|
63 | 63 | | AI-Scientist | 端到端自动科学发现系统,覆盖想法生成、实验、论文和同行评审 | agent | <!--stars:SakanaAI/AI-Scientist-->⭐ 14.4k<!--/stars--> | [GitHub](https://github.com/SakanaAI/AI-Scientist) | - | [Nature 2024](https://modelscope.cn/papers/2408.06292/) | |
64 | 64 | | AI-Scientist-v2 | 基于 agentic tree search 的自动科研系统,不依赖人类模板,面向更通用的 ML 研究探索 | agent | <!--stars:SakanaAI/AI-Scientist-v2-->⭐ 7k<!--/stars--> | [GitHub](https://github.com/SakanaAI/AI-Scientist-v2) | - | [arXiv 2025](https://modelscope.cn/papers/2504.08066/) | |
65 | 65 | | EvoScientist | 多 agent AI scientist 系统,强调持久记忆、技能演化和端到端科研协作 | agent | <!--stars:EvoScientist/EvoScientist-->⭐ 4.5k<!--/stars--> | [GitHub](https://github.com/EvoScientist/EvoScientist) | [Demo](https://evoscientist.ai) | [arXiv 2025](https://modelscope.cn/papers/2603.08127/) | |
66 | | -| nature-skills | 符合 nature 论文学术表达和科研绘图的 Skill | skill | <!--stars:Yuan1z0825/nature-skills-->⭐ 36.5k<!--/stars--> | [GitHub](https://github.com/Yuan1z0825/nature-skills) · [魔搭 Skills](https://modelscope.cn/collections/stn54999/nature-skills) | - | - | |
| 66 | +| nature-skills | 符合 nature 论文学术表达和科研绘图的 Skill | skill | <!--stars:Yuan1z0825/nature-skills-->⭐ 36.7k<!--/stars--> | [GitHub](https://github.com/Yuan1z0825/nature-skills) · [魔搭 Skills](https://modelscope.cn/collections/stn54999/nature-skills) | - | - | |
67 | 67 | | AutoResearchClaw | 自主、自进化的多阶段研究流水线,从研究想法推进到论文产物 | agent | <!--stars:aiming-lab/AutoResearchClaw-->⭐ 14.1k<!--/stars--> | [GitHub](https://github.com/aiming-lab/AutoResearchClaw) | [Demo](https://openclaw.ai) | [arXiv 2025](https://modelscope.cn/papers/2605.22662/) | |
68 | 68 | | autoresearch | Andrej Karpathy 的自主 ML 研究代理,在单 GPU 上自动运行实验并改进模型 | agent | <!--stars:karpathy/autoresearch-->⭐ 94.5k<!--/stars--> | [GitHub](https://github.com/karpathy/autoresearch) | - | - | |
69 | | -| Auto Claude Code Research in Sleep | 自动化 Claude Code 科研工作流项目,面向自动实验与代码执行 | skill | <!--stars:wanshuiyin/Auto-claude-code-research-in-sleep-->⭐ 15k<!--/stars--> | [GitHub](https://github.com/wanshuiyin/Auto-claude-code-research-in-sleep) | - | - | |
| 69 | +| Auto Claude Code Research in Sleep | 自动化 Claude Code 科研工作流项目,面向自动实验与代码执行 | skill | <!--stars:wanshuiyin/Auto-claude-code-research-in-sleep-->⭐ 15.1k<!--/stars--> | [GitHub](https://github.com/wanshuiyin/Auto-claude-code-research-in-sleep) | - | - | |
70 | 70 | | AgentLaboratory | 端到端的自主研究工作流程,由 LLM 驱动的专业代理支持完成从文献综述到报告撰写的全流程 | agent | <!--stars:SamuelSchmidgall/AgentLaboratory-->⭐ 5.8k<!--/stars--> | [GitHub](https://github.com/SamuelSchmidgall/AgentLaboratory) | - | - | |
71 | 71 | | Aether | 基于 OpenCode 的开源项目,面向科研人员提供 web 与桌面端统一的 AI 研究工作环境 | agent/应用 | <!--stars:Science-Discovery/Aether-->⭐ 71<!--/stars--> | [GitHub](https://github.com/Science-Discovery/Aether) | - | - | |
72 | 72 | | EurekAgent | 环境工程驱动的自主科研系统,面向可度量任务协调 Claude Code 会话提出方案、实现代码、隔离评测并迭代优化 | agent | <!--stars:THU-Team-Eureka/EurekAgent-->⭐ 80<!--/stars--> | [GitHub](https://github.com/THU-Team-Eureka/EurekAgent) | - | [arXiv 2026](https://arxiv.org/abs/2606.13662) | |
73 | 73 | | InternAgent | 面向长程自主科学发现的统一 agent 框架,支持假设生成、自动实验执行、论文复现、记忆模块和 Deep Research | agent | <!--stars:InternScience/InternAgent-->⭐ 1.4k<!--/stars--> | [GitHub](https://github.com/InternScience/InternAgent) | [Website](https://discovery.intern-ai.org.cn/home) | [arXiv 2026](https://arxiv.org/abs/2602.08990) | |
74 | | -| Tashan Research Skills | 国科大他山团队自研的 16 个科研 skills,按文献证据、研究构思、成果表达、协作沉淀、工具测评五类整理,每个技能自带脚本、模板和测试 | skill | <!--stars:TashanGKD/tashan-research-skills-->⭐ 14<!--/stars--> | [GitHub](https://github.com/TashanGKD/tashan-research-skills) | - | - | |
| 74 | +| Tashan Research Skills | 国科大他山团队自研的 16 个科研 skills,按文献证据、研究构思、成果表达、协作沉淀、工具测评五类整理,每个技能自带脚本、模板和测试 | skill | <!--stars:TashanGKD/tashan-research-skills-->⭐ 15<!--/stars--> | [GitHub](https://github.com/TashanGKD/tashan-research-skills) | - | - | |
75 | 75 |
|
76 | 76 | --- |
77 | 77 |
|
|
81 | 81 |
|
82 | 82 | | 项目名称 | 描述 | 类型 | Stars | 链接 | Demo | Paper | |
83 | 83 | |---|---|---|---|---|---|---| |
84 | | -| SciAgentsDiscovery | MIT 开源的科学发现多智能体系统,结合知识图谱与 agent 协作用于跨学科假说生成 | agent | <!--stars:lamm-mit/SciAgentsDiscovery-->⭐ 633<!--/stars--> | [GitHub](https://github.com/lamm-mit/SciAgentsDiscovery) | - | [arXiv 2024](https://arxiv.org/abs/2409.05556) | |
85 | | -| AutoDiscovery | AllenAI 开源的开放式科学发现框架,通过 Bayesian Surprise 从数据中发现可验证假说 | benchmark/workflow | <!--stars:allenai/autodiscovery-neurips-->⭐ 202<!--/stars--> | [GitHub](https://github.com/allenai/autodiscovery-neurips) | - | [NeurIPS 2024](https://arxiv.org/abs/2406.13266) | |
| 84 | +| SciAgentsDiscovery | MIT 开源的科学发现多智能体系统,结合知识图谱与 agent 协作用于跨学科假说生成 | agent | <!--stars:lamm-mit/SciAgentsDiscovery-->⭐ 634<!--/stars--> | [GitHub](https://github.com/lamm-mit/SciAgentsDiscovery) | - | [arXiv 2024](https://arxiv.org/abs/2409.05556) | |
| 85 | +| AutoDiscovery | AllenAI 开源的开放式科学发现框架,通过 Bayesian Surprise 从数据中发现可验证假说 | benchmark/workflow | <!--stars:allenai/autodiscovery-neurips-->⭐ 203<!--/stars--> | [GitHub](https://github.com/allenai/autodiscovery-neurips) | - | [NeurIPS 2024](https://arxiv.org/abs/2406.13266) | |
86 | 86 |
|
87 | 87 | --- |
88 | 88 |
|
|
110 | 110 |
|
111 | 111 | | 项目名称 | 描述 | 类型 | Stars | 链接 | Demo | Paper | |
112 | 112 | |---|---|---|---|---|---|---| |
113 | | -| Curie | 自动化且强调严谨性的科学实验 agent,从假说澄清到实验执行、分析和报告 | agent/workflow | <!--stars:Just-Curieous/Curie-->⭐ 369<!--/stars--> | [GitHub](https://github.com/Just-Curieous/Curie) | - | [arXiv 2025](https://arxiv.org/abs/2502.16069) | |
| 113 | +| Curie | 自动化且强调严谨性的科学实验 agent,从假说澄清到实验执行、分析和报告 | agent/workflow | <!--stars:Just-Curieous/Curie-->⭐ 370<!--/stars--> | [GitHub](https://github.com/Just-Curieous/Curie) | - | [arXiv 2025](https://arxiv.org/abs/2502.16069) | |
114 | 114 |
|
115 | 115 | --- |
116 | 116 |
|
|
142 | 142 |
|
143 | 143 | | 项目名称 | 描述 | 类型 | Stars | 链接 | Demo | Paper | |
144 | 144 | |---|---|---|---|---|---|---| |
145 | | -| Academic Research Skills | 覆盖学术写作、润色、投稿检查和发表流程的 Claude Code skill 套件,也覆盖文献调研 | skill | <!--stars:Imbad0202/academic-research-skills-->⭐ 43.4k<!--/stars--> | [GitHub](https://github.com/Imbad0202/academic-research-skills) | - | - | |
146 | | -| RefChecker | 学术参考文献验证工具,可检查引用真实性、元数据错误和疑似伪造引用 | tool | <!--stars:markrussinovich/refchecker-->⭐ 472<!--/stars--> | [GitHub](https://github.com/markrussinovich/refchecker) | - | - | |
| 145 | +| Academic Research Skills | 覆盖学术写作、润色、投稿检查和发表流程的 Claude Code skill 套件,也覆盖文献调研 | skill | <!--stars:Imbad0202/academic-research-skills-->⭐ 43.5k<!--/stars--> | [GitHub](https://github.com/Imbad0202/academic-research-skills) | - | - | |
| 146 | +| RefChecker | 学术参考文献验证工具,可检查引用真实性、元数据错误和疑似伪造引用 | tool | <!--stars:markrussinovich/refchecker-->⭐ 474<!--/stars--> | [GitHub](https://github.com/markrussinovich/refchecker) | - | - | |
147 | 147 | | Research Paper Lifecycle Skills | 面向 Agent 的论文全生命周期 Skill 套件,覆盖文献综述、引用核验、投稿检查、审稿回复、artifact、幻灯片和海报 | skill | <!--stars:ShaishavMaisuria/research-paper-lifecycle-skills-->⭐ 34<!--/stars--> | [GitHub](https://github.com/ShaishavMaisuria/research-paper-lifecycle-skills) | [Website](https://shaishavmaisuria.github.io/research-paper-lifecycle-skills/) | - | |
148 | 148 |
|
149 | 149 | --- |
|
0 commit comments