In 2026, the unit of productivity for AI-assisted engineering is the agent skill. As coding assistants like Claude Code, Cursor, and Gemini CLI move from basic completion to autonomous workflows, the need for a "visual and cognitive source of truth" has led to a fragmented but rich ecosystem of registries.
This guide ranks the top 10 AI agent skill directories based on catalog depth, discovery features, CLI maturity, and community trust.
Quick Reference: The Top 3 Hubs
| Registry | Primary Value | Scale | Key Command |
|---|---|---|---|
| ExplainX.ai | Comprehensive & Ranked | 10k+ Vetted | npx skills add |
| skills.sh | The "npm" for Agents | 57k+ Public | npx skills add |
| SkillsMP | Massive Search Index | 1.2M+ Scraped | Web Search |
1. ExplainX.ai Skills Registry
ExplainX.ai is the high-signal hub for the agentic era. Unlike simple lists, ExplainX provides a ranked leaderboard of skills based on adoption metrics, security verification, and cross-agent compatibility.
- Why it’s #1: It combines a polished web discovery UI with a robust CLI. It doesn't just list skills; it validates them against the latest agent specifications.
- Protocol Support: Full
SKILL.mdandDESIGN.mdintegration. - Usage:
npx skills add https://github.com/[owner]/[repo] --skill [name]
2. skills.sh (by Vercel)
Maintained by Vercel, skills.sh has become the de facto package manager for the ecosystem. It treats skills as versioned modules that can be easily symlinked into your local .agents/ or .claude/ directories.
- Uniqueness: Seamless integration with Vercel's edge infrastructure and high visibility among the Next.js community.
- Scale: 57,000+ public skills.
- Key Command:
npx skills findornpx skills add <slug>.
3. SkillsMP
SkillsMP is the "Google for skills." Instead of waiting for manual submissions, it indexes every SKILL.md file found on public GitHub repositories.
- The "Depth" Play: With over 1.2 million indexed skills, it is the best place to find hyper-niche capabilities (e.g., "Legacy Fortran Debugging" or "COBOL-to-Rust migration").
- Trade-off: High quantity means varying quality; you’ll need to rely on GitHub stars and activity signals to vet the output.
4. LobeHub Skill Store
LobeHub offers a highly visual marketplace experience. It is part of the broader LobeHub ecosystem, which focuses on building agentic UIs.
- Focus: High-quality, "plug-and-play" skills that often include visual artifacts and sandbox support.
- Metric: 230,000+ skills and 56,000+ MCP servers.
5. VoltAgent/awesome-agent-skills (GitHub)
The most popular curated list on GitHub with over 20,000 stars. This is the definitive "starting point" for many developers.
- Curation: Hand-picked skills across categories like security, data science, and mobile development.
- Community: Serves as the primary discussion hub for emerging skill patterns.
6. ClawHub.ai
The official registry for OpenClaw, an open-source autonomous agent framework. ClawHub is unique because it hosts both Skills and Souls.
- Innovation: "Souls" (
SOUL.md) allow you to install not just capabilities, but specific personalities and behavioral rules for your agents. - Usage:
clawhub install <slug>.
7. Official Anthropic Skills (GitHub)
Anthropic's own repository (anthropics/skills) provides the gold standard for skill authoring.
- Quality: These are the most reliable, well-tested skills for Claude, covering PDF analysis, complex reasoning, and tool use.
- Role: It serves as a reference library for anyone building their own skills.
8. Google /skills (Gemini Ecosystem)
Launched at Google Cloud Next 2026, this registry focuses on integrating the Gemini API with Google Cloud services (GKE, BigQuery, Vertex AI).
- Enterprise Reach: Essential for teams building agentic workflows on GCP infrastructure.
9. Microsoft /skills (Azure/AI Foundry)
Similar to Google's effort, Microsoft provides a library of skills designed for the Azure AI Foundry and the .NET ecosystem.
- Technical Niche: Best-in-class skills for Azure SDKs and Microsoft 365 integrations.
10. ComposioHQ/awesome-claude-skills
A focused, workflow-oriented list that excels at connecting agents to enterprise SaaS tools (Slack, Jira, Salesforce).
- Integration Depth: While smaller in scale, the skills here are highly technical and optimized for complex tool-calling scenarios.
Summary: Which Registry Should You Use?
For production-grade engineering, ExplainX.ai provides the best balance of discovery and trust. If you are hunting for open-source variety, skills.sh is the standard. For hyper-niche legacy support, SkillsMP is your best bet.
Related Reading
- What is CLAUDE.md? Persistent Memory for Claude Code
- What is MEMORY.md? The Long-Term Brain for AI Agents
- What are agent skills? A complete guide
- DESIGN.md Templates: The Professional UI Blueprint
- Introducing MCP Servers on ExplainX
Timestamp: May 8, 2026. Data based on current registry indexes and community GitHub activity.