agentskill-mcp
Official# agentskill-mcp
MCP server for [agentskill.sh](https://agentskill.sh) -- search, discover, and install AI agent skills from the largest skills directory (110k+ skills).
## Tools
| Tool | Description |
|------|-------------|
| `search_skills` | Search skills by keyword with platform, category, and security score filters |
| `get_skill` | Get full details including SKILL.md, security analysis, and quality review |
| `install_skill` | Install a skill to the local skills directory with security checks |
| `get_trending` | Get hot (24h), trending (7d), top (all time), or latest skills |
| `browse_skillsets` | Browse curated skill collections bundled by workflow or role |
| `install_skillset` | Install all skills from a skillset at once |
| `rate_skill` | Rate a skill 1-5 to help others discover the best skills |
| `check_updates` | Check if installed skills have newer versions available |
## Installation
### Claude Code
```bash
claude mcp add agentskill -- npx -y agentskill-mcp
```
### Claude Desktop
Add to `~/Library/Application Support/Claude/claude_desktop_config.json`:
```json
{
"mcpServers": {
"agentskill": {
"command": "npx",
"args": ["-y", "agentskill-mcp"]
}
}
}
```
### Cursor / Windsurf / Cline / Other MCP Clients
```json
{
"mcpServers": {
"agentskill": {
"command": "npx",
"args": ["-y", "agentskill-mcp"]
}
}
}
```
## Usage Examples
Once connected, your AI agent can:
- **Search**: "Find skills for SEO optimization"
- **Filter**: "Search for marketing skills with security score above 80"
- **Browse**: "Show me trending skills for Claude Code"
- **Skillsets**: "Browse curated skillsets for frontend development"
- **Install**: "Install the compound-engineering/frontend-design skill"
- **Install collection**: "Install the seo-starter skillset"
- **Rate**: "Rate compound-engineering/frontend-design 5/5"
- **Updates**: "Check for updates on my installed skills"
## Supported Platforms
Skills are available for 20+ platforms: Claude Code, Cursor, GitHub Copilot, Windsurf, Codex, Gemini CLI, Hermes, ChatGPT, Cline, VS Code, OpenCode, Aider, AMP, Goose, Roo Code, Trae, and more.
## License
MIT
TDQS
Scored across 4 tools
Each tool has a clearly distinct purpose with no overlap: get_skill retrieves detailed information for a specific skill, get_trending lists popular skills, install_skill handles installation, and search_skills performs keyword-based searches. The descriptions reinforce these distinct roles, making misselection unlikely.
All tool names follow a consistent verb_noun pattern using snake_case (e.g., get_skill, install_skill). The verbs (get, get, install, search) are appropriate and predictable, creating a coherent naming convention throughout the set.
With 4 tools, this server is well-scoped for its purpose of managing AI agent skills. Each tool serves a clear and necessary function (discovery, retrieval, installation), and none feel redundant or missing, making the count appropriate for the domain.
The tool set covers core workflows for skill discovery and installation effectively, including search, trending, detailed retrieval, and installation. A minor gap exists in lifecycle management, such as updating or uninstalling skills, but agents can work around this with the provided tools for most use cases.