This project provides a "cognitive operating system" for AI agents, based on the principles and mental models of Charlie Munger. It aims to move beyond simple quote aggregation by offering a runnable framework that applies Munger's decision-making heuristics and thinking patterns to user queries, enabling more nuanced and wisdom-informed AI responses. The target audience includes users of AI agents compatible with the Agent Skills protocol seeking to leverage Munger's analytical approach.

How It Works

The skill distills over 50 sources, including Munger's books, speeches, and interviews, to extract 5 core mental models, 8 decision heuristics, and a distinct "expression DNA." It employs Munger's specific frameworks, such as inverse thinking, the Lollapalooza effect, incentive mechanism diagnosis, circle of competence, and the three-bucket classification system, to analyze problems and generate responses. The core innovation lies in its functional application of these models within an AI agent, rather than merely reciting Munger's quotes.

Quick Start & Requirements

• Installation: Recommended: npx skills add alchaincyf/munger-skill. Manual installation via git clone is also supported for specific runtimes (e.g., Claude Code, Codex CLI, Cursor, OpenClaw, Hermes Agent).
• Prerequisites: Requires an AI agent runtime compatible with the open Agent Skills protocol (e.g., Claude Code, Codex, Cursor, OpenClaw, Hermes Agent, CodeBuddy, Gemini CLI).
• Links: Agent Skills, skills.sh.

Highlighted Details

• Implements core Munger mental models: Multi-disciplinary thinking, inverse thinking, Lollapalooza effect, circle of competence, and incentive structures.
• Features 8 decision heuristics, including the "three-bucket" approach (Yes/No/Too Hard) and "avoiding stupidity."
• Captures Munger's "expression DNA," including sentence structure, vocabulary, analogies, and dry humor.
• Acknowledges Munger's "internal tensions," such as selective rationality (e.g., on crypto/AI) and potential tech blindness.
• Generated using the nuwa-skill framework.

Maintenance & Community

The project is developed by 花叔 (Huashu), an AI Native Coder. Related projects and other distilled skills (e.g., Steve Jobs, Elon Musk, Naval Ravikant) are available. Contact and project links include official websites (bookai.top, huasheng.ai), Twitter (@AlchainHust), Bilibili, YouTube, and Xiaohongshu.

Licensing & Compatibility

The skill is released under the MIT License, permitting free use, modification, and distribution, including for commercial purposes and integration into closed-source applications.

Limitations & Caveats

The project implicitly inherits Munger's known biases and limitations, such as "selective rationality" regarding certain technologies (AI, crypto) and potential "tech blindness." The effectiveness of the skill is dependent on the capabilities and implementation of the host AI agent runtime. The README notes Munger's "internal tensions," including significant investment losses (e.g., Alibaba) and ideological stances that may represent areas of cognitive limitation.