mind-map
npx skills add https://github.com/mblode/agent-skills --skill mind-map
Agent 安装分布
Skill 文档
Mind Map
Analyze source material and produce a Mermaid mindmap diagram with 2-4 levels of branching hierarchy.
Reference Files
| File | Read When |
|---|---|
references/mermaid-syntax.md |
Default: mind map node shapes, indentation, icons, known limitations |
references/generation-guidelines.md |
Default: node text constraints, depth control, source-type strategies |
references/mermaid-reference.md |
Generating non-mind-map Mermaid diagrams or needing broader syntax |
Workflow
Copy this checklist to track progress:
Mind map progress:
- [ ] Step 1: Determine source type
- [ ] Step 2: Set scope (depth, budget, central concept)
- [ ] Step 3: Extract structure
- [ ] Step 4: Generate Mermaid mind map
- [ ] Step 5: Validate output
- [ ] Step 6: Present to user
Step 1: Determine source type
Identify which source type applies and run the corresponding discovery actions:
- Topic exploration: User names a concept. Ask one clarifying question about scope or audience if ambiguous. No file scanning needed.
- Codebase overview: Scan project structure. Read key config files (
package.json,pyproject.toml, etc.). Map modules, entry points, and dependencies. - File/document summary: Read the target file. Extract key themes, sections, and relationships between ideas.
- Feature planning / debugging: Gather requirements or symptoms. Decompose into components, phases, or possible causes.
Step 2: Set scope
Choose a depth level using the table in references/generation-guidelines.md (overview, standard, or detailed). Ask the user only if unclear. Default to standard (3 levels, 15-30 nodes).
Set the central concept: one clear noun or short phrase (1-3 words) for the root node. Total node budget: 40 max.
Step 3: Extract structure
- Identify 3-7 main branches from the source material (5 is ideal).
- For each branch, identify 2-5 sub-nodes per the depth setting.
- Group related concepts under the same branch.
- Each node must be a single keyword or short phrase, never a sentence.
- See
references/generation-guidelines.mdfor node text word limits and keyword extraction techniques.
Step 4: Generate Mermaid mind map
Load references/mermaid-syntax.md for exact syntax. Generate inside a fenced mermaid code block.
Use node shapes intentionally (see references/generation-guidelines.md for the full shape strategy):
- Root:
((circle))for the central concept - Main branches:
[square]for categories,(rounded)for processes - Leaf nodes: default for general items,
{cloud}for uncertain items
Step 5: Validate
Run these checks before presenting. If any fail, revise and re-check:
- Total node count (including root) is 40 or fewer
- Maximum depth does not exceed selected level
- No node text exceeds 6 words
- Root node text is 1-3 words
- Each main branch has 2-7 children
- No single-child branches (merge upward if found)
- No unescaped
()[]{}in plain-text nodes (will break Mermaid parsing) - Branch depths are balanced (within 1 level of each other)
Step 6: Present output
Output the mind map in a fenced mermaid code block:
```mermaid
mindmap
root((Central Concept))
[Branch 1]
Sub-topic A
Sub-topic B
[Branch 2]
Sub-topic C
(Process D)
```
Note where it renders: GitHub markdown, Mermaid Live Editor, VS Code with Mermaid extension.
If the user requests a file, write to a .md file containing the fenced block.
Anti-patterns
- Sentences as node labels (split into parent concept with child details instead)
- Depth beyond 4 levels (Mermaid becomes unreadable; split into multiple maps)
- More than 40 nodes (split into multiple focused maps)
- Adding icons unless requested (require Font Awesome, may not render)
- Mixing
mindmapwith other Mermaid diagram types in one code block
Related skills
creating-presentationsfor slide decks that complement visual overviewsplan-featurefor implementation plans when a mind map reveals feature scopedefine-architecturefor detailed architecture decisions beyond visual mapping