causal-loop-microsim-generator

This skill generates interactive Causal Loop Diagram (CLD) MicroSims using the vis-network JavaScript library. Use this skill when users need to create causal loop diagrams for systems thinking education, showing feedback loops, reinforcing and balancing dynamics. The skill creates a complete MicroSim package with index.md, main.html, JavaScript, JSON data, and CSS files in the /docs/sims/ directory. This skill should be used when users request creating CLDs, causal diagrams, feedback loop visualizations, or systems thinking diagrams.

$ Instalar

git clone https://github.com/dmccreary/claude-skills /tmp/claude-skills && cp -r /tmp/claude-skills/skills/archived/causal-loop-microsim-generator ~/.claude/skills/claude-skills

// tip: Run this command in your terminal to install the skill


name: causal-loop-microsim-generator description: This skill generates interactive Causal Loop Diagram (CLD) MicroSims using the vis-network JavaScript library. Use this skill when users need to create causal loop diagrams for systems thinking education, showing feedback loops, reinforcing and balancing dynamics. The skill creates a complete MicroSim package with index.md, main.html, JavaScript, JSON data, and CSS files in the /docs/sims/ directory. This skill should be used when users request creating CLDs, causal diagrams, feedback loop visualizations, or systems thinking diagrams.

Causal Loop MicroSim Generator

Overview

This skill generates interactive Causal Loop Diagram (CLD) MicroSims for educational purposes in systems thinking. CLDs visualize cause-and-effect relationships, feedback loops, and system dynamics using nodes (variables) and edges (causal relationships with positive or negative polarity).

When to Use This Skill

Use this skill when users request:

  • Creating a causal loop diagram
  • Visualizing feedback loops
  • Building systems thinking diagrams
  • Generating CLD visualizations
  • Creating reinforcing or balancing loop diagrams
  • Building system dynamics visualizations

Workflow

Step 1: Gather Requirements

Collect the following information from the user:

  1. MicroSim name (kebab-case, e.g., ai-flywheel, climate-feedback)
  2. Title for the diagram
  3. Description of the system being modeled
  4. Nodes (variables in the system) with their labels and descriptions
  5. Edges (causal relationships) with polarity (positive/negative)
  6. Loops (reinforcing R or balancing B) with descriptions

If the user provides a text description, parse it to identify:

  • Key variables (become nodes)
  • Causal relationships (become edges with polarity)
  • Feedback loops (reinforcing or balancing)

Step 2: Generate the MicroSim Files

Create the following directory structure in /docs/sims/{{MICROSIM_NAME}}/:

{{MICROSIM_NAME}}/
├── index.md           # Documentation page
├── main.html          # HTML container
├── {{MICROSIM_NAME}}.js   # JavaScript code using vis-network
├── data.json          # Node and edge data
└── style.css          # Custom CSS styles

Step 3: File Generation Details

3.1 data.json

Generate the JSON data file following the CLD schema. Refer to assets/rules.md for the complete JSON schema and best practices.

Key structure:

{
  "metadata": {
    "id": "{{MICROSIM_NAME}}-cld",
    "title": "Title",
    "archetype": "archetype-name",
    "description": "Description",
    "version": "1.0.0"
  },
  "nodes": [...],
  "edges": [...],
  "loops": [...]
}

Node positioning guidelines:

  • Canvas center is approximately (300, 300)
  • Space nodes 150-200 pixels apart
  • Arrange nodes in a logical flow (clockwise for reinforcing, counter-clockwise for balancing)
  • For 4-node loops: use positions like (300,150), (450,300), (300,450), (150,300)

3.2 main.html

Create the HTML file using the template in assets/templates/main.html. The HTML should:

  • Load vis-network from CDN
  • Include the CSS file
  • Reference the JavaScript file
  • Have a container div for the network

3.3 {{MICROSIM_NAME}}.js

Generate JavaScript using vis-network library. Refer to assets/templates/microsim.js for the template.

Key features to implement:

  • Load data from data.json
  • Configure node appearance (box shape, colors, fonts)
  • Configure edge appearance (arrows, polarity colors: green for +, red for -)
  • Disable physics for manual positioning
  • Add click handlers for showing details
  • Support URL parameters for iframe embedding

3.4 style.css

Create CSS for the MicroSim layout. Use the template in assets/templates/style.css.

3.5 index.md

Create the documentation page with:

  • Title and description
  • Learning objectives
  • Iframe embed of the MicroSim
  • Link to full-screen version
  • Explanation of the system dynamics

Step 4: Update mkdocs.yml

Add the new MicroSim to the navigation in mkdocs.yml:

  1. Find the MicroSims: section in the nav
  2. Add a new entry in alphabetical order: - {{Title}}: sims/{{MICROSIM_NAME}}/index.md

Important: Maintain alphabetical ordering of all MicroSim entries.

Step 5: Remind User About Screenshot

After generating all files, remind the user:

Screenshot Required: Please take a screenshot of the MicroSim and save it as {{MICROSIM_NAME}}.png in the /docs/sims/{{MICROSIM_NAME}}/ directory. This image will be used for social sharing and documentation.

CLD Design Best Practices

Refer to assets/rules.md for detailed rules on:

  • JSON schema specification
  • Node positioning algorithms
  • Edge polarity and curve directions
  • Loop labeling conventions
  • vis-network configuration options

Resources

assets/

  • rules.md - Comprehensive CLD generation rules and JSON schema
  • templates/main.html - HTML template
  • templates/microsim.js - JavaScript template
  • templates/style.css - CSS template
  • templates/index.md - Documentation template
  • templates/data.json - Example JSON data structure

Example Usage

User request: "Create a CLD showing how increased AI usage leads to more training data, which improves model accuracy, which increases AI usage."

Generated MicroSim:

  • Name: ai-usage-loop
  • Nodes: AI Usage, Training Data, Model Accuracy
  • Edges: All positive polarity forming a reinforcing loop
  • Loop: R - AI Improvement Cycle