$git clone https://github.com/intellectronica/skillz> ⚠️ Experimental proof‑of‑concept. Potentially unsafe. Treat skills like untrusted code and run in sandboxes/containers. Use at your own risk.
| 1 | # Skillz |
| 2 | |
| 3 | ## 👌 **Use _skills_ in any agent** _(Codex, Copilot, Cursor, etc...)_ |
| 4 | |
| 5 | [](https://pypi.org/project/skillz/) |
| 6 | [](https://pypi.org/project/skillz/) |
| 7 | |
| 8 | > ⚠️ **Experimental proof‑of‑concept. Potentially unsafe. Treat skills like untrusted code and run in sandboxes/containers. Use at your own risk.** |
| 9 | |
| 10 | **Skillz** is an MCP server that turns [Claude-style skills](https://github.com/anthropics/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. |
| 11 | |
| 12 | > 💡 You can find skills to install at the **[Skills Supermarket](http://skills.intellectronica.net/)** directory. |
| 13 | |
| 14 | ## Quick Start |
| 15 | |
| 16 | To run the MCP server in your agent, use the following config (or equivalent): |
| 17 | |
| 18 | ```json |
| 19 | { |
| 20 | "skillz": { |
| 21 | "command": "uvx", |
| 22 | "args": ["skillz@latest"] |
| 23 | } |
| 24 | } |
| 25 | ``` |
| 26 | |
| 27 | with the skills residing at `~/.skillz` |
| 28 | |
| 29 | _or_ |
| 30 | |
| 31 | ```json |
| 32 | { |
| 33 | "skillz": { |
| 34 | "command": "uvx", |
| 35 | "args": ["skillz@latest", "/path/to/skills/direcotry"] |
| 36 | } |
| 37 | } |
| 38 | ``` |
| 39 | |
| 40 | or Docker |
| 41 | |
| 42 | You can run Skillz using Docker for isolation. The image is available on Docker Hub at `intellectronica/skillz`. |
| 43 | |
| 44 | To run the Skillz MCP server with your skills directory mounted using Docker, configure your agent as follows: |
| 45 | |
| 46 | Replace `/path/to/skills` with the path to your actual skills directory. Any arguments after `intellectronica/skillz` in the array are passed directly to the Skillz CLI. |
| 47 | |
| 48 | ```json |
| 49 | { |
| 50 | "skillz": { |
| 51 | "command": "docker", |
| 52 | "args": [ |
| 53 | "run", |
| 54 | "-i", |
| 55 | "--rm", |
| 56 | "-v", |
| 57 | "/path/to/skills:/skillz", |
| 58 | "intellectronica/skillz", |
| 59 | "/skillz" |
| 60 | ] |
| 61 | } |
| 62 | } |
| 63 | ``` |
| 64 | |
| 65 | ## Gemini CLI Extension |
| 66 | |
| 67 | A Gemini CLI extension is available at [intellectronica/gemini-cli-skillz](https://github.com/intellectronica/gemini-cli-skillz). |
| 68 | |
| 69 | Install it with: |
| 70 | |
| 71 | ```bash |
| 72 | gemini extensions install https://github.com/intellectronica/gemini-cli-skillz |
| 73 | ``` |
| 74 | |
| 75 | This extension enables Anthropic-style Agent Skills in Gemini CLI using the skillz MCP server. |
| 76 | |
| 77 | ## Usage |
| 78 | |
| 79 | Skillz looks for skills inside the root directory you provide (defaults to |
| 80 | `~/.skillz`). Each skill lives in its own folder or zip archive (`.zip` or `.skill`) |
| 81 | that includes a `SKILL.md` file with YAML front matter describing the skill. Any |
| 82 | other files in the skill become downloadable resources for your agent (scripts, |
| 83 | datasets, examples, etc.). |
| 84 | |
| 85 | An example directory might look like this: |
| 86 | |
| 87 | ```text |
| 88 | ~/.skillz/ |
| 89 | ├── summarize-docs/ |
| 90 | │ ├── SKILL.md |
| 91 | │ ├── summarize.py |
| 92 | │ └── prompts/example.txt |
| 93 | ├── translate.zip |
| 94 | ├── analyzer.skill |
| 95 | └── web-search/ |
| 96 | └── SKILL.md |
| 97 | ``` |
| 98 | |
| 99 | When packaging skills as zip archives (`.zip` or `.skill`), include the `SKILL.md` |
| 100 | either at the root of the archive or inside a single top-level directory: |
| 101 | |
| 102 | ```text |
| 103 | translate.zip |
| 104 | ├── SKILL.md |
| 105 | └── helpers/ |
| 106 | └── translate.js |
| 107 | ``` |
| 108 | |
| 109 | ```text |
| 110 | data-cleaner.zip |
| 111 | └── data-cleaner/ |
| 112 | ├── SKILL.md |
| 113 | └── clean.py |
| 114 | ``` |
| 115 | |
| 116 | ### Directory Structure: Skillz vs Claude Code |
| 117 | |
| 118 | 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` or `.skill` files (as shown in the examples above). |
| 119 | |
| 120 | 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` or `.skill` files are not supported. |
| 121 | |
| 122 | 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. |
| 123 | |
| 124 | **Claude Code–compatible layout:** |
| 125 | |
| 126 | ```text |
| 127 | skills/ |
| 128 | ├── hello-world/ |
| 129 | │ ├── SKILL.md |
| 130 | │ └── run.sh |
| 131 | └── summarize-text/ |
| 132 | ├── SKILL.md |
| 133 | └── run.py |
| 134 | ``` |
| 135 | |
| 136 | **Skillz-only layout examples** (not compatible with Claude Code): |
| 137 | |
| 138 | ```text |
| 139 | skills/ |
| 140 | ├── text-tools/ |
| 141 | │ └── summarize-text/ |
| 142 | │ |