A curated list of Python libraries, tools, and resources for structured reasoning, mental models, and decision-making.
Mental models are cognitive frameworks that help humans and AI agents interpret problems, make decisions, and reason about complex systems. This list focuses on executable tools — libraries, SDKs, and platforms that let you apply mental models programmatically, not just read about them.
- Python Libraries
- Claude Code Skills
- Curated Lists
- LLM Reasoning Research
- Neurosymbolic & Logic Reasoning
- Decision-Making Tools
- Content & Learning Resources
- Contributing
Libraries that provide programmatic access to mental models or reasoning operators.
- base120 - 120 reasoning operators across 6 transformation families (Perspective, Inversion, Composition, Decomposition, Recursion, Systems). Stdlib-only Python SDK, CLI, MCP server, append-only ledger. Zero dependencies.
- pyreason - Explainable inference software supporting annotated, real-valued, graph-based and temporal logic. Neurosymbolic reasoning framework.
- semantic-router - Superfast AI decision making and intelligent processing of multi-modal data. Route inputs to decisions using semantic similarity.
Mental model and reasoning skill packs for Claude Code and other Agent Skills-compatible tools.
- cc-thinking-skills - 28 eval-informed mental models and critical-thinking skills for Claude Code, GitHub Copilot, Codex, Cursor.
- mental-models-os - 80 operational mental models as Claude Code skills with model chaining.
- claude-skills-mental-models - Mental models and cognitive frameworks as Claude Code skills for better decision-making.
- premortem - Premortem skill for Claude Code — finds the concrete ways a plan could fail before commitment. Multi-agent silent scan, mitigation triplets.
Awesome-lists of mental models, cognitive biases, and thinking tools.
- awesome-concepts - Awesome list about all kinds of interesting topics: Laws, Principles, Mental Models, Cognitive Biases.
- awesome-mental-models - Curated list of awesome mental models, heuristics, intuition pumps, atomic habits, tools for thought.
Libraries for LLM reasoning research (training, inference, evaluation). These are in a different category from mental-models libraries but are often relevant.
- llm-reasoners - A library for advanced large language model reasoning. Includes reasoning algorithms, evaluation frameworks, and training tools.
- RAT-retrieval-augmented-thinking - Improves AI responses by leveraging DeepSeek's reasoning capabilities to guide other models through a structured thinking process.
- thinkchain - Advanced Claude streaming interface with interleaved thinking, dynamic tool discovery, and MCP integration.
Libraries that combine neural networks with symbolic logic for reasoning.
- pyreason - Explainable inference software supporting annotated, real-valued, graph-based and temporal logic. (Also listed in Python Libraries.)
- KAG - Logical form-guided reasoning and retrieval framework based on OpenSPG engine and LLMs.
Tools for structured decision-making, planning, and risk analysis.
- semantic-router - Superfast AI decision making and intelligent processing of multi-modal data. (Also listed in Python Libraries.)
- PageIndex - Document index for vectorless, reasoning-based RAG.
Non-executable resources for learning about mental models and reasoning.
- Farnam Street - Mental Models - Detailed explanations of ~100 mental models with examples and application guidance.
- Poor Charlie's Almanack - Charlie Munger's wisdom on building a "latticework of mental models."
- Thinking, Fast and Slow - Daniel Kahneman's framework for System 1 and System 2 thinking.
Contributions are welcome! Please read the contribution guidelines before submitting a pull request.
- Python libraries: Must be installable via
pipor source, with a public GitHub repo. Mental-models or reasoning-focused. - Claude Code skills: Must be a skill pack that includes mental models or reasoning frameworks.
- Curated lists: Must focus on mental models, cognitive biases, or thinking tools.
- LLM reasoning research: Must be a Python library for LLM reasoning (training, inference, or evaluation).
- Content resources: Must be freely accessible and high-quality.
- Fork this repository
- Add your project to the appropriate section, sorted by star count (descending)
- Include the project name, link, and a one-line description
- Include the star badge:
 - Submit a pull request
To the extent possible under law, the contributors of this list have waived all copyright and related or neighboring rights to this work.