AgentHubAgentHub
AI DailyAll →
无需游戏引擎!马斯克称特斯拉 FSD v15 实景视图仅动用 AI4 芯片一个 TRIP 加速器一半算力IT之家 · 92中国第一家:小鹏汇天 A868 电动飞行汽车倾转过渡试飞成功IT之家 · 91终抵千万里程碑:曝华为 Mate 80 系列手机终端销量破 1000 万台,标准版占比近 58%IT之家 · 90韩国高考 AI 作弊事件发酵,教育部长引咎辞职IT之家 · 89特斯拉 FSD v15 全新可视化 UI 引热议,消息称是在硬件劣势的情况做了 3D 高斯全场景重建IT之家 · 88九年长篇今日完结:国漫《斗破苍穹》剧场版《双帝之战》定档午间 12 点开播IT之家 · 87品牌全球销量破百万:长城坦克首支单曲《百万分之一》定档 11 月 25 日“唱响”IT之家 · 86OSTA O1 Pro 为何搭载麒麟 9030S?华为星芒刘赫总结三大原因,价格考虑也是关键IT之家 · 85Anthropic 数学家 AI 攻克黎曼猜想?Lean 社区通过核验,数十人报告代码库编译正确IT之家 · 842026 脑机接口报告:全球企业增至 1175 家,中国三大城市群集聚 190 余家IT之家 · 83奕境曾清林回应“汽车博主直播喊话车企拆车”话题:会落实,时间和方式定了第一时间公开IT之家 · 82试题经过大模型测试?国家中医药管理局认证中心:未与任何大模型企业合作,亦未开展测评IT之家 · 81
✦AgentHub · AI Agent Capability Store · 40,000+ Tools

AgentHub·Give your AI agent superpowers

MCP Server & Agent Skill directory — copy, paste, connected

Discover plugins for Cursor, Claude Code, VS Code and Trae: browsing, GitHub, databases and more.

Capabilities you can add to your AI

All scenarios →

No jargon needed — each card is a superpower. Open one and we'll show you exactly what to install.

Trending capabilities today

Full rankings →

What everyone is installing right now · by installs and usage

agent-browser

SkillSkillsMP1M

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.

source

Featured Prompts

More Prompts →

Spotlight refreshes hourly

What is AgentHub?

Your all-in-one workbench for MCP and Agent Skills—with 1-click configs, China MCP hub, and scenario-based playbooks.

Developer Architecture & Standards

Why AgentHub: The Foundation for AI Agents & MCP Developers

System capabilities, client setup guides, and developer FAQs

Expand

The Model Context Protocol (MCP), open-sourced by Anthropic, is the open standard uniting LLMs with external tools. AgentHub indexes 40,000+ vetted tools and configs for Cursor, Claude Code, and VS Code.

1

System-level Extensibility

Break free from pure chat to empower AI agents with real actions across desktop, web, and infrastructure:

  • Local file system & codebase indexing, refactoring, and static analysis
  • Real browser automation, page screenshots, form filling, and E2E testing
  • Production database (Postgres, MySQL, Redis) direct query & analytics
  • Git version control, automatic commits, and GitHub PR reviews
2

Zero-friction Client Setup

Automatically generated configurations for all leading AI developer environments:

  • Cursor IDE: 1-click JSON import & .cursorrules guidelines
  • Claude Code: Terminal CLI commands & environment variables injection
  • VS Code: Native compatibility with Roo Code and Cline extensions
  • ByteDance Trae & global developer tools full integration
3

Official Protocol & Security

Fully aligned with Anthropic's Model Context Protocol specification & security best practices:

  • Specification Aligned: 100% compliant with Anthropic official MCP standards
  • Local Privacy: Environment variables & API keys remain strictly on your machine
  • Transparent Auditing: Verified GitHub provenance and static vulnerability scanning
  • Community Maintained: Continuous health checks and active removal of deprecated tools

Developer Frequently Asked Questions

What is Model Context Protocol (MCP) and how does it differ from traditional APIs?

Model Context Protocol (MCP) is an open standard open-sourced by Anthropic that enables AI models to securely interact with local and remote data sources and tools. Unlike proprietary APIs, MCP offers a unified protocol contract so that one server works across Cursor, Claude Code, and VS Code seamlessly.

How do I install an MCP server into Cursor or Claude Code from AgentHub?

Find the desired tool on AgentHub, go to its detail page, and copy the ready-to-use JSON configuration snippet. Open Cursor Settings > Features > MCP or Claude Code config file, paste the snippet, and reload your environment to start using the tool immediately.

What is an Agent Skill and how does it collaborate with MCP Servers?

An Agent Skill (such as ClawHub specifications) provides prompt workflows, domain instructions, and behavior rules (SKILL.md), while MCP servers provide underlying tool execution interfaces. Together, Skills handle planning and logic while MCP servers handle actions.

Is it safe to install tools from AgentHub?

AgentHub indexes open-source and community-vetted tools with direct links to GitHub repositories and permission manifests. All credentials and API keys stay on your local machine and are never transmitted to our servers.

Partners & Friends

Apply for Link
Featured & Recognized
Featured on toolfame.comFeatured on 新趣集Featured on doforai.toolsFeatured on Twelve ToolsFeatured on ShowMeBestAIFeatured on EasyLaunch