A Codex Skill for applying the MungerModels.com Charlie Munger mental-model knowledge graph to decisions, investments, business analysis, risk reviews, organization design, and learning plans.
This is a graph-based analysis skill, not a pure roleplay prompt. It helps Codex select and combine relevant mental models across disciplines, then turn them into diagnostic questions, failure modes, and practical checklists.
SKILL.md- skill trigger metadata and operating workflow.references/graph-analysis.md- parsed graph structure, central models, scenario playbooks, and source URL patterns.agents/openai.yaml- UI metadata for Codex skill discovery.
The skill is based on the public Munger Models knowledge graph:
- 232 mental models
- 14 disciplines
- 9 application scenarios
- 1,757 graph edges on the graph page
It highlights central models such as 安全边际, 护城河, 逆向思维, 能力圈, 社会认同倾向, Lollapalooza倾向, 复利效应, and 概率思维与期望值.
Clone this repository into your Codex skills directory:
git clone https://github.com/MichaelRochonnn/munger-models.git ~/.codex/skills/munger-models-graphThen start a new Codex session so the skill index can refresh.
Example prompts:
Use $munger-models-graph to analyze whether this business has a durable moat.
用 $munger-models-graph 帮我复盘这个投资决策,重点看安全边际、能力圈和失败模式。
用芒格思维模型图谱分析这个组织激励设计有没有代理问题。
For concrete decisions, the skill typically produces:
- Verdict
- Model Stack
- Cross-Disciplinary Read
- Inversion / Failure Modes
- Checklist
- Confidence / Too-hard boundary
Primary source map:
- Homepage: https://mungermodels.com/
- Full index: https://mungermodels.com/all
- Graph: https://mungermodels.com/graph
- Model detail pages:
https://mungermodels.com/models/{model-id} - Scenario pages:
https://mungermodels.com/scenarios/{scenario-id}
When exact model details or quotes matter, open and cite the relevant source page instead of relying only on the bundled summary.
If you have the Codex skill-creator validation script available, run:
python3 ~/.codex/skills/.system/skill-creator/scripts/quick_validate.py ~/.codex/skills/munger-models-graphExpected result:
Skill is valid!