xmind

# xmind 🧠

Generate and read **XMind** `.xmind` files using the published MCP server `xmind-generator-mcp` (npm).

This skill is chat-first:
- “Generate an XMind for this (optionally: save to …)” → generates a `.xmind` and sends it back.
- “Read this XMind and tell me what it’s about” → explains the map first, then offers a Markdown export.

## When to use
- User wants: “generate an XMind file from this outline / plan / PRD / test plan”.
- Output should be a real `.xmind` file that opens in XMind.

## Input format (for generation)
The MCP tool `generate-mind-map` accepts **Schema A** JSON:
```json
{
  "title": "Root topic",
  "filename": "mindmap-name-no-date",
  "topics": [
    {
      "title": "Topic title (keep core info in the title)",
      "note": "Optional: use sparingly; only when the title would be too long",
      "labels": ["optional"],
      "markers": ["Arrow.refresh"],
      "children": [{"title": "Child topic"}]
    }
  ]
}
```

## How the assistant should call the MCP (stdio via mcporter)
### Generate XMind
Use `npx xmind-generator-mcp@0.1.2` as the MCP server command:
```bash
mcporter call --stdio "npx -y xmind-generator-mcp@0.1.2" generate-mind-map --args '{...}'
```

### Read XMind (Markdown export)
```bash
mcporter call --stdio "npx -y xmind-generator-mcp@0.1.2" read-mind-map --args '{"inputPath":"/path/to/file.xmind","style":"A"}'
```

## Read behavior in chat (UX requirement)
When a user sends an `.xmind` file and asks to “read/understand” it:
1) First, explain in detail what the mind map is about (summary + structure + key points + actionable items + gaps).
2) Only after that, ask whether the user wants an export.
3) If the user wants an export, default to **Markdown**.

- `generate --output` can be a directory (recommended) or a full `.xmind` path.
- If outputPath is omitted, the MCP server defaults to its configured `outputPath` environment variable (see below).

## Chat-first workflow (what the assistant should do)
Trigger phrases:
- “Gene...