wondel-skills-mcp

# Wondel.ai Skills MCP

Let your agent ask which skill to use — then hand it over.

An open, **read-only** MCP server over 50 book-based agent skills and 12 guided journeys, from [Wondel.ai](https://wondel.ai). It routes a task to the right framework, tells your agent which ones to skip and why, and serves the real SKILL.md in-session. No install, no account, no auth.

## Connect

Streamable HTTP, one URL, nothing to configure.

```bash
claude mcp add --transport http wondel-skills https://skills.wondel.ai/mcp
```

In any client that takes an `mcpServers` block — Cursor, Codex, OpenClaw and the rest:

```json
{
  "mcpServers": {
    "wondel-skills": {
      "type": "http",
      "url": "https://skills.wondel.ai/mcp"
    }
  }
}
```

## Tools

| Tool | What it does |
|---|---|
| `recommend_skills(task, stack?, phase?, max?)` | Route a task to the right skills — which to load first, which to skip (quoted from each skill's own "not for" line), and a guided journey when the task spans a whole project. Start here. |
| `search_skills(query, kind?, category?, limit?)` | Search by keyword or plain task language ("my UI looks amateur", "pricing", "dark mode"). |
| `get_skill(slug, sections?)` | Cheap metadata for one skill (~300 tokens): definition, when to use, when not to, related slugs. |
| `load_skill(slug, part?)` | The skill's actual instructions — the upstream SKILL.md, ready to follow in-session. |
| `get_journey(slug? \| goal?)` | One guided journey as a phase plan: ordered skills, the decision question each phase answers, and go/no-go gates. |

Every tool is read-only and annotated as such.

Info: https://skills.wondel.ai/mcp · Contact: contact@wondel.ai · [Wondel.ai](https://wondel.ai) — product studio building consumer AI experiences (text, image, video, voice), Warsaw, Poland.