/kb-visualize | knowledge-base

Visualize knowledge ontology: $ARGUMENTS <ultrathink> A picture is worth a thousand links. Visualization reveals patterns invisible in linear text, making the implicit explicit. </ultrathink> <megaexpertise type="ontology-builder"> The assistant should use the ontology-builder agent to create visual knowledge graphs using Graphviz or Mermaid, with optional filtering by category. </megaexpertise> <context> Filter: $ARGUMENTS Knowledge base path: .claude/knowledge/ Output path: .claude/knowledge/ontology/ Available filters: metadata, qa, code_index, patterns, plans, concepts, memory_anchors, or "all" </context> <requirements>

Parse all knowledge entries and relationships
Build graph data structure
Generate visual representation (Graphviz DOT or Mermaid)
Color-code by entry type
Show relationship types with different edge styles
Perform gap analysis
Provide actionable insights
Save to file for viewing </requirements>

<actions> 1. Parse Arguments: - Extract type filter if provided (default: "all") - Validate filter is recognized category

Launch Ontology Builder Agent:

Use the Task tool with subagent_type="general-purpose" to invoke the ontology-builder agent.

Prompt:
"Generate visual ontology graph for knowledge base.
Filter: $ARGUMENTS (or 'all' if no filter)

Please:
1. Scan all knowledge entries in .claude/knowledge/
2. Parse YAML frontmatter for metadata and relationships
3. Extract wiki-style [[links]] from content
4. Build graph data structure:
   - Nodes: knowledge entries (with uuid, title, type, tags, status)
   - Edges: relationships (with type and directionality)
5. Generate Graphviz DOT format with:
   - Color coding by type (metadata=blue, patterns=orange, etc.)
   - Different edge styles (solid, dashed, bold) for relationship types
   - Node styling for status (dashed=archived, dotted=deprecated)
6. Create output directory: .claude/knowledge/ontology/
7. Save DOT file and generate PNG/SVG using Graphviz
8. Perform gap analysis:
   - Isolated entries (0 connections)
   - Poorly connected entries (< 2 connections)
   - Missing documentation areas
   - Broken references
9. Generate text-based report with:
   - Summary statistics
   - Gap analysis findings
   - Relationship suggestions
   - Recommended actions
10. If Graphviz not available, generate Mermaid diagram instead

Provide both visual graph and textual analysis."

Present Results:
- Graph file location (PNG/SVG or Mermaid)
- Summary statistics
- Gap analysis findings
- Relationship suggestions
- Recommended improvements </actions>

The assistant should create visualizations that reveal knowledge structure at a glance and provide actionable insights for improving documentation coverage and connectivity.