mental-models（心理模型）

Mental Models

Apply structured thinking frameworks to decisions, problems, and planning.

When This Triggers

• "Help me think through X"
• "What mental model should I use?"
• "Apply first principles to this"
• "Do a pre-mortem on this plan"
• "I need to analyze this decision"
• Any specific model name (SWOT, 5 Whys, etc.)

Workflow

Step 1: Run Mental Models Scanner

Run the script to get all available models:

python 00-system/mental-models/scripts/select_mental_models.py --format brief

This returns JSON with all 59 models across 12 categories.

Optional filters:

# Filter by category
python 00-system/mental-models/scripts/select_mental_models.py --category cognitive --format brief

# List format (names only, grouped by category)
python 00-system/mental-models/scripts/select_mental_models.py --format list

Step 2: Identify Context and Offer Options

Based on user's situation, offer 2-3 relevant models:

For decisions: Decision Matrix, Pre-Mortem, Cost-Benefit, Inversion For problems: First Principles, Root Cause (5 Whys), Fishbone For planning: Scenario Planning, Stakeholder Mapping, OKR For creativity: Design Thinking, SCAMPER, Lateral Thinking For risk: Pre-Mortem, Force Field, Red Team, Black Swan For communication: Pyramid Principle, BLUF, Steel Manning For learning: Feynman Technique, Deliberate Practice

Present options with brief descriptions (3-7 words each).

Step 3: Load Specific Model File

After user selects, load the individual model file:

File structure: 00-system/mental-models/models/{category}/{model-slug}.md

Example:

User picks: "First Principles + Pre-Mortem"

AI loads:
→ Read: 00-system/mental-models/models/cognitive/first-principles.md
→ Read: 00-system/mental-models/models/diagnostic/pre-mortem.md

Step 4: Apply Model Questions

Use the question templates from the loaded model file to guide the user through structured thinking.

Each model file contains:

• Purpose: What the model does
• When to Use: Best situations for this model
• Questions to Ask: Ready-to-use prompts
• Process: Step-by-step application
• Output: What you get from using it

Keep it collaborative - this is a conversation, not an interrogation.

Quick Reference

12 Categories (59 Models):

1. Cognitive - First Principles, Systems Thinking, Lateral Thinking, Inversion, Second-Order Thinking, Analogous Reasoning
2. Collaborative - Six Hats, MECE, Stakeholder Mapping, Devil's Advocate
3. Diagnostic - 5 Whys, Fishbone, Pre-Mortem, Force Field, Fault Tree
4. Strategic - Scenario Planning, OODA, Jobs to Be Done, Blue Ocean, PESTLE, Porter's Five Forces
5. Analytical - Decision Matrix, SWOT, Cost-Benefit, Pareto, Assumption Testing, Sensitivity Analysis
6. Creative - Design Thinking, SCAMPER, Morphological, Random Entry, Constraint Removal
7. Operational - Kanban, Value Stream, OKR, Lean Canvas, Theory of Constraints
8. Validation - Hypothesis Testing, Prototyping, Red Team, A/B Testing
9. Time & Resource - Eisenhower Matrix, Time Boxing, Opportunity Cost, Sunk Cost, Resource Mapping
10. Communication - Pyramid Principle, BLUF, Situation-Complication-Resolution, Steel Manning
11. Learning - Feynman Technique, Spaced Repetition, Deliberate Practice, T-Shaped Skills
12. Probability & Risk - Expected Value, Margin of Safety, Black Swan, Bayesian Updating, Regret Minimization

Notes

• Always offer choice, never prescribe
• Run script first to see all available models
• Load individual model files only after user selects
• Combine models when appropriate (e.g., First Principles + Pre-Mortem)
• Adapt formality to user's context