Skip to content

About

A comprehensive open-source latticework of mental models for human decision-making and AI agent reasoning.

Resources

Contributing

Stars

1 star

Watchers

0 watching

Forks

Latest commit

Β 

History

4 Commits

Folders and files

NameName
Last commit message
Last commit date
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

The Mental Models Latticework

License: MIT Models Architecture MCP CI Validation

"You can't really know anything if you just remember isolated facts and try and bang 'em back. If the facts don't hang together on a latticework of theory, you don't have them in a usable form."

An open-source, multidisciplinary knowledge base and decision-support system synthesizing foundational models across physics, evolutionary biology, economics, cognitive psychology, operations research, and mathematics.

Designed from first principles to serve both human decision-makers and autonomous AI agents without fragmentation.


🧭 The 5 Pillars of Utility

flowchart TD
    subgraph Truth ["1. Core Truth (Single Source)"]
        M["models/**/*.md\n(Rigorous Markdown + YAML Frontmatter)"]
    end

    subgraph Tooling ["Tooling & Interface Layer"]
        CLI["CLI Interface (cli/mm)\nFast terminal lookup & diagnosis"]
        AGENT["AI Skill & MCP Server\n(skills/ & mcp/)\nChain-of-thought protocols"]
        GRAPH["Interactive Knowledge Graph\n(docs/graph.html)\nD3.js & Obsidian Wikilinks"]
        ANKI["Spaced Repetition (Anki)\n(scripts/export_anki.py)\nAutomated flashcard decks"]
        MATRIX["Problem-to-Model Matrix\nActionable decision playbooks"]
    end

    subgraph Users ["End Beneficiaries"]
        H["Human Decision Makers\n(Engineers, Founders, Leaders)"]
        AI["Autonomous AI Agents\n(Antigravity, Claude, Cursor)"]
        K["Knowledge Graphs\n(Obsidian, Logseq)"]
    end

    M --> CLI --> H
    M --> MATRIX --> H
    M --> AGENT --> AI
    M --> GRAPH --> K & H
    M --> ANKI --> H
Loading

1. The Human Decision Playbook & Problem-to-Model Matrix

Every model file contains a field-tested Diagnostic Checklist and Real-World Case Studies (Technology, Business, High-Stakes Decisions). When facing a dilemma, consult the matrix below:

If You Are Facing... Consult Primary Model Complementary Pair Counter-Balance
Deleting or refactoring legacy systems [[Chesterton's Fence]] [[First-Principles Thinking]] [[Second-Order Thinking]]
Big-bang rewrites vs incremental evolution [[Gall's Law]] [[Theory of Constraints]] [[Chesterton's Fence]]
High-stakes risk with asymmetric downside [[Ergodicity & Absorbing Barriers]] [[Inversion]] [[Margin of Safety]]
Runaway positive or negative spirals [[Feedback Loops]] [[Network Effects]] [[Theory of Constraints]]
Prioritizing roadmap or capital allocation [[Opportunity Cost]] [[Expected Value]] [[Comparative Advantage]]
Resource allocation across fat tails [[Power Laws & The Pareto Principle]] [[Asymmetric Payoffs]] [[Theory of Constraints]]
Slow organizational delivery & backlogs [[Theory of Constraints]] [[Feedback Loops]] [[Leverage]]
Interpersonal friction or vendor breakages [[Hanlon's Razor]] [[Occam's Razor]] [[Proximate vs Root Cause]]
Rapid competitor adaptation or arms race [[The Red Queen Effect]] [[The OODA Loop]] [[Economic Moats]]
Hesitation to cut losses on failing projects [[Sunk Cost Fallacy]] [[Loss Aversion]] [[Opportunity Cost]]
Uncertain forecasts or planning fallacies [[Base Rate Fallacy]] [[Map vs Territory]] [[Anchoring & Adjustment]]
Designing systems under hostile volatility [[Antifragility]] [[Defense in Depth]] [[Margin of Safety]]
Shared platform or multi-tenant degradation [[Tragedy of the Commons]] [[Goodhart's Law]] [[Principal-Agent Problem]]
Navigating live crises under incomplete data [[The Fog of War]] [[The OODA Loop]] [[Margin of Safety]]
Hiring, credentials, and counterfeit claims [[Costly Signaling & The Handicap Principle]] [[Circle of Competence]] [[Principal-Agent Problem]]

2. AI Agent Reasoning Engine (Skill & MCP Server)

Models are machine-readable with strict YAML frontmatter and dedicated AI Agent Reasoning Protocols.

  • Antigravity Skill: Located at skills/mental-models/SKILL.md. Equip your agent to run multi-model debiasing passes before major refactors.
  • Model Context Protocol (MCP): Run mcp/server.py over stdio with Claude Desktop, Cursor, or Antigravity:
    {
      "mcpServers": {
        "mental-models": {
          "command": "python3",
          "args": ["/path/to/mental-models/mcp/server.py"]
        }
      }
    }

3. "The Latticework" Knowledge Graph

Models do not exist in isolation. Every file connects to others via standard wikilinks ([[Model Name]]), compatible with Obsidian and Logseq.

  • Interactive Force-Directed Graph: Open docs/graph.html directly in any browser for an interactive D3.js visualization showing 100 interconnected nodes and 235 cross-disciplinary links.
  • Regenerate Graph:
    python3 scripts/generate_graph.py

4. Zero-Dependency Terminal CLI (mm)

Quickly search, diagnose, and review models from your terminal with zero external pip dependencies:

# List all 100 models organized by category
./cli/mm list

# Search models across titles, summaries, and triggers
./cli/mm search "uncertainty"

# Diagnose a strategic or technical problem statement
./cli/mm diagnose "Our team wants to replace a legacy billing service"

# Show complete details, checklist, and triggers for a model
./cli/mm show chestertons-fence

# Pull a random model for daily reflection
./cli/mm random

5. Spaced Repetition (Anki Flashcards)

Internalize models into long-term intuition. The automated export script extracts conceptual mechanisms and diagnostic scenarios into ready-to-import Anki cards:

python3 scripts/export_anki.py

Outputs exports/mental_models_anki.tsv containing 200 flashcards ready for 1-click import into Anki Desktop / Mobile.


πŸ“š Categorized Library Index (100 Models)

🧠 Core Thinking & Reasoning (18 Models)

  • Abductive Reasoning (Inference to the Best Explanation): The form of logical inference that starts with an incomplete set of observations and forms the most likely, parsimonious,...
  • Chesterton's Fence: The principle that reforms, deletions, or removals of existing structures, rules, or code should never be made until the...
  • Circle of Competence: The practice of identifying the realistic boundaries of your genuine knowledge and expertise, ruthlessly operating within...
  • Falsifiability: The principle that for any hypothesis, theory, or statement to be considered genuinely scientific or predictive, it must be...
  • First-Principles Thinking: The practice of deconstructing a problem down to its most fundamental, irreducible truths and building up a solution from...
  • Hanlon's Razor: Never attribute to malice that which is adequately explained by carelessness, ignorance, incompetence, or cognitive overload.
  • Inversion: The practice of approaching problems backwards, focusing on avoiding failure, catastrophe, and stupidity rather than...
  • Lateral Thinking: The practice of solving problems through an indirect, non-linear, and creatively diagonal approach, deliberately stepping...
  • The Lindy Effect: The theory that the future life expectancy of non-perishable entities (technologies, books, ideas, institutions) is...
  • Map vs. Territory: The fundamental realization that an abstraction, model, diagram, or metric of reality is not the reality itself, and that...
  • Necessary vs. Sufficient Conditions: The fundamental logical distinction between a condition that must be present for an event to occur (necessary) and a...
  • Occam's Broom: The intellectually dishonest practice where an advocate or theorist subtly sweeps inconvenient, contradictory facts and...
  • Occam's Razor: Among competing hypotheses or design solutions that predict or achieve identical outcomes, the one requiring the fewest...
  • Proximate vs. Root Cause: The crucial distinction between the immediate, surface event that triggered a failure (proximate cause) and the deeper,...
  • Reductio ad Absurdum: The method of disproving a proposition or architectural claim by demonstrating that its logical consequences, when followed...
  • Second-Order Thinking: The discipline of evaluating decisions not just by their immediate, intended consequences, but by the subsequent ripple...
  • Steelmanning (The Principle of Charity): The practice of constructing the strongest, most compelling possible version of your opponent's or alternative's argument...
  • Thought Experiments (Gedankenexperiment): The practice of exploring the implications of an idea, hypothesis, or paradox using pure disciplined mental visualization...

πŸ”„ Systems & Complexity (15 Models)

  • Antifragility: The property of systems that increase in capability, resilience, or robustness as a consequence of stressors, volatility,...
  • Ashby's Law of Requisite Variety: The cybernetic law stating that for a control system or organization to maintain stability and deal effectively with...
  • Braess's Paradox: The counter-intuitive observation that adding extra capacity, links, or routes to a congested, decentralized network can...
  • The Bullwhip Effect: The phenomenon where small fluctuations in retail customer demand become progressively and exponentially amplified as...
  • The Cobra Effect (Perverse Incentives): The systemic phenomenon where an attempted solution, reward, or incentive designed to solve a problem inadvertently rewards...
  • Emergence & Self-Organization: The phenomenon where a collection of simple agents interacting according to basic local rules gives rise to complex,...
  • Feedback Loops (Reinforcing & Balancing): The circular causality where the output of a system circles back to either amplify (reinforcing) or stabilize (balancing)...
  • Gall's Law: The rule of thumb that a complex system that works is invariably found to have evolved from a simple system that worked;...
  • Goodhart's Law: When a measure becomes a target, it ceases to be a good measure, because actors will optimize to maximize the target while...
  • Homeostasis vs. Allostasis: The distinction between maintaining stability by holding internal states constant around a rigid setpoint (homeostasis)...
  • Hysteresis & Path Dependence: The property of physical and human systems where the current state depends not only on the current environment and inputs,...
  • Le Chatelier's Principle: The universal systems law that when a system at dynamic equilibrium is subjected to an external disturbance, stress, or...
  • Local vs. Global Optima: The dilemma in optimization where a system reaches the peak of a low hill (a local optimum) where any immediate step in any...
  • Theory of Constraints: The principle that any manageable system is limited in achieving more of its goals by a very small number of constraints...
  • Tragedy of the Commons: The systemic dilemma where individuals acting independently and rationally according to their own self-interest deplete,...

πŸ“Š Economics, Strategy & Games (15 Models)

  • Adverse Selection & The Market for Lemons: The market failure where asymmetric information between buyers and sellers causes high-quality goods, healthy participants,...
  • Asymmetric Payoffs: Situations where the potential upside of an action or investment dwarfs the strictly capped, finite downside, creating an...
  • Comparative Advantage & Specialization: The economic principle that individuals, teams, or nations maximize collective and personal prosperity by specializing in...
  • Creative Destruction: The process of industrial and technological mutation that incessantly revolutionizes the economic structure from within,...
  • Economic Moats & Competitive Defensibility: The structural, sustainable competitive advantages that protect a business or platform from competitor encroachment,...
  • Economies of Scale vs. Diseconomies of Scale: The cost advantage that arises with increased output of a product, where fixed costs are spread across more units (economies...
  • Gresham's Law: The monetary principle that "bad money drives out good money" when legal tender laws mandate an equal face value between...
  • Moral Hazard: The lack of incentive to guard against risk and loss when one is insulated from its negative consequences, typically because...
  • Nash Equilibrium & Prisoner's Dilemma: A state in a game or competitive system where no player has an incentive to unilaterally change their chosen strategy, even...
  • Network Effects: The economic dynamic where a product, platform, or service becomes exponentially more valuable to existing and prospective...
  • Opportunity Cost: The true cost of any choice is not merely the financial outlay, but the lost value of the next best alternative that must be...
  • Principal-Agent Problem: The structural conflict of interest that arises when one party (the agent) is authorized to make decisions on behalf of...
  • Rent-Seeking & Tollbooth Economics: The practice of expending resources, political lobbying, or legal maneuvering to manipulate the social or economic...
  • Sunk Cost Fallacy: The irrational cognitive bias of continuing an endeavor or allocating further resources based on the accumulated historical...
  • Zero-Sum vs. Non-Zero-Sum Games: The distinction between interactions where one party's gain is strictly balanced by another party's exact loss (zero-sum,...

πŸ‘€ Psychology & Human Behavior (13 Models)

  • Anchoring & Adjustment: The cognitive bias where an individual relies excessively on the first piece of information encountered (the "anchor") when...
  • Availability Heuristic: The mental shortcut where people evaluate the frequency, probability, or importance of an event based on how easily...
  • Commitment & Consistency: The deeply ingrained psychological desire to remain consistent with what we have previously said, done, or publicly...
  • Confirmation Bias: The pervasive psychological tendency to search for, interpret, favor, and recall information in a way that confirms one's...
  • The Curse of Knowledge: The cognitive bias where an individual, once they have acquired deep knowledge or mastery in a subject, finds it virtually...
  • Dunning-Kruger Effect: The cognitive bias where individuals with low competence or knowledge in a specific domain drastically overestimate their...
  • Framing Effect: The cognitive bias where people draw radically different conclusions and make divergent choices depending on how identical...
  • Fundamental Attribution Error: The cognitive tendency to explain others' negative behaviors by citing internal character flaws or disposition, while...
  • Hindsight Bias ("I Knew It All Along"): The psychological tendency for people to perceive past events as having been completely predictable and obvious before they...
  • Hyperbolic Discounting & Present Bias: The cognitive tendency to place disproportionately high value on immediate rewards right now, while steeply and...
  • Loss Aversion: The psychological phenomenon where the pain of losing something is cognitively experienced as roughly twice as intense as...
  • The Scarcity Principle: The psychological heuristic where people assign significantly higher subjective value, urgency, and desirability to...
  • Social Proof & Informational Cascades: The psychological phenomenon where people assume the actions, beliefs, and behaviors of others in an attempt to reflect...

🎲 Probability & Mathematics (12 Models)

  • Base Rate Fallacy: The cognitive tendency to ignore general prior statistical distribution rates in favor of specific, vivid, but...
  • Black Swan Theory: The theory of rare, unforeseen events that lie outside the realm of regular expectations, carry an extreme, catastrophic or...
  • Combinatorial Explosion: The mathematical phenomenon where the number of possible states, combinations, permutations, or interactions in a system...
  • Compounding & Exponential Growth: The mathematical process where the returns, gains, or outputs of an asset, knowledge base, or habit are reinvested to...
  • Ergodicity & Absorbing Barriers: The critical distinction between the average outcome of a group across parallel worlds (ensemble average) and the average...
  • Expected Value: The anticipated value of an investment, decision, or action calculated by multiplying each possible outcome by its...
  • The Gambler's Fallacy & The Monte Carlo Fallacy: The erroneous belief that if a particular random event has occurred more frequently than normal in the past, it is somehow...
  • The Kelly Criterion & Position Sizing: A mathematical formula for sizing a series of bets or capital investments to maximize the long-term compound geometric...
  • Law of Large Numbers: The mathematical theorem stating that as the number of independent, identically distributed trials increases, the empirical...
  • Power Laws & The Pareto Principle (80/20): The mathematical relationship where a tiny percentage of participants, causes, or inputs account for a vastly...
  • Regression to the Mean: The statistical phenomenon that if an extreme, outlier measurement or performance is observed on a first trial, subsequent...
  • Survivorship Bias: The logical error of focusing exclusively on the successful "survivors" of a selective process while completely ignoring the...

βš™οΈ Physics & Engineering (12 Models)

🧬 Evolution & Biological Systems (8 Models)

  • Carrying Capacity: The maximum population size or throughput that an environment or system can sustainably sustain without exhausting its...
  • Co-evolution & Symbiosis: The process where two or more distinct species, technologies, or organizations reciprocally affect each other's evolution...
  • Ecological Niches & The Competitive Exclusion Principle: The principle that two species competing for the exact same ecological niche and identical finite resources cannot coexist...
  • Evolutionary Mismatch: The phenomenon where biological or behavioral traits that evolved to maximize fitness in an ancestral environment become...
  • Natural Selection & Survival of the Fittest: The universal process where differential reproductive success driven by environmental selection pressure acts upon heritable...
  • Punctuated Equilibrium: The evolutionary pattern where complex systems experience long periods of static stability (stasis) interrupted by brief,...
  • The Red Queen Effect: The principle that in a co-evolving, competitive system, continuous adaptation, innovation, and exertion are required simply...
  • Costly Signaling & The Handicap Principle: The principle that for a signal or communication to be credible and trusted between actors with conflicting incentives, it...

⚑ Operations & High-Stakes Strategy (7 Models)

  • Attrition vs. Maneuver: The fundamental strategic dichotomy between grinding down an opponent through brute mass and symmetric resource depletion...
  • Center of Gravity (Schwerpunkt): The primary hub of power, cohesion, and movement upon which an entire system depends; concentrating overwhelming force...
  • Defense in Depth: The strategy of employing multiple layered, independent, and redundant defensive controls so that the failure or compromise...
  • The Fog of War & Epistemic Uncertainty: The profound uncertainty, ambiguity, misinformation, and cognitive confusion that invariably characterizes real-time...
  • Interior vs. Exterior Lines: The geometric principle where a centrally positioned force (interior lines) moves resources, communicates, and concentrates...
  • Mission Command (Auftragstaktik): A decentralized operational doctrine where leadership defines the overarching intent and constraints ("what" and "why"),...
  • The OODA Loop: A four-step decision cycle (Observe, Orient, Decide, Act) emphasizing that the competitor or system capable of processing...

πŸ› οΈ Repository Architecture

mental-models/
β”œβ”€β”€ .github/workflows/ci.yml       # Automated CI schema & CLI test suite
β”œβ”€β”€ cli/
β”‚   β”œβ”€β”€ mm                         # Executable CLI entrypoint (ANSI formatted)
β”‚   └── mm_core.py                 # Core parsing, search, and diagnosis engine
β”œβ”€β”€ docs/
β”‚   └── graph.html                 # Standalone interactive D3.js knowledge graph
β”œβ”€β”€ exports/
β”‚   └── mental_models_anki.tsv     # Generated Anki flashcard deck (200 cards)
β”œβ”€β”€ mcp/
β”‚   └── server.py                  # JSON-RPC Model Context Protocol server
β”œβ”€β”€ models/                        # THE SINGLE SOURCE OF TRUTH (100 models)
β”‚   β”œβ”€β”€ core_thinking/             # 18 models
β”‚   β”œβ”€β”€ economics_strategy/        # 15 models
β”‚   β”œβ”€β”€ evolution_biology/         # 8 models
β”‚   β”œβ”€β”€ operations_strategy/       # 7 models
β”‚   β”œβ”€β”€ physics_engineering/       # 12 models
β”‚   β”œβ”€β”€ probability_math/          # 12 models
β”‚   β”œβ”€β”€ psychology_cognition/      # 13 models
β”‚   └── systems_complexity/        # 15 models
β”œβ”€β”€ scripts/
β”‚   β”œβ”€β”€ export_anki.py             # Generates Spaced Repetition decks
β”‚   β”œβ”€β”€ generate_graph.py          # Builds force-directed D3 graph
β”‚   └── validate_models.py         # Validates frontmatter & link schema
β”œβ”€β”€ skills/
β”‚   └── mental-models/SKILL.md     # Antigravity / AI Agent skill specification
β”œβ”€β”€ templates/
β”‚   └── model-template.md          # Standard schema for adding new models
β”œβ”€β”€ CONTRIBUTING.md
β”œβ”€β”€ LICENSE                        # MIT License
└── README.md

πŸ§ͺ Schema Validation & CI

All model contributions must conform to the strict schema checked by our validation script:

python3 scripts/validate_models.py

Checks performed:

  • Valid YAML frontmatter (id, title, domain, category, summary, triggers, counter_models, paired_models).
  • Presence of all required sections: Core Intuition, Diagnostic Checklist, Real-World Case Studies, Failure Modes, The Latticework, AI Protocol.
  • Integrity of inter-model references.

🀝 Contributing

Contributions of new mental models or refinements to existing ones are warmly welcome! Please review CONTRIBUTING.md and use templates/model-template.md as the blueprint.


πŸ“„ License

This project is licensed under the MIT License Β© 2026 Gabriel Rondon.

About

A comprehensive open-source latticework of mental models for human decision-making and AI agent reasoning.

Resources

Contributing

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages