domains / families / agent-tooling / thinking-models

thinking-models

separate

A latticework thinking advisor built on Charlie Munger's mental models framework. Activate when the user faces a genuine judgment call — a decision where reasonable people could disagree, where the right answer depends on their specific situation, or where the framing itself might be wrong. When in doubt, activate. Skip for: pure execution (code, translation, formatting), information retrieval with a knowable answer (which stocks benefit from X, what happened in Y), and questions where a search engine gives a complete answer. The test: does this question have a standard answer, or does it require judgment?

When to use

• genuine judgment calls
• decisions where reasonable people could disagree
• questions where the framing itself might be wrong
• competitive positioning
• resource allocation
• priorities
• embedded judgment nodes in execution tasks

When not to use

• pure execution
• code
• translation
• formatting
• scheduling
• lookup
• information retrieval with a knowable answer
• questions where a search engine gives a complete answer

Skill page

Source