Report Date: 2026-03-09
Analysis Window: 30 days
Languages Tracked: python, typescript, rust, go, javascript
Top N: 50
Overview
根据截至2026年3月9日的30天数据显示，GitHub AI开源领域的趋势明显由“Agent”类项目主导，无论在项目数量还是近期热度上都占据绝对优势。在技术栈方面，TypeScript和Python是开发者首选的编程语言，Rust也显示出强劲的势头。
Top 50 Repositories
| Rank | Repository | Category | Language | Heat_7 | Heat_30 | Score |
|---|---|---|---|---|---|---|
| 1 | openclaw | typescript | 0 | 106,172 | 3,115 | |
| 2 | shannon | agent | typescript | 6,900 | 21,399 | 367 |
| 3 | pi-mono | agent | typescript | 0 | 13,206 | 44 |
| 4 | langextract | llm | python | 0 | 10,002 | 33 |
| 5 | learn-claude-code | agent | typescript | 3,901 | 7,115 | 530 |
| 6 | PageIndex | rag | python | 0 | 6,730 | 22 |
| 7 | OpenSandbox | agent | python | 3,959 | 5,821 | 189 |
| 8 | dexter | agent | typescript | 0 | 5,134 | 17 |
| 9 | chrome-devtools-mcp | agent | typescript | 0 | 4,669 | 15 |
| 10 | codex | agent | rust | 1,437 | 4,637 | 195 |
| 11 | superset | agent | typescript | 3,132 | 4,374 | 148 |
| 12 | AstrBot | agent | python | 0 | 4,182 | 13 |
| 13 | OpenViking | agent | python | 0 | 3,991 | 13 |
| 14 | monty | rust | 0 | 3,860 | 12 | |
| 15 | gh-aw | agent | go | 0 | 3,547 | 11 |
| 16 | litellm | llm | python | 0 | 2,821 | 9 |
| 17 | ruvector | rust | 753 | 2,682 | 41 | |
| 18 | claudian | llm | typescript | 468 | 1,385 | 24 |
| 19 | agents | agent | typescript | 0 | 1,360 | 4 |
| 20 | SillyTavern | llm | javascript | 366 | 1,245 | 47 |
| 21 | qdrant | rust | 0 | 780 | 2 | |
| 22 | agent-deck | llm | go | 235 | 697 | 12 |
| 23 | bifrost | go | 0 | 673 | 2 | |
| 24 | ai-dev-kit | agent | python | 0 | 636 | 2 |
| 25 | meetily | rust | 0 | 565 | 1 | |
| 26 | rig | llm | rust | 0 | 485 | 1 |
| 27 | skills | agent | javascript | 126 | 463 | 6 |
| 28 | screenpipe | rust | 0 | 452 | 1 | |
| 29 | google_workspace_mcp | python | 0 | 414 | 1 | |
| 30 | pi-skills | agent | javascript | 0 | 372 | 1 |
| 31 | page-agent | agent | typescript | 0 | 0 | 429 |
| 32 | skills | agent | python | 7,152 | 0 | 306 |
| 33 | impeccable | javascript | 0 | 0 | 265 | |
| 34 | Qwen-Agent | agent | python | 1,735 | 0 | 228 |
| 35 | GitNexus | agent | typescript | 3,952 | 0 | 169 |
| 36 | ai-hedge-fund | python | 0 | 0 | 165 | |
| 37 | agent-skills | agent | javascript | 707 | 0 | 145 |
| 38 | ruflo | agent | typescript | 2,922 | 0 | 125 |
| 39 | beads | agent | go | 0 | 0 | 100 |
| 40 | worktrunk | agent | rust | 591 | 0 | 59 |
| 41 | webnovel-writer | llm | python | 0 | 0 | 56 |
| 42 | agentscope | agent | python | 1,163 | 0 | 49 |
| 43 | Flowise | agent | typescript | 1,105 | 0 | 47 |
| 44 | AReaL | agent | python | 969 | 0 | 41 |
| 45 | openai-cua-sample-app | agent | typescript | 0 | 0 | 41 |
| 46 | claude-hud | agent | javascript | 327 | 0 | 41 |
| 47 | hermes-agent | agent | python | 941 | 0 | 40 |
| 48 | goose | agent | rust | 842 | 0 | 36 |
| 49 | anything-llm | javascript | 715 | 0 | 30 | |
| 50 | LMCache | llm | python | 632 | 0 | 27 |
Category Breakdown
(11 projects)
未分类项目（11个）同样值得关注，其中包含了像openclaw（个人AI助手）和qdrant（向量数据库）等基础工具，这些项目跨越了单一的技术标签。
Average Heat_7: 133
Average Score: 330
- openclaw - Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞
- monty - A minimal, secure Python interpreter written in Rust for use by AI
- ruvector - RuVector is a High Performance, Real-Time, Self-Learning, Vector Graph Neural Network, and Database built in Rust.
- qdrant - Qdrant - High-performance, massive-scale Vector Database and Vector Search Engine for the next generation of AI. Also available in the cloud https://cloud.qdrant.io/
- bifrost - Fastest enterprise AI gateway (50x faster than LiteLLM) with adaptive load balancer, cluster mode, guardrails, 1000+ models support & <100 µs overhead at 5k RPS.
- meetily - Privacy first, AI meeting assistant with 4x faster Parakeet/Whisper live transcription, speaker diarization, and Ollama summarization built on Rust. 100% local processing. no cloud required. Meetil...
- screenpipe - screenpipe turns your computer into a personal AI that knows everything you've done. record. search. automate. all local, all private, all yours.
- google_workspace_mcp - Control Gmail, Google Calendar, Docs, Sheets, Slides, Chat, Forms, Tasks, Search & Drive with AI - Comprehensive Google Workspace / G Suite MCP Server & CLI Tool
- impeccable - The design language that makes your AI harness better at design.
- ai-hedge-fund - An AI Hedge Fund Team
- anything-llm - The all-in-one AI productivity accelerator. On device and privacy first with no annoying setup or configration.
agent (30 projects)
Agent是当前最热门的赛道，占据了榜单中30个席位，其平均热度（1395.37）远超其他类别，表明构建自主代理、编码助手和Agent框架是当前开源AI创新的核心。
Average Heat_7: 1,395
Average Score: 113
- shannon - Shannon Lite is a fully autonomous AI pentester for web apps and APIs. 96.15% (100/104 exploits) on a hint-free, source-aware variant of the XBOW benchmark.
- pi-mono - AI agent toolkit: coding agent CLI, unified LLM API, TUI & web UI libraries, Slack bot, vLLM pods
- learn-claude-code - Bash is all you need - A nano Claude Code–like agent, built from 0 to 1
- OpenSandbox - OpenSandbox is a general-purpose sandbox platform for AI applications, offering multi-language SDKs, unified sandbox APIs, and Docker/Kubernetes runtimes for scenarios like Coding Agents, GUI Agent...
- dexter - An autonomous agent for deep financial research
- chrome-devtools-mcp - Chrome DevTools for coding agents
- codex - Lightweight coding agent that runs in your terminal
- superset - IDE for the AI Agents Era - Run an army of Claude Code, Codex, etc. on your machine
- AstrBot - Agentic IM Chatbot infrastructure that integrates lots of IM platforms, LLMs, plugins and AI feature, and can be your openclaw alternative. ✨
- OpenViking - OpenViking is an open-source context database designed specifically for AI Agents(such as openclaw). OpenViking unifies the management of context (memory, resources, and skills) that Agents need th...
- gh-aw - GitHub Agentic Workflows
- agents - Build and deploy AI Agents on Cloudflare
- ai-dev-kit - Databricks Toolkit for Coding Agents provided by Field Engineering
- skills - A collection of AI agent skills for working with Expo projects and Expo Application Services
- pi-skills - Skills for pi coding agent (compatible with Claude Code and Codex CLI)
- page-agent - JavaScript in-page GUI agent. Control web interfaces with natural language.
- skills - Public repository for Agent Skills
- Qwen-Agent - Agent framework and applications built upon Qwen>=3.0, featuring Function Calling, MCP, Code Interpreter, RAG, Chrome extension, etc.
- GitNexus - GitNexus: The Zero-Server Code Intelligence Engine - GitNexus is a client-side knowledge graph creator that runs entirely in your browser. Drop in a GitHub repo or ZIP file, and get an interactive ...
- agent-skills - Collection of Apify Agent Skills
- ruflo - 🌊 The leading agent orchestration platform for Claude. Deploy intelligent multi-agent swarms, coordinate autonomous workflows, and build conversational AI systems. Features enterprise-grade arch...
- beads - Beads - A memory upgrade for your coding agent
- worktrunk - Worktrunk is a CLI for Git worktree management, designed for parallel AI agent workflows
- agentscope - Build and run agents you can see, understand and trust.
- Flowise - Build AI Agents, Visually
- AReaL - Lightning-Fast RL for LLM Reasoning and Agents. Made Simple & Flexible.
- openai-cua-sample-app - Learn how to use CUA (our Computer Using Agent) via the API on multiple computer environments.
- claude-hud - A Claude Code plugin that shows what's happening - context usage, active tools, running agents, and todo progress
- hermes-agent - The agent that grows with you
- goose - an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM
llm (8 projects)
LLM领域虽然仍有8个项目上榜，但热度（212.6）远低于Agent，这可能反映出开发重点正从基础模型转向更高层次的应用框架和工具。
Average Heat_7: 212
Average Score: 26
- langextract - A Python library for extracting structured information from unstructured text using LLMs with precise source grounding and interactive visualization.
- litellm - Python SDK, Proxy Server (AI Gateway) to call 100+ LLM APIs in OpenAI (or native) format, with cost tracking, guardrails, loadbalancing and logging. [Bedrock, Azure, OpenAI, VertexAI, Cohere, Anthr...
- claudian - An Obsidian plugin that embeds Claude Code as an AI collaborator in your vault
- SillyTavern - LLM Frontend for Power Users.
- agent-deck - Terminal session manager for AI coding agents. One TUI for Claude, Gemini, OpenCode, Codex, and more.
- rig - ⚙️🦀 Build modular and scalable LLM Applications in Rust
- webnovel-writer - 基于 Claude Code 的长篇网文辅助创 作系统，解决 AI 写作中的「遗忘」和「幻觉」问题，支持 200 万字量级 连载创作。
- LMCache - Supercharge Your LLM with the Fastest KV Cache Layer
rag (1 projects)
RAG类别仅有一个项目，且近期热度为零，这表明RAG可能正从一个独立的趋势领域，演变为大型Agent或LLM框架中的一个集成功能。
Average Heat_7: 0
Average Score: 22
- PageIndex - 📑 PageIndex: Document Index for Vectorless, Reasoning-based RAG
Dark Horse Projects
“黑马项目”榜单揭示了一批近期获得极高关注度（score）的新兴项目。其中既有阿里巴巴（OpenSandbox）和Anthropic（skills）等巨头的作品，也有独立的创新项目，显示出Agent工具链的快速发展和激烈竞争。这些项目热度的突发性（例如anthropics/skills近期热度极高）表明该领域创新迭代速度非常快。
| Repository | Score | Heat_30 | Heat_7 | Category |
|---|---|---|---|---|
| openclaw | 3,115 | 106,172 | 0 | |
| shannon | 367 | 21,399 | 6,900 | agent |
| learn-claude-code | 530 | 7,115 | 3,901 | agent |
| OpenSandbox | 189 | 5,821 | 3,959 | agent |
| codex | 195 | 4,637 | 1,437 | agent |
| superset | 148 | 4,374 | 3,132 | agent |
| page-agent | 429 | 0 | 0 | agent |
| skills | 306 | 0 | 7,152 | agent |
| impeccable | 265 | 0 | 0 | |
| Qwen-Agent | 228 | 0 | 1,735 | agent |
| GitNexus | 169 | 0 | 3,952 | agent |
| ai-hedge-fund | 165 | 0 | 0 | |
| agent-skills | 145 | 0 | 707 | agent |
| ruflo | 125 | 0 | 2,922 | agent |
| beads | 100 | 0 | 0 | agent |
Consecutive Appearances
“常青树项目”列表则反映了市场的持续兴趣。像openclaw/openclaw和KeygraphHQ/shannon等项目连续7周上榜，说明它们已不仅仅是短期热点，而是建立了稳固的用户和社区基础。这些项目的稳定性对于生态系统中的其他开发者至关重要。
| Repository | Weeks in Top | Category | Current Heat_7 |
|---|---|---|---|
| openclaw | 7 | 0 | |
| shannon | 7 | agent | 6,900 |
| pi-mono | 7 | agent | 0 |
| langextract | 5 | llm | 0 |
| PageIndex | 7 | rag | 0 |
| OpenSandbox | 2 | agent | 3,959 |
| dexter | 7 | agent | 0 |
| chrome-devtools-mcp | 4 | agent | 0 |
| codex | 7 | agent | 1,437 |
| superset | 3 | agent | 3,132 |
| AstrBot | 3 | agent | 0 |
| monty | 2 | 0 | |
| litellm | 3 | llm | 0 |
| ruvector | 2 | 753 | |
| claudian | 2 | llm | 468 |
| agents | 3 | agent | 0 |
| agent-deck | 3 | llm | 235 |
| meetily | 2 | 0 | |
| rig | 3 | llm | 0 |
| skills | 3 | agent | 126 |
| screenpipe | 4 | 0 | |
| google_workspace_mcp | 2 | 0 | |
| pi-skills | 2 | agent | 0 |
| skills | 7 | agent | 7,152 |
| GitNexus | 3 | agent | 3,952 |
| ruflo | 2 | agent | 2,922 |
| beads | 2 | agent | 0 |
| anything-llm | 2 | 715 |
Highlighted Repositories
openclaw/openclaw
该项目以极高的分数（3115）和30日热度（106172）位居榜首，定位为个人AI助手，代表了Agent技术的一个重要应用方向。尽管近7日热度为零，其巨大的存量关注度使其成为本周期内无可争议的领导者。
KeygraphHQ/shannon
作为排名第二的Agent项目，shannon专注于自主AI网络渗透测试，这是一个高度专业化且价值明确的应用场景。其持续的高热度（6900）表明市场对特定领域Agent解决方案的强烈需求。
alibaba/OpenSandbox
由阿里巴巴推出的OpenSandbox解决了AI Agent应用中的一个关键痛点：安全执行环境。作为一个通用的沙盒平台，它为编码、评估和训练等场景提供了基础支持，其高涨的近期热度（3959）反映了该基础设施的重要性。
anthropics/skills
尽管总排名不高，但该项目由核心AI公司Anthropic发布，且近7日热度（7152）极高，这通常预示着一个重要的新技术或框架的发布。它专注于Agent技能，是未来最值得关注的项目之一。
pydantic/monty
此项目是一个用Rust编写的、用于AI的极简安全Python解释器。它代表了AI领域的一个重要子趋势：为支持AI应用而构建高性能、安全的基础设施，尤其是在AI Agent需要执行代码的场景下。
Methodology
Data Sources:
- GitHub Trending pages
- GitHub API stargazer timestamps
Metrics:
- Heat_7: Stars gained in last 7 days
- Heat_30: Stars gained in last 30 days
- Score: Weighted scoring combining short-term heat and sustained growth
- Formula: 0.6 × stars_1d + 0.3 × (stars_7d / 7) + 0.1 × (stars_30d / 30)
- Emphasizes recent activity (60%) while considering sustained trends (40%)
Filtering:
- Include keywords: %!s(int=30)
- Exclude keywords: %!s(int=30)
- Categories: %!s(int=30)
Ranking:
- Sort by Heat_30 (descending)
- Tie-break by Score (descending)
- Select top %!d(string=ai, llm, agent, gpt, rag, langchain, machine-learning, deep-learning, neural, transformer, embedding, vector)
Limitations:
- Trending data limited to GitHub's trending algorithm
- Star history may be incomplete for repos with >40k stars (API pagination limits)
- LLM-generated commentary is interpretive, not prescriptive
- Weekly snapshots may miss short-lived trends%!(EXTRA string=tutorial, awesome, learning, course, book, string=agent, llm, rag, vision, infra, int=50)
Generated at: 2026-03-09T00:28:00Z