Metadata-Version: 2.4
Name: skeleton-ui-mcp-server
Version: 0.1.0
Summary: MCP server exposing Skeleton UI documentation as tools for coding agents
Project-URL: Repository, https://github.com/dmglab/skeleton-ui-mcp-server
Project-URL: Issues, https://github.com/dmglab/skeleton-ui-mcp-server/issues
Author: Daniel Gehrlein
License-Expression: MIT
Keywords: coding-agents,documentation,mcp,skeleton-ui,svelte
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Topic :: Software Development :: Documentation
Requires-Python: >=3.14
Requires-Dist: mcp[cli]>=1.27.0
Requires-Dist: toon-format>=0.9.0b1
Provides-Extra: fetch
Requires-Dist: httpx>=0.28.1; extra == 'fetch'
Requires-Dist: python-slugify>=8.0.4; extra == 'fetch'
Description-Content-Type: text/markdown

# Skeleton UI MCP Server

> This project was built with the assistance of coding agents (Claude Code).

An MCP server that exposes the [Skeleton UI](https://www.skeleton.dev) documentation as tools for coding agents.
Documentation is served from pre-fetched static files — no live network calls during tool use.

## Tools

| Tool                         | Description                                                                                       |
|------------------------------|---------------------------------------------------------------------------------------------------|
| `search_docs(query, limit?)` | Keyword search across titles, slugs, excerpts, and groups. Use this first to find the right slug. |
| `list_docs_by_group(group)`  | List all docs in a category (e.g. `"Framework Components"`).                                      |
| `list_all_docs()`            | Full index of all 58 docs. Prefer the tools above.                                                |
| `get_docs_for(slug)`         | Returns full Markdown documentation for a slug.                                                   |

## Claude Desktop Configuration

### Using uvx (recommended)

No installation needed — `uvx` fetches and runs the package automatically:

```json
{
  "mcpServers": {
    "skeleton-ui-docs": {
      "command": "uvx",
      "args": ["skeleton-ui-mcp-server"]
    }
  }
}
```

### From source

```bash
git clone <repo>
cd skeleton-ui-mcp-server
uv sync
```

```json
{
  "mcpServers": {
    "skeleton-ui-docs": {
      "command": "uv",
      "args": [
        "--directory",
        "/absolute/path/to/skeleton-ui-mcp-server",
        "run",
        "skeleton-ui-mcp-server"
      ]
    }
  }
}
```

## Refreshing the Static Files

The `skeleton_ui_mcp_server/static/` directory contains pre-fetched documentation from `skeleton.dev`. To refresh it manually:

```bash
uv run --extra fetch python index.py
```

This re-fetches all pages listed in `skeleton_ui_mcp_server/static/_llms.txt` and regenerates the individual JSON files.

A GitHub Actions workflow (`.github/workflows/update-static.yml`) runs this automatically every night at 03:00 UTC and commits any changes back to the repository. It can also be triggered manually via **Actions → Update static docs → Run workflow**. The integrity tests (`tests/test_audit.py`) run as part of the workflow before committing — if they fail the commit is skipped.

## Quality Assurance

Install dev dependencies first:

```bash
uv sync --group dev
```

| Command                                            | What it does                                        |
|----------------------------------------------------|-----------------------------------------------------|
| `uv run --group dev pytest -v`                     | Run all tests (tool smoke tests + static integrity) |
| `uv run --group dev pytest tests/test_tools.py -v` | MCP tool tests only                                 |
| `uv run --group dev pytest tests/test_audit.py -v` | Static data + `_llms.txt` format tests only         |
| `uv run --group dev pyright`                       | Type-check `server.py` and `index.py`               |
| `uv run --group dev pip-audit`                     | Scan dependencies for known vulnerabilities         |
