研究
把 AI 编码工具变成研究助手的 Agent Skills:网络抓取、学术文献综述、实证分析、带来源的综合归纳。
14 个技能
Browserbase's official collection of agent skills to access the web.
3,732Sep 29, 2026🔬 A curated collection of 23,000+ agent skills for empirical research across 8 social science disciplines. | 精选 23,000+ AI Agent 技能库,覆盖8大社会科学学科的实证研究。CoPaper.AI 20分钟完成一篇可复现的规范实证论文,并支持用户上传 Skills。-- Maintained by CoPaper.AI from Stanford REAP.
4,460Sep 30, 2026🔬 A curated collection of 23,000+ agent skills for empirical research across 8 social science disciplines. | 精选 23,000+ AI Agent 技能库,覆盖8大社会科学学科的实证研究。CoPaper.AI 20分钟完成一篇可复现的规范实证论文,并支持用户上传 Skills。-- Maintained by CoPaper.AI from Stanford REAP.
Star: 4,460
最近更新: Sep 30, 2026
本项目围绕吴恩达老师在DeepLearning.AI出品的agent-skills-with-anthropic系列课程,为学习者打造中文翻译与知识整理教程。项目提供课程内容翻译、知识点梳理和示例代码解读等内容,欢迎大家Star!
1,531Aug 27, 2026本项目围绕吴恩达老师在DeepLearning.AI出品的agent-skills-with-anthropic系列课程,为学习者打造中文翻译与知识整理教程。项目提供课程内容翻译、知识点梳理和示例代码解读等内容,欢迎大家Star!
Star: 1,531
最近更新: Aug 27, 2026
8 Claude skills · 100+ recon capabilities · 80 secret-regex patterns · 80+ dorks · 9 read-only credential validators · 27 attack-path templates · ~10,000 lines of structured tradecraft. Drop-in SKILL.md files that turn Claude into a god-mode external recon operator for authorized red-team and bug-bounty engagements.
2,718Aug 30, 20268 Claude skills · 100+ recon capabilities · 80 secret-regex patterns · 80+ dorks · 9 read-only credential validators · 27 attack-path templates · ~10,000 lines of structured tradecraft. Drop-in SKILL.md files that turn Claude into a god-mode external recon operator for authorized red-team and bug-bounty engagements.
Star: 2,718
最近更新: Aug 30, 2026
Codex skill: aggressive web research via OpenAI-compatible Grok endpoint (2api)
181Mar 13, 2026Codex skill: aggressive web research via OpenAI-compatible Grok endpoint (2api)
Star: 181
最近更新: Mar 13, 2026
Star: 1,405
最近更新: Jun 8, 2026
Turn any AI agent into an AI Scientist. The #1 Agent Skills library for science, used by 250,000+ scientists worldwide. 177 ready-to-use validated skills plus 100+ scientific databases covering biology, chemistry, medicine, and drug discovery. Compatible with Cursor, Claude Code, Codex, Pi, Antigravity, and the open...
47,326Oct 1, 2026Turn any AI agent into an AI Scientist. The #1 Agent Skills library for science, used by 250,000+ scientists worldwide. 177 ready-to-use validated skills plus 100+ scientific databases covering biology, chemistry, medicine, and drug discovery. Compatible with Cursor, Claude Code, Codex, Pi, Antigravity, and the open...
Star: 47,326
最近更新: Oct 1, 2026
SkillOpt is a text-space optimizer that trains reusable natural-language skills for frozen LLM agents through trajectory-driven edits, validation-gated updates, and deployable best_skill.md artifacts.
17,943Sep 30, 2026SkillOpt is a text-space optimizer that trains reusable natural-language skills for frozen LLM agents through trajectory-driven edits, validation-gated updates, and deployable best_skill.md artifacts.
Star: 17,943
最近更新: Sep 30, 2026
A comprehensive collection of Agent Skills for context engineering, multi-agent architectures, and production agent systems. Use when building, optimizing, or debugging agent systems that require effective context management.
18,060Oct 1, 2026A comprehensive collection of Agent Skills for context engineering, multi-agent architectures, and production agent systems. Use when building, optimizing, or debugging agent systems that require effective context management.
Star: 18,060
最近更新: Oct 1, 2026
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
63,366Oct 1, 2026AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
Star: 63,366
最近更新: Oct 1, 2026
A .codex SKILL for issue-driven ML/AI arXiv review papers: scaffold LaTeX + verify BibTeX citations end-to-end.
360Dec 29, 2025A .codex SKILL for issue-driven ML/AI arXiv review papers: scaffold LaTeX + verify BibTeX citations end-to-end.
Star: 360
最近更新: Dec 29, 2025
Unofficial Python API and agentic skill for Google Gemini Notebook. Full programmatic access to NotebookLM's features—including capabilities the web UI doesn't expose—via Python, CLI, and AI agents like Claude Code, Codex, and OpenClaw.
19,590Oct 2, 2026Unofficial Python API and agentic skill for Google Gemini Notebook. Full programmatic access to NotebookLM's features—including capabilities the web UI doesn't expose—via Python, CLI, and AI agents like Claude Code, Codex, and OpenClaw.
Star: 19,590
最近更新: Oct 2, 2026
Curated list of Agent Skills centered on people, relationships, commemorative scenes, and methodological perspectives
4,677Aug 19, 2026Curated list of Agent Skills centered on people, relationships, commemorative scenes, and methodological perspectives
Star: 4,677
最近更新: Aug 19, 2026
相关教程
通过 Grok 端点给 Codex 或 Claude 接上实时联网搜索
一个 181 星的 skill,把你代理的查询——版本号、API 文档、报错信息——路由到你自己的 OpenAI 兼容 Grok 端点,并返回带来源的 JSON。
跟着 DeepLearning.AI 课程文件通过实例学 Agent Skills
《Agent Skills with Anthropic》短课程的官方资料仓库,包含这门课程的提示词、笔记和数据文件——克隆下来,对比同一个任务在有技能与无技能时的差别。
claude-osint:把 Claude 变成授权侦察分析师的 8 个技能
一个 2,583 星的仓库,用 8 个 SKILL.md 文件把外部侦察方法论交给 Claude,只服务于你拥有或获得书面授权评估的资产,并对主动利用划出硬边界。
agent-skills-with-anthropic:吴恩达 Agent Skills 课程中文配套
Datawhale 的这个 1,502 星学习配套把吴恩达在 DeepLearning.AI 的 Agent Skills 短课程翻译成中文,逐课补充知识点梳理和代码解读,每课都配一名指定核查人——是中文使用者最快的入门通道。
Apify Agent Skills:描述数据需求,让 agent 去抓取
Apify 官方的 2,365 星 skill 包,让编码 agent 在 130 多个精选爬虫里挑选并运行 Actor,找不到还能回退到 3 万多个 Actor 的 Apify Store——五个 skill 覆盖抓取、Actor 构建和 SDK 集成。
auto-empirical-research-skills:带基准测试的 1,096 个计量研究 skill
一份 3.6k 星的 Stanford REAP × CoPaper.AI 合集,把实证研究方法——从数据清洗到期刊投稿——打包成 agent skill,并用 19 个基准任务和公开评分器检验你的 agent 是否真的跑对了。
browserbase/skills:给 Claude Code 一整套浏览器自动化工具箱
一份 3.7k 星的官方 skill 集合,把 Claude Code 变成能干活的浏览器 agent——支持带验证码处理的远程会话、DevTools trace 捕获、对抗式 UI 测试,以及一条命令 QA 你的本地开发服务器。
用 Context Engineering Skills 构建生产级 AI Agent
17 个围绕 context engineering 与 harness engineering 的 skill——覆盖 context 基础、多 agent 模式、记忆系统、工具设计、评估与自改进回路,用于生产 AI agent。
用 notebooklm-py 以编程方式访问 NotebookLM
一套非官方的 Python API、CLI、MCP server 与 agent skill,用于 Google NotebookLM(现称 Gemini Notebook)——生成音频概览、视频讲解、幻灯片、信息图、测验与报告,并为 agent 提供跨 session 记忆。
用微软 SkillOpt 自动生成 Agent Skill
SkillOpt 通过轨迹驱动的编辑,为冻结的 LLM agent 训练可复用的 best_skill.md 产物——在 GPT-5.5、Codex CLI、Claude Code CLI 上相对 SearchQA 的实测增益 +23.5 准确率点,且无需微调模型。
用 Scientific Agent Skills 把任意 AI Agent 变成科学家
一套 161 个现成 skill,让你的 AI agent 直接调用 PubChem、RDKit、AlphaFold 等科学数据库与 Python 包,把科研流水线变成一次对话。
掌握上下文工程:构建更好的 AI Agent 系统
如何使用 muratcankoylan/agent-skills-for-context-engineering 为生产级 AI Agent 设计、优化与评估上下文窗口。
用 Last 30 Days Skill 在几分钟内调研任意主题
实战指南:使用 mvanhorn/last30days-skill,让 AI Agent 在 Reddit、X、YouTube、HN 等平台并行调研。