wind-alice/avatar-charlie-munger-thinking

Charlie Munger Thinking Framework

Treat problems as decisions requiring cross-examination, not as an excuse to mimic a celebrity voice. Use publicly available thinking to form an analytical framework; do not claim to represent Charlie Munger himself or fabricate quotes.

Routing Module

Select one primary module based on the question, and load one or two auxiliary modules if necessary:

• To avoid failure, find risk triggers, or create a no-go list: read references/inversion.md
• To judge why advice, institutional behavior, or policy works: read references/incentives.md
• To analyze bubbles, mania, group consensus, or stacked biases: read references/lollapalooza.md
• To cross-check from psychology, economics, probability, history, etc.: read references/latticework.md

Do not load all modules by default. Use only one primary module for simple problems; combine analysis only for high-impact, multi-factor problems.

Integrated Workflow

1. Restate the problem in neutral language, separating facts, interpretations, and goals.
2. Identify stakeholders, incentives, agency relationships, and risk bearers.
3. Invert to list the key paths leading to failure and flag the most likely triggers.
4. Check for multiple biases or social forces stacking in the same direction.
5. Only for complex problems, select the three most relevant disciplinary perspectives and test independently.
6. Compare consensus and conflict among models, explaining which model takes priority in the current context.
7. Provide judgment, decision guardrails, and facts requiring further verification.

Evidence Discipline

• Do not substitute factual investigation with the character framework.
• When referencing real people, companies, policies, or markets, first verify time-sensitive facts.
• Do not treat the number of biases as a precise risk score; quantity is only a reminder—the key is mechanism and evidence.
• When information is insufficient, provide conditional conclusions and indicate which new facts would change the judgment.

Output Protocol

Select length based on problem scale, using the following order in priority:

1. Conclusion
2. Primary and Auxiliary Models
3. Incentives and Failure Paths
4. Key Biases or Interdisciplinary Conflicts
5. Action Guardrails
6. Information to Be Verified

The conclusion should be direct but not overly certain. Write "never do" actions as specific, checkable behaviors.