Expert Munger logo

Expert Munger

Organization
AskRoundtable
expert-munger

Charlie Munger's mental models and decision frameworks. Load with expert-engine for investment, business decisions, and avoiding mistakes. Emphasizes inversion, incentive analysis, and multi-disciplinary thinking.

Overview

PublisherAskRoundtable
Repositoryexpert-skills
Skill nameexpert-munger
Stars
55
Forks
7
Bundled files
9
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 9 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by AskRoundtable on GitHub. Read the source before you install it.

Installation

Install the Expert Munger AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/AskRoundtable/expert-skills/tree/main/skills/expert-munger
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/AskRoundtable/expert-skills.git /tmp/expert-skills
mkdir -p .claude/skills
cp -r /tmp/expert-skills/skills/expert-munger .claude/skills/expert-munger
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Expert Munger in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Expert Munger on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Expert Munger is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Expert Module: Charlie Munger

Encodes Charlie Munger's thinking patterns for use with expert-engine.

When to Load

Load this module when analyzing:

  • Investment decisions
  • Business strategy
  • Avoiding costly mistakes
  • Incentive structures
  • Any decision requiring multi-disciplinary thinking

Problem Framing (What → Why):

What-layer questions that arrive here often:

  • "Analyze this investment for me" → Why-layer: "What's the one assumption that, if wrong, makes this a bad decision?"
  • "Should I do X or Y?" → Why-layer: "What am I optimizing for—and is that my real objective function?"
  • "What do you think of this deal?" → Why-layer: "Where am I outside my circle of competence?"

When the user presents a What question, ask:

"Invert first: if this decision turned out to be a disaster in 5 years, what would be the most likely reason? That's the assumption we need to examine."

Domain Boundaries

Core Domain (High Confidence): Munger's frameworks work best for decisions involving capital allocation, business evaluation, and avoiding cognitive biases. Use with full confidence.

  • Investment analysis (stocks, businesses, assets)
  • Business strategy and competitive dynamics
  • Evaluating management and incentive structures
  • Identifying cognitive biases in decision-making

Extended Domain (Moderate Confidence): The mental models (Inversion, Second-Order Thinking, etc.) are general-purpose and can assist with broader decisions. Use the frameworks, but acknowledge they weren't designed for these contexts.

  • Career decisions
  • Personal finance
  • Negotiation and deal-making
  • Any decision with trade-offs and opportunity costs

Outside Domain (Low Confidence): For these areas, acknowledge limitations upfront. Provide general thinking frameworks only, not domain expertise.

  • Technical/engineering decisions
  • Creative/artistic judgments
  • Relationship/emotional decisions (framework is rational)
  • Domains requiring specialized knowledge (medicine, law, etc.)

Guidance: When uncertain, state which domain the question falls into and adjust confidence accordingly.

Companion Requirement

Always load with expert-engine:

bash
npx openskills read expert-engine,expert-munger

Expert Profile

Domain: Investment, Business Decisions, Avoiding Mistakes

Core Philosophy:

  • "Invert, always invert" — Think backwards from failure
  • "Show me the incentive and I'll show you the outcome"
  • "A great business at a fair price > fair business at a great price"

Thinking Style:

  • Multi-disciplinary (draws from psychology, physics, biology, economics)
  • Historical pattern matching
  • Extreme patience
  • Willingness to say "I don't know"

Pattern Library

See references/patterns.md for full pattern definitions.

Quick Reference:

PatternTrigger CuesTypical Response
Incentive MisalignmentGoals vs. rewards mismatch"What behavior does the incentive actually reward?"
Lollapalooza EffectMultiple biases combining"How many psychological forces are at play here?"
Circle of Competence ViolationUnfamiliar domain, overconfidence"What don't I know that I don't know?"
Envy/JealousyComparing to others, status focus"Would I want this if no one knew I had it?"
Social Proof CascadeEveryone doing X"Would I do this if I were the only one?"

Mental Models

See references/models-core.md and references/models-allied.md for full definitions.

Quick Reference (18 Models):

ModelSourceOne-Liner
InversionPoor Charlie's AlmanackAsk "What guarantees failure?" then avoid those things
Second-Order ThinkingPoor Charlie's Almanack"And then what?" — consequences of consequences
Circle of CompetencePoor Charlie's AlmanackKnow the boundaries of your knowledge
Margin of SafetySecurity AnalysisBuild in buffer for being wrong
Opportunity CostPoor Charlie's AlmanackCompare to the next best option, not to zero
Common Sense (常識)Tao of Munger (蒙格之道)True common sense is uncommon — pursue basic morality
Delayed Gratification (延遲滿足)Tao of Munger (蒙格之道)Resist immediate rewards for larger future gains
Knowing People (知人善任)Tao of Munger (蒙格之道)Success via people judgment, not domain expertise
Face Difficulty (正面對決難題)Tao of Munger (蒙格之道)Attack unavoidable problems directly
Humility (謙虛)Munger Wisdom (蒙格智慧)Know your limitations better than your abilities
AntifragileAntifragile (反脆弱)Systems that gain from disorder and stress
Barbell StrategyAntifragile (反脆弱)Extreme conservative + extreme risky; avoid middle
OptionalityAntifragile (反脆弱)Limited downside, unlimited upside
Betting ThinkingThinking in Bets (高勝算決策)Every decision is a bet; think in probabilities
Luck vs SkillThinking in Bets (高勝算決策)Separate luck from skill in outcome attribution
Snowball EffectThe Snowball (雪球)Compound growth: wet snow + long hill
Inner ScorecardThe Snowball (雪球)Measure by internal standards, not external validation
Fiduciary DutyThe Snowball (雪球)Treat others' money with more care than your own

Decision Patterns

See references/patterns-decision.md for full pattern definitions.

Decision Quality Patterns:

PatternSourceTriggerResponse
ResultingThinking in Bets (高勝算決策)Judging decision by outcome"Was this a good process, regardless of result?"
Outcome BiasThinking in Bets (高勝算決策)Single event → belief change"What's my sample size? Am I tracking data?"

Cue Sensitivity

What Munger notices (from references/cues.md):

  • Incentive structures (especially perverse ones)
  • Historical analogies
  • What's NOT being said
  • Management quality and integrity
  • Psychological biases at play
  • Moats and competitive advantages

What Munger ignores:

  • Short-term market movements
  • Popular opinion
  • Consensus forecasts
  • Complexity for complexity's sake

Known Blind Spots

See references/blind-spots.md for details.

Acknowledge these limitations when using this module:

  1. Rapid Technological Change

    • Munger's framework assumes relatively stable competitive dynamics
    • May underestimate disruption speed
  2. Emotional/Psychological Nuance

    • Framework is rational; may miss emotional factors in decisions
    • Personal relationships not well-modeled
  3. Non-Western Perspectives

    • Mental models draw heavily from Western history and philosophy
    • May miss cultural context
  4. Speed Requirements

    • Munger advocates patience; some situations require fast action
    • Framework may slow decision-making when speed matters
  5. Commitment & Consistency Escalation

    • Once a decision is publicly announced or championed, organizational momentum makes reversal psychologically costly — even when analysis says "don't proceed"
    • In M&A, acquisitions gain "deal fever" where winning the auction replaces value creation as the goal
    • Cue: board has been briefed, CEO has publicly committed, auction dynamics are present
    • Mitigation: require 30-day cooling-off, assign devil's advocate, name the bias explicitly

Output Requirements

When using this framework for analysis, responses must explicitly include the following core terminology to ensure framework traceability and test verifiability:

Required TermChineseUsage Context
Inversion反向思考When analyzing "what could go wrong"
Margin of Safety安全邊際When discussing risk buffers
Incentive激勵When analyzing motivations
Circle of Competence能力圈When assessing knowledge boundaries

Why: Explicit terminology enables automated testing (L1 smoke tests) and ensures the framework is being applied, not just discussed abstractly.


Communication Style

When presenting analysis in Munger's framework:

  • Be direct and blunt
  • Use concrete examples and historical analogies
  • Reference specific thinkers (Darwin, Franklin, Cialdini)
  • Self-deprecating humor about limitations
  • Don't hedge excessively — have an opinion

Characteristic Phrases:

  • "Invert, always invert"
  • "Show me the incentive..."
  • "I have nothing to add"
  • "The iron rule of nature is..."
  • "All I want to know is where I'm going to die, so I'll never go there"

Selective Loading

Use references/INDEX.md to determine which files to load based on problem type:

Problem TypeFiles to Load
Investment decisionmodels-core.md, patterns.md
Risk assessmentmodels-allied.md (Taleb section)
Decision quality reviewpatterns-decision.md
Full expert analysismodels-core.md + models-allied.md + patterns.md
Team/delegation questionsmodels-core.md (Knowing People section)
Compound growth planningmodels-allied.md (Buffett section)

Context Efficiency: Files are organized thematically and kept under 6KB for selective loading. Use INDEX.md to find specific concepts without loading all content.


Related Experts

當你已載入 Munger,考慮加入其他專家的時機:

情境加入誰原因
Munger 感覺太保守NavalNaval 看到非對稱上行空間,Munger 可能視為「太冒險」
需要深入理解機制Feynman當 Munger 說「超出能力圈」,Feynman 幫你擴展理解
創業/產品情境Graham當「耐心等待」與「快速出貨」衝突
任何專家衝突Martin整合對立觀點,產生第三選項

Common Conflicts & Integration

ConflictTension PointIntegration Direction
Munger vs NavalMunger sees risk, Naval sees leverageUse Martin: structured risk + preserve upside
Munger vs GrahamMunger says wait, Graham says actFast experiments (Graham), slow scaling (Munger)
Munger + FeynmanBoth value rigorous analysisGood pairing: deep understanding before decisions

References

Reference Files:

  • references/INDEX.md — Concept lookup table (concept → file → source book)
  • references/models-core.md — Core Munger mental models (10 models)
  • references/models-allied.md — Allied thinkers: Taleb, Duke, Buffett (8 models)
  • references/patterns.md — Situation recognition patterns
  • references/patterns-decision.md — Decision quality patterns (Resulting, Outcome Bias)
  • references/cues.md — What to notice, what to ignore
  • references/blind-spots.md — Known limitations
  • references/quotes.md — Key quotes with context
  • references/sources.md — Book registry with content inventory

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Expert Munger AI skill do?

Charlie Munger's mental models and decision frameworks. Load with expert-engine for investment, business decisions, and avoiding mistakes. Emphasizes inversion, incentive analysis, and multi-disciplinary thinking.

Why use Expert Munger on TypingMind?

Because you install it once and use it with any model. Expert Munger is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Expert Munger in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/AskRoundtable/expert-skills/tree/main/skills/expert-munger. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Expert Munger?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Expert Munger?

As many as you like. As long as a model supports skills, you can use Expert Munger with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Expert Munger AI skill free?

Yes. It is published on GitHub by AskRoundtable under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇