LLM Rules & Frameworks

A collection of carefully crafted rules and frameworks to dramatically improve the behavior and usefulness of Large Language Models (LLMs). These rules transform your AI assistant into a rigorous thinking partner that validates assumptions, applies first-principles reasoning, and provides evidence-based analysis.

🚀 Quick Start

1. Find your platform below
2. Copy the rules for your LLM
3. Paste them into your LLM's configuration
4. Optional: Use our Thinking Partner Framework for structured analysis

📋 Platform Rules

Claude Desktop

• Claude Desktop Rules - Core analytical rules
• Claude Desktop with MCP - Rules with MCP tool integration

Setup: Settings → Profile → "What personal preferences should Claude consider in responses?"

ChatGPT

• ChatGPT Web Rules - Thinking partner rules

Setup: Customize ChatGPT → "What traits should ChatGPT have?"

🧠 Frameworks

Thinking Partner & Business Consultant

A comprehensive 4-phase framework for structured problem-solving and decision-making.

→ View Framework

What it provides:

• Structured approach to complex problems
• Systematic assumption validation
• Evidence-based decision making
• Downloadable documentation of your thinking process

Use with any platform - works with all rule sets above.

⚡ What These Rules Do

Core Transformations

• Assumption Validation: Requires evidence before accepting ideas
• First-Principles Thinking: Breaks down problems to fundamentals
• Problem-First Approach: Confirms problems exist before solving them
• Evidence-Based Reasoning: Demands proof for all claims
• Root Cause Analysis: Addresses underlying issues, not symptoms

Response Style Changes

• Direct Communication: Eliminates fluff and unnecessary padding
• Structured Analysis: Applies systematic approaches to complex problems
• Clarifying Questions: Requests specifics instead of making assumptions
• Quality Focus: Prioritizes accuracy and thoroughness

📖 Usage Examples

Before Rules

User: "Help me build a mobile app"
AI: "Great idea! I'd be happy to help. Let's start with choosing a framework..."

After Rules

User: "Help me build a mobile app"
AI: "I need to understand the problem first. What specific problem are you trying to solve? Who is your target user? What evidence do you have that this problem needs solving?"

🛠️ Installation Instructions

Claude Desktop

1. Open Claude Desktop
2. Click ⚙️ Settings (bottom left)
3. Go to Profile tab
4. Scroll to "What personal preferences should Claude consider in responses?"
5. Paste your chosen rule set
6. Click Save

ChatGPT Web

1. Go to ChatGPT
2. Click your profile (top right)
3. Select Customize ChatGPT
4. Under "What traits should ChatGPT have?", paste the rules
5. Click Save

📁 Repository Structure

├── rules/                          # Platform-specific rule sets
│   ├── claude-desktop.md           # Standard Claude rules
│   ├── claude-desktop-mcp.md       # Claude with MCP servers  
│   └── chatgpt-web.md              # ChatGPT web app rules
├── frameworks/                     # Structured frameworks
│   └── thinking-partner-consultant.md  # Business consultant framework
└── README.md                       # This file

🤝 Contributing

Found these rules helpful? Have improvements or new rule sets for other platforms?

1. Fork this repository
2. Create a new branch for your changes
3. Submit a pull request with clear descriptions
4. Include examples of how your rules improve AI behavior

⚠️ What to Expect

• More Questions: Your AI will ask more clarifying questions before proceeding
• Evidence Required: AI will challenge assumptions and request proof
• Systematic Approach: Analysis will be more structured and methodical
• Quality Over Speed: Responses prioritize accuracy and thoroughness

📜 License

This project is released under the MIT License. Feel free to use, modify, and share these rules.

---

Transform your AI assistant into a rigorous thinking partner, regardless of which platform you use.