skillz 0.1.13

MCP server that exposes Claude-style skills to any MCP client.

Skillz

👌 Use skills in any agent (Codex, Copilot, Cursor, etc...)

⚠️ Experimental proof‑of‑concept. Potentially unsafe. Treat skills like untrusted code and run in sandboxes/containers. Use at your own risk.

Skillz is an MCP server that turns Claude-style skills (SKILL.md plus optional resources) into callable tools for any MCP client. It discovers each skill, exposes the authored instructions and resources, and can run bundled helper scripts.

💡 You can find skills to install at the Awesome Skills directory.

Quick Start

To run the MCP server in your agent, use the following config (or equivalent):

{
  "skillz": {
    "command": "uvx",
    "args": ["skillz@latest"]
  }
}

with the skills residing at ~/.skillz

or

{
  "skillz": {
    "command": "uvx",
    "args": ["skillz@latest", "/path/to/skills/direcotry"]
  }
}

or Docker

You can run Skillz using Docker for isolation. First, build the Docker image:

# Build the Docker image (requires Docker Buildx)
docker buildx build -t skillz .

To run the Skillz MCP server with your skills directory mounted using Docker, configure your agent as follows:

Replace /path/to/skills with the path to your actual skills directory. Any arguments after skillz in the array are passed directly to the Skillz CLI.

{
  "skillz": {
    "command": "docker",
    "args": [
      "run",
      "-i",
      "--rm",
      "-v",
      "/path/to/skills:/skillz",
      "skillz",
      "/skillz"
    ]
  }
}

Usage

Skillz looks for skills inside the root directory you provide (defaults to ~/.skillz). Each skill lives in its own folder or zip archive that includes a SKILL.md file with YAML front matter describing the skill. Any other files in the skill become downloadable resources for your agent (scripts, datasets, examples, etc.).

An example directory might look like this:

~/.skillz/
├── summarize-docs/
│   ├── SKILL.md
│   ├── summarize.py
│   └── prompts/example.txt
├── translate.zip
└── web-search/
    └── SKILL.md

When packaging skills as zips, include the SKILL.md either at the root of the archive or inside a single top-level directory:

translate.zip
├── SKILL.md
└── helpers/
    └── translate.js

data-cleaner.zip
└── data-cleaner/
    ├── SKILL.md
    └── clean.py

Directory Structure: Skillz vs Claude Code

Skillz supports a more flexible skills directory than Claude Code. In addition to a flat layout, you can organize skills in nested subdirectories and include skills packaged as .zip files (as shown in the examples above).

Claude Code, on the other hand, expects a flat skills directory: every immediate subdirectory is a single skill. Nested directories are not discovered, and .zip files are not supported.

If you want your skills directory to be compatible with Claude Code (for example, so you can symlink one skills directory between the two tools), you must use the flat layout.

Claude Code–compatible layout:

skills/
├── hello-world/
│   ├── SKILL.md
│   └── run.sh
└── summarize-text/
    ├── SKILL.md
    └── run.py

Skillz-only layout examples (not compatible with Claude Code):

skills/
├── text-tools/
│   └── summarize-text/
│       ├── SKILL.md
│       └── run.py
└── image-processing.zip

You can use skillz --list-skills (optionally pointing at another skills root) to verify which skills the server will expose before connecting it to your agent.

CLI Reference

skillz [skills_root] [options]

Flag / Option	Description
positional skills_root	Optional skills directory (defaults to ~/.skillz).
--transport {stdio,http,sse}	Choose the FastMCP transport (default stdio).
--host HOST	Bind address for HTTP/SSE transports.
--port PORT	Port for HTTP/SSE transports.
--path PATH	URL path when using the HTTP transport.
--list-skills	List discovered skills and exit.
--verbose	Emit debug logging to the console.
--log	Mirror verbose logs to /tmp/skillz.log.

---

Made with 🫶 by @intellectronica