Agent Skills Hub
agentskillshub.devSecurity-reviewed MCP server directory
Find, compare, and install security-reviewed MCP servers and AI agent skills for Claude Code workflows.
Overview
AgentSkillsHub is a curated directory of MCP servers and AI agent skills for developers building with Claude Code and similar AI coding assistants. It provides security grading, installation guidance, and compatibility context for each listing. Teams use it to evaluate and adopt agent tools with confidence before adding them to local or production workflows.
How it works
The directory includes 1213+ security-scanned MCP skills, workflows, and integrations. Each skill receives a security grade from A to F based on permission scope, maintenance signals, and code review indicators. The scanner checks for 45+ vulnerability patterns including arbitrary code execution, API key leakage, and unbounded file access. Listings also include installation commands, required environment variables, and quick-start guides.
Who it is for
AgentSkillsHub is for developers and teams building with Claude Code, Cursor, Codex, Opencode, and similar AI coding assistants. It helps shortlist MCP servers, Claude Code skills, and agent workflow extensions before install. Popular use cases include GitHub integration, code review automation, Stripe payment processing, Shopify store management, and Zapier-style automation. The site also tracks OpenClaw skills with a leaderboard ranked by community adoption and security audit score.
Updates and trends
The directory is updated continuously as new MCP servers are released and existing ones are updated. Security grades are reviewed whenever upstream packages publish major changes. Users can filter by recently added or recently updated to find the latest additions. The site also publishes weekly trend pulses covering MCP registries, workflow automation, browser agents, memory systems, and research agents.
Pricing and limits
All skills in the directory are free to use. The site charges $99 for premium security audits and verification badges, which include manual static analysis and a security certificate. The 'VERIFIED SECURE' badge is available for skills that pass the audit. The directory also includes 120 official skills and reports that 100% of scanned skills are safe, with 2,555 OpenClaw skills (85%) passing high-safety grade in the latest audit.
Key features
Security grading
Security grades (A-F) assigned to every skill based on permission scope, maintenance signals, and code review indicators.
Vulnerability scanning
Static analysis engine scans for 45+ vulnerability patterns including arbitrary code execution, credential leakage, and unbounded file access.
Searchable directory
Searchable directory of 1213+ security-scanned MCP skills, workflows, and integrations, filterable by category, grade, and recency.
Install guidance
Each listing includes installation commands, required environment variables, and quick-start guides for Claude Code and similar assistants.
Editorial review
Editorial team reviews every listing with permission audit, README analysis, and operational risk triage before publication.
Use cases
Development tools
Find and install MCP servers for GitHub integration, code review automation, and deployment pipelines.
Finance and operations
Discover skills for Stripe payment processing, invoice generation, and expense tracking.
E-commerce
Locate skills for Shopify store management, inventory updates, and product SEO.
Marketing
Find skills for social media scheduling, SEO analysis, and content generation.
Workflow automation
Explore Zapier-style automation, data transformation, and API orchestration skills.
Pricing & limits
From the official site · checked 10/14/2026Frequently asked questions about Agent Skills Hub
- What is AgentSkillsHub?
- AgentSkillsHub is a curated directory of MCP servers and AI agent skills for developers building with Claude Code and similar AI coding assistants. Each listing includes installation guidance, security scoring, and compatibility context.
- What is an MCP server?
- An MCP (Model Context Protocol) server is a lightweight integration that gives AI coding assistants like Claude Code structured access to external tools and data, exposing operations as typed tools the agent can invoke during a coding session.
- How do I install an MCP skill in Claude Code?
- Use the command 'claude mcp add <name> -- npx <package>' in your terminal. Each skill page includes the exact install command, required environment variables, and a quick-start guide.
- Are all MCP servers safe to use?
- Safety varies. AgentSkillsHub assigns a security grade (A through F) based on permission scope, maintenance signals, and code review indicators. Always check the grade and permissions before installing, especially in team or production environments.
- How often is the directory updated?
- The directory is updated continuously as new MCP servers are released and existing ones are updated. Security grades are reviewed whenever upstream packages publish major changes.
