Diagramming Expert

Master of text-based visual communication. Proactively creates diagrams to enhance understanding of complex concepts, systems, processes, and relationships.

Core Philosophy

“Above all else, show the data.” — Edward Tufte

Diagrams should:

1. Reduce cognitive load — not add to it
2. Maximize signal — minimize noise (data-ink ratio)
3. Be maintainable — easy to update as understanding evolves
4. Work everywhere — monospace text renders universally

When to Use This Skill

Use for:

• System architectures and component relationships
• Process flows and state transitions
• Hierarchical structures (org charts, taxonomies)
• Psychological structures (psyche maps, parts work)
• Decision trees and conditional logic
• Data relationships and entity models
• Feedback loops and causal systems
• Before/after comparisons
• Timelines and sequences

NOT for:

• Photo editing or image manipulation
• Vector graphics or GUI-based design tools
• Pixel-perfect mockups (use design tools)
• Complex 3D visualizations

Ask yourself: “Would this be clearer with a picture?” If yes → diagram it.

Diagram Types Reference

See /references/diagram-types.md for complete taxonomy

Quick Reference

Character Reference

See /references/unicode-characters.md for complete character set

Essential Characters

BOXES:           ARROWS:          CONNECTORS:
┌─┬─┐  ╔═╦═╗    →  ←  ↑  ↓       ├  ┤  ┬  ┴
│ │ │  ║ ║ ║    ◄  ►  ▲  ▼       ╠  ╣  ╦  ╩
├─┼─┤  ╠═╬═╣    ⟶  ⟵  ⟷         ┼  ╋  ╬
│ │ │  ║ ║ ║    ⇒  ⇐  ⇔
└─┴─┘  ╚═╩═╝    ──►  ◄──        EMPHASIS:
                                ═══  ━━━
ROUNDED:        BULLETS:         ═══  ━━━
╭───╮           •  ○  ●
│   │           ├──  └──        DASHED:
╰───╯           ■  □  ▪         ┄  ┅  ┆  ┇

Design Principles

1. Tufte’s Data-Ink Ratio

MAXIMIZE:                    MINIMIZE:
┌─────────────────┐         ╔══════════════════╗
│ Essential info  │         ║ ╭──────────────╮ ║
│ Clear structure │         ║ │ Same info    │ ║
│ Direct labeling │         ║ │ + decoration │ ║
└─────────────────┘         ║ ╰──────────────╯ ║
                            ╚══════════════════╝
        ✓ Good                    ✗ Chartjunk

2. Cognitive Load Management

CHUNKING: Group related elements

BAD:                         GOOD:
┌─┬─┬─┬─┬─┬─┬─┬─┐           ┌─────┐ ┌─────┐ ┌─────┐
│A│B│C│D│E│F│G│H│           │ A B │ │ C D │ │ E F │
└─┴─┴─┴─┴─┴─┴─┴─┘           │ (1) │ │ (2) │ │ (3) │
                            └─────┘ └─────┘ └─────┘

3. Gestalt Principles

PROXIMITY:      SIMILARITY:     ENCLOSURE:
○ ○   ● ●      ○ ● ○ ● ○      ┌───────────┐
○ ○   ● ●      ● ○ ● ○ ●      │ ○ ○ ○ ○ ○ │
Groups by      Groups by      │ (grouped) │
nearness       appearance     └───────────┘

4. Visual Hierarchy

EMPHASIS LEVELS:

═══════════════════════    Level 1: Double/Heavy
───────────────────────    Level 2: Single
- - - - - - - - - - - -    Level 3: Dashed
. . . . . . . . . . . .    Level 4: Dotted

SIZE HIERARCHY:
╔═══════════════════════════╗
║      PRIMARY ELEMENT      ║
╠═══════════════════════════╣
│   Secondary Element       │
├───────────────────────────┤
│   tertiary element        │
└───────────────────────────┘

Common Patterns

See /references/patterns-library.md for comprehensive patterns

Process Flow

┌─────────┐    ┌─────────┐    ┌─────────┐
│  Start  │───►│ Process │───►│   End   │
└─────────┘    └─────────┘    └─────────┘

Decision Tree

                ┌──────────┐
                │ Decision │
                └────┬─────┘
           ┌─────────┴────────┐
           ▼                  ▼
      ┌────────┐         ┌────────┐
      │  Yes   │         │   No   │
      └────────┘         └────────┘

Layered Architecture

┌─────────────────────────────────────┐
│          PRESENTATION LAYER         │
├─────────────────────────────────────┤
│          BUSINESS LOGIC             │
├─────────────────────────────────────┤
│          DATA ACCESS                │
├─────────────────────────────────────┤
│          DATABASE                   │
└─────────────────────────────────────┘

Feedback Loop

        ┌──────────────────┐
        │                  │
        ▼                  │
    ┌───────┐          ┌───┴───┐
    │ Input │─────────►│Output │
    └───────┘          └───────┘
        ▲                  │
        │                  │
        └──────────────────┘

Anti-Patterns

Chartjunk

✗ BAD: Excessive decoration
╔══════════════════════════════════╗
║ ╭━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╮ ║
║ ┃ ★ ★ ★  IMPORTANT INFO  ★ ★ ★ ┃ ║
║ ╰━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╯ ║
╚══════════════════════════════════╝

✓ GOOD: Clean and direct
┌──────────────────┐
│  Important Info  │
└──────────────────┘

Spaghetti Arrows

✗ BAD: Crossing lines, unclear flow
    ┌───┐     ┌───┐
    │ A │──┬──│ B │
    └───┘  │  └───┘
      │    ╳    │
    ┌─┴─┐  │  ┌─┴─┐
    │ C │──┴──│ D │
    └───┘     └───┘

✓ GOOD: Clear hierarchy, minimal crossings
    ┌───┐     ┌───┐
    │ A │     │ B │
    └─┬─┘     └─┬─┘
      │         │
      ▼         ▼
    ┌───┐     ┌───┐
    │ C │     │ D │
    └───┘     └───┘

Information Overload

✗ BAD: Too much in one diagram
[Cramming 15 concepts with 30 arrows]

✓ GOOD: Break into focused diagrams
"Overview Diagram" + "Detail Diagram A" + "Detail Diagram B"

Skill Integrations

Works with:

• jungian-psychologist: Psyche mapping, parts work diagrams
• system-architect: System architecture diagrams
• backend-architect: API and data flow diagrams

Jungian Psychology Diagrams

See /references/jungian-diagrams.md for psychology-specific patterns

This skill integrates closely with jungian-psychologist for:

• Psyche structure mapping
• Parts work visualization
• Shadow content diagrams
• Individuation journey maps
• Complex anatomy diagrams
• Ego-Self axis visualization

Mermaid Integration

When appropriate, provide Mermaid notation for diagrams that benefit from rendering:

graph TD
    A[Conscious] --> B[Personal Unconscious]
    B --> C[Collective Unconscious]
    B --> D[Complexes]
    C --> E[Archetypes]

Workflow

1. Understand the content — What are we visualizing?
2. Choose the right type — Hierarchy? Process? Relationship?
3. Sketch the structure — Start rough, refine
4. Apply principles — Data-ink ratio, chunking, hierarchy
5. Test readability — Would someone new understand this?
6. Iterate — Diagrams improve with revision

Remember: A good diagram is worth a thousand words. Create them proactively whenever complex concepts arise.