs/hono](https://github.com/honojs/hono) — 超轻量 Web 框架，Node/Deno/Bun/Cloudflare Workers 通吃。 -- [nuxt/nuxt](https://github.com/nuxt/nuxt) — Vue 全栈框架，Vue 版的 Next.js。 -- [prisma/prisma](https://github.com/prisma/prisma) — TypeScript ORM，类型安全的数据库访问层。 -- [pocketbase/pocketbase](https://github.com/pocketbase/pocketbase) — 单文件后端：数据库+认证+文件存储，一个二进制跑起来。 +- [fastapi/fastapi](https://github.com/fastapi/fastapi) `⭐100.3k · 2026-07-07` — Python 高性能 API 框架，类型驱动，AI 生成的代码质量高。 +- [nestjs/nest](https://github.com/nestjs/nest) `⭐76.1k · 2026-07-09` — Node.js 企业级后端框架，结构清晰适合 AI 遵循。 +- [supabase/supabase](https://github.com/supabase/supabase) `⭐105.9k · 2026-07-09` — 开源 Firebase 替代品，数据库+认证+存储一条龙。 +- [vercel/next.js](https://github.com/vercel/next.js) `⭐140.4k · 2026-07-09` — React 全栈框架，AI 最熟悉的前端框架之一。 +- [django/django](https://github.com/django/django) `⭐88k · 2026-07-08` — Python 全家桶式 Web 框架，自带 Admin/ORM，AI 语料极多。 +- [expressjs/express](https://github.com/expressjs/express) `⭐69.3k · 2026-07-06` — Node.js 最经典的 Web 框架，简单直接。 +- [gin-gonic/gin](https://github.com/gin-gonic/gin) `⭐88.9k · 2026-06-26` — Go 高性能 Web 框架，写 API 服务的首选。 +- [spring-projects/spring-boot](https://github.com/spring-projects/spring-boot) `⭐81.1k · 2026-07-07` — Java 企业级标准，开箱即用。 +- [honojs/hono](https://github.com/honojs/hono) `⭐31.3k · 2026-07-07` — 超轻量 Web 框架，Node/Deno/Bun/Cloudflare Workers 通吃。 +- [nuxt/nuxt](https://github.com/nuxt/nuxt) `⭐60.6k · 2026-07-04` — Vue 全栈框架，Vue 版的 Next.js。 +- [prisma/prisma](https://github.com/prisma/prisma) `⭐46.7k · 2026-07-08` — TypeScript ORM，类型安全的数据库访问层。 +- [pocketbase/pocketbase](https://github.com/pocketbase/pocketbase) `⭐59.5k · 2026-07-08` — 单文件后端：数据库+认证+文件存储，一个二进制跑起来。 ## 🔌 MCP 与 Agent 生态 给 AI 装外挂的基础设施。 -- [modelcontextprotocol/servers](https://github.com/modelcontextprotocol/servers) — MCP 官方服务器合集，文件系统、数据库、搜索等能力即插即用。 -- [punkpeye/awesome-mcp-servers](https://github.com/punkpeye/awesome-mcp-servers) — 社区 MCP 服务器大全，找外挂先来这。 -- [langchain-ai/langgraph](https://github.com/langchain-ai/langgraph) — 构建有状态多 Agent 应用的框架。 -- [microsoft/autogen](https://github.com/microsoft/autogen) — 微软的多 Agent 对话框架。 -- [microsoft/playwright-mcp](https://github.com/microsoft/playwright-mcp) — 官方 Playwright MCP，让 AI 直接操控浏览器。 -- [github/github-mcp-server](https://github.com/github/github-mcp-server) — GitHub 官方 MCP，让 AI 管理仓库/issue/PR。 -- [PrefectHQ/fastmcp](https://github.com/PrefectHQ/fastmcp) — 用 Python 快速构建 MCP 服务器和客户端的框架。 -- [crewAIInc/crewAI](https://github.com/crewAIInc/crewAI) — 角色分工的多 Agent 协作框架。 -- [mem0ai/mem0](https://github.com/mem0ai/mem0) — 给 AI Agent 加一层长期记忆。 -- [browserbase/stagehand](https://github.com/browserbase/stagehand) — AI 浏览器自动化框架，自然语言与代码混合控制。 -- [openai/openai-agents-python](https://github.com/openai/openai-agents-python) — OpenAI 官方多 Agent 编排 SDK。 -- [google/adk-python](https://github.com/google/adk-python) — Google 的 Agent 开发套件（ADK）。 +- [modelcontextprotocol/servers](https://github.com/modelcontextprotocol/servers) `⭐88.3k · 2026-07-06` — MCP 官方服务器合集，文件系统、数据库、搜索等能力即插即用。 +- [punkpeye/awesome-mcp-servers](https://github.com/punkpeye/awesome-mcp-servers) `⭐90.5k · 2026-07-04` — 社区 MCP 服务器大全，找外挂先来这。 +- [langchain-ai/langgraph](https://github.com/langchain-ai/langgraph) `⭐36.9k · 2026-07-09` — 构建有状态多 Agent 应用的框架。 +- [microsoft/autogen](https://github.com/microsoft/autogen) `⭐59.6k · 2026-04-15` — 微软的多 Agent 对话框架。 +- [microsoft/playwright-mcp](https://github.com/microsoft/playwright-mcp) `⭐34.9k · 2026-06-29` — 官方 Playwright MCP，让 AI 直接操控浏览器。 +- [github/github-mcp-server](https://github.com/github/github-mcp-server) `⭐31.3k · 2026-07-08` — GitHub 官方 MCP，让 AI 管理仓库/issue/PR。 +- [PrefectHQ/fastmcp](https://github.com/PrefectHQ/fastmcp) `⭐26.1k · 2026-07-09` — 用 Python 快速构建 MCP 服务器和客户端的框架。 +- [crewAIInc/crewAI](https://github.com/crewAIInc/crewAI) `⭐55.2k · 2026-07-09` — 角色分工的多 Agent 协作框架。 +- [mem0ai/mem0](https://github.com/mem0ai/mem0) `⭐60.5k · 2026-07-09` — 给 AI Agent 加一层长期记忆。 +- [browserbase/stagehand](https://github.com/browserbase/stagehand) `⭐23.4k · 2026-07-09` — AI 浏览器自动化框架，自然语言与代码混合控制。 +- [openai/openai-agents-python](https://github.com/openai/openai-agents-python) `⭐27.8k · 2026-07-09` — OpenAI 官方多 Agent 编排 SDK。 +- [google/adk-python](https://github.com/google/adk-python) `⭐20.5k · 2026-07-09` — Google 的 Agent 开发套件（ADK）。 --- diff --git a/check.py b/check.py index 6915b0b..fefd15f 100644 --- a/check.py +++ b/check.py @@ -1,7 +1,9 @@ #!/usr/bin/env python3 """体检 README.md 里收录的 GitHub 仓库：星数 >= 1000 且最近 6 个月内有提交。 -用法: GITHUB_TOKEN=xxx python3 check.py (token 可选，无 token 限流 60 次/小时) +用法: GITHUB_TOKEN=xxx python3 check.py [--update] + --update 把 star 数和最后更新日期写回 README 条目（CI 每周自动执行） + (token 可选，无 token 限流 60 次/小时) """ import json import os @@ -14,6 +16,9 @@ MIN_STARS = 1000 MAX_IDLE_DAYS = 183 SELF = "patrick-eu/ARMORY" +FILES = ("README.md", "README.en.md") +# 只匹配收录条目（以 "- [" 开头的 GitHub 链接），可选捕获已有的 `⭐…` 标注以便替换 +LINK = re.compile(r"(- \[[^\]]+\]\(https://github\.com/([\w.-]+/[\w.-]+)[^)]*\))( `⭐[^`]*`)?") def listed_repos(readme: str) -> list[str]: @@ -33,15 +38,34 @@ def fetch(repo: str) -> dict: return json.load(resp) +def fmt_stars(n: int) -> str: + return f"{n / 1000:.1f}k".replace(".0k", "k") if n >= 1000 else str(n) + + +def annotate(path: str, stats: dict) -> None: + with open(path, encoding="utf-8") as f: + src = f.read() + + def sub(m: re.Match) -> str: + if info := stats.get(m.group(2)): + return f"{m.group(1)} `⭐{fmt_stars(info[0])} · {info[1]}`" + return m.group(1) + + out = LINK.sub(sub, src) + if out != src: + with open(path, "w", encoding="utf-8") as f: + f.write(out) + + def main() -> int: root = os.path.dirname(__file__) or "." text = "".join( - open(os.path.join(root, name), encoding="utf-8").read() - for name in ("README.md", "README.en.md") + open(os.path.join(root, name), encoding="utf-8").read() for name in FILES ) repos = listed_repos(text) now = datetime.now(timezone.utc) failures = [] + stats = {} for repo in repos: try: data = fetch(repo) @@ -51,6 +75,7 @@ def main() -> int: stars = data["stargazers_count"] pushed = datetime.fromisoformat(data["pushed_at"].replace("Z", "+00:00")) idle = (now - pushed).days + stats[repo] = (stars, pushed.date().isoformat()) problems = [] if stars < MIN_STARS: problems.append(f"仅 {stars} stars") @@ -61,6 +86,10 @@ def main() -> int: if problems: failures.append(f"{repo}: {status}") print(f"\n共 {len(repos)} 个仓库，{len(failures)} 个不达标") + if "--update" in sys.argv: + for name in FILES: + annotate(os.path.join(root, name), stats) + print("已把 star 数与更新日期写回 README") if failures: print("\n".join(failures), file=sys.stderr) return 1