面向 long horizon agents 的增量引擎 🌟 喜欢的话请点 Star!Incremental engine for long horizon agents 🌟 Star if you like it!
趋势榜
实时增长榜 · 周综述每周日由 spark 生成
按日期回看
2026-09-06 新收录仓库 37 个
面向 Model Context Protocol (MCP) servers 的社区驱动 registry 服务。A community driven registry service for Model Context Protocol (MCP) servers.
Observal 是一个用于 coding agent 扩展的自托管 registry,内置 insight engine。安装 Observal、定义作用域,即可与同伴共享你的 Skills、MCPs 和 Agents。Observal is self-hosted registry for your coding agent extensions with a built in insight engine. Setup Observal, define the scope and share your Skills, MCPs and Agents with your peers.
由 AI 驱动的虚拟高管团队 —— 单一连贯的高管 persona,由 8 个 Claude 专家 agent 支撑(FastAPI + Next.js)。AI-powered virtual executive team — a single coherent executive persona backed by 8 specialist Claude agents (FastAPI + Next.js).
一款将 Claude Code、Cursor、Codex 与 Antigravity 的用量上限固定显示在屏幕边缘的 macOS 应用。A macOS app that pins usage limits from Claude Code, Cursor, Codex, and Antigravity to a screen edge.
:keyboard: 精选的超越 LaTeX 的优秀工具、演示与资源列表:keyboard: A curated list of awesome tools, demos and resources to go beyond LaTeX
自托管 AI 工作站:自主屏幕 Agent、三层神经路由、并行 Agent 集群、视频生成、4K/8K 升频、RAG、语音接口、60+ 工具执行引擎——全部在你的硬件上本地运行。The self-hosted AI workstation. Autonomous screen agents, 3-tier neural routing, parallel agent swarms, video generation, 4K/8K upscaling, RAG, voice interface, 60+ tool execution engine — all running locally on your hardware.
邮件客户端与服务商网关,让你打造最佳的邮件体验。An email client & provider gateway so you can build your best email experience.
[COLM 2025] 论文 Know Me, Respond to Me: Benchmarking LLMs for Dynamic User Profiling and Personalized Responses at Scale——大语言模型在动态用户画像与大规模个性化响应任务上的基准测试。[COLM 2025] Know Me, Respond to Me: Benchmarking LLMs for Dynamic User Profiling and Personalized Responses at Scale
在 ComfyUI 中运行 MiniMax-H3 VAE 的 ONNX/TRT 版本,速度最高提升至 1.7 倍。Running ONNX/TRT version of the MiniMax-H3 VAE in ComfyUI, which increase speed by up to 1.7x
[ICLR/AAAI/KDD/EMNLP2026] 基于开源 LLM 的数据分析 Agents[ICLR/AAAI/KDD/EMNLP2026] Open-Source LLM-Based Data Analysis Agents
面向 agentic 界面的 React design system。提供每个免费 BoardUI 组件的源代码,并以使用你自己 model key 的可运行 AI chat 应用作为主页。React design system for agentic interfaces. Every free BoardUI component as source, with a working AI chat app on your own model key as the homepage.
面向 Hyprland(基于 Omarchy)的按工作区平铺比例方案。拖动分隔条即可移动窗口 —— 25/50/25、master and stack、thirds、grid。支持命名布局、按工作区分配与可切换的 profile。Per-workspace tiling ratios for Hyprland, on Omarchy. Drag the dividers and your windows move — 25/50/25, master and stack, thirds, grid. Named layouts, per-workspace assignment, switchable profiles.
基于 DeepSeek Harness 构建的 Agent 原生交易终端:三栏 GUI 同时呈现美股、A股、港股与加密资产,配备 19+ 可热插拔连接器,默认 dry-run 并对每笔实盘订单要求人工审批。BYOK,不进行数据再分发。Agent-native trading terminal built on DeepSeek Harness. Crypto, US, CN and HK in one three-column GUI, 19+ hot-swappable connectors, dry-run by default with human approval on every live order. BYOK, no data redistribution.
TRX Drainer Tool——面向 TRON 钱包的 TRX 与 TRC20 token 自动 drainer,具备实时余额监控、多钱包批量转出、TronLink/Trust Wallet 兼容以及隐身的能量优化交易路由。TRX Drainer Tool — Automated TRX and TRC20 token drainer for TRON wallets with real-time balance monitoring, multi-wallet sweeping, TronLink/Trust Wallet compatibility, and stealth energy-optimized transaction routing
Linux 2000:适用于 X11 的类 Windows 2000 桌面 —— 窗口管理器、Explorer、applets,以及 XP 和 Windows 7 主题。与 Microsoft 无关。Linux 2000: a Windows 2000-like desktop for X11 -- window manager, Explorer, applets, XP and Windows 7 themes. Not affiliated with Microsoft.
高性能视频生成网关与会话协调器。High-performance video generation gateway and session coordinator
为 Claude Code 提供专用 agents、MCP servers、skills 和可视化 dashboard。MIT 许可。Specialized agents, MCP servers, skills, and a visual dashboard for Claude Code. MIT.
通过 Ollama 使用本地 LLM 自动生成文献综述:检索学术数据库、获取全文论文并撰写 APA 第 7 版格式综述;模型在结构上被禁止编造引文,每条引文均逐字核验并按 APA 第 7 版排版。Automated literature reviews using local LLMs via Ollama. Searches academic databases, acquires full-text papers, and writes APA 7th reviews where the model is structurally prevented from fabricating quotes — every quote is verified verbatim and is formated in APA7th.
Top Paper Retriever —— 仅检索顶刊的学术论文搜索(CCF A/B · 中科院 1/2 区 · flagship journals),提供 library、CLI 和 MCP server 三种形态。Top Paper Retriever — top-only academic paper search (CCF A/B · 中科院 1/2 区 · flagship journals) as a library, CLI and MCP server
AI Agent 的精确一次执行层——自托管、原生 MCP、永久免费且无限制。开源核心(社区版采用 MIT/Apache-2.0)。The exactly-once execution layer for AI agents — self-hosted, MCP-native, unlimited & free forever. Open-core (Community MIT/Apache-2.0).
面向 Zotero 的可解释论文推荐插件,支持可编辑的兴趣画像、AI 摘要与一键导入。An explainable paper recommendation plugin for Zotero, with editable interest profiles, AI summaries, and one-click imports.
由 Jujutsu 驱动的本地优先 Overleaf 协作工具。Local-first Overleaf collaboration, powered by Jujutsu
PaperClaw 是面向公共管理学者的 AI 研究助手,可帮助用户检索相关文献、定位文本关键概念、获取全文证据、回答研究问题并提出潜在想法,覆盖从检索到洞察的全流程。PaperClaw is an AI research assistant for public administration scholars. It helps users find relevant literature, locate key concepts in texts, retrieve full-text evidence, answer research questions, and suggest potential ideas, improving efficiency from search to insight.
Agentic 多 Agent AI 系统,通过 RAG 检索、来源质量评分,以及 planning → research → verification → decision 流水线,将研究问题转化为有证据支持、可验证的决策。Agentic multi-agent AI system that turns research questions into evidence-backed, verifiable decisions using RAG retrieval, source-quality scoring, and a planning → research → verification → decision pipeline.
使用 RevMan 9.16 和 Rayyan 评估糖尿病相关 COVID-19 死亡率的 15 项研究(6800 万+ 患者)的系统综述与 meta 分析。PROSPERO: CRD42024498934Systematic review and meta-analysis of 15 studies (68M+ patients) evaluating diabetes-associated COVID-19 mortality using RevMan 9.16 and Rayyan. PROSPERO: CRD42024498934
面向 swarmstatus.com 的开放性研究问题与证据综述。Open research questions and evidence reviews for swarmstatus.com
旨在辅助学术与科研任务的 AI 系统:能自主检索文献、总结论文、整理参考文献,并借助自然语言处理理解研究问题、检索相关信息。AI system designed to assist with academic and scientific research tasks. It can autonomously search for literature, summarize papers, and organize references. Using natural language processing, it understands research questions and retrieves relevant information
面向生命科学研究的系统综述、证据综合与 meta 分析平台。Systematic review, evidence-synthesis and meta-analysis platform for life-sciences research.
统计学学生,专注于可复现研究与数据项目。Statistics student focused on reproducible research and data projects.
有证据支撑的学术写作与修订技能:贡献优先的规划、引文支持、手稿审校,以及 LaTeX/Word 交付。Evidence-backed academic writing and revision skill: contribution-first planning, citation support, manuscript audits, and LaTeX/Word delivery.
2026-09-06 增长榜定格 当日周增 Top 20
| 名次 | 名称 | 方向 | 周增 | Stars |
|---|---|---|---|---|
| 1 — | odysseus-dev/odysseus | Agent 智能体 | +152,621 | 84,138 |
| 2 — | lnkiai/m3e-canvas | 工程化 | +11,186 | 3,871 |
| 3 — | anywhere-labs/deepseek-harness-desktop | 评测基准 | +8,281 | 13,495 |
| 4 — | stablyai/orca | Agent 智能体 | +5,299 | 47,228 |
| 5 — | s1dashu/ip-as-logo-skill | Agent 智能体 | +4,646 | 3,727 |
| 6 — | MiniMax-AI/MiniMax-H3 | LLM 基础设施 | +4,445 | 4,611 |
| 7 — | diegosouzapw/OmniRoute | Agent 智能体 | +4,343 | 53,672 |
| 8 — | holaboss-ai/holaOS | Agent 智能体 | +3,798 | 9,436 |
| 9 — | safishamsi/graphify | 数据与向量库 | +3,752 | 76,856 |
| 10 — | NousResearch/hermes-agent | Agent 智能体 | +3,386 | 230,735 |
| 11 — | zhu1090093659/dsh-web-ui | 评测基准 | +3,318 | 4,541 |
| 12 — | firecrawl/firecrawl | RAG 检索增强 | +2,975 | 165,271 |
| 13 — | firecrawl/anydoc | 工程化 | +2,870 | 13,720 |
| 14 — | MadsLorentzen/ai-job-search | 评测基准 | +2,740 | 40,537 |
| 15 — | koala73/worldmonitor | Agent 智能体 | +2,239 | 82,478 |
| 16 — | yjh051108/dsh-routing-suite | LLM 基础设施 | +2,149 | 7,031 |
| 17 — | hoainho/img2threejs | 多模态 | +2,051 | 4,176 |
| 18 — | DietrichGebert/ponytail | Agent 智能体 | +1,869 | 100,293 |
| 19 — | anthropics/skills | Agent 智能体 | +1,722 | 167,649 |
| 20 — | addyosmani/agent-skills | Agent 智能体 | +1,631 | 85,866 |