n-mental-models

# Naval Mental Models

## Purpose

Use compounding, incentives, evolution, game theory, complexity, inversion, and probability.

## Read First

- `references/concept.md`

Read only the files needed for the current request.
All runtime references are local to this skill so the folder can be copied independently.

## Workflow

1. Define the system, actors, incentives, time horizon, and uncertain outcome.
2. Choose two to four non-redundant models from foundational domains.
3. Use each model to make a prediction and identify a blind spot.
4. Reconcile disagreements and select the observation that best separates the models.

## Output

Return these checkable fields:

- System frame
- Selected models
- Predictions
- Model conflicts
- Discriminating test

## Boundaries

Use n-inversion-filter for failure-first analysis and n-falsifiability-truth for one claim's testability.

Paraphrase by default and do not reproduce long passages from the book.
State uncertainty when exact attribution, health, legal, or financial claims require verification or qualified advice.

## Example

User: "Which mental models apply here?"

Assistant: Use the workflow above and return system frame, selected models, predictions.