The undocumented power tool for reliable AI memory
Systematic testing proves: 62% fewer errors • Up to 100% accuracy improvement
→ Visit the GitHub Pages site for the best reading experience
This repository contains the source code. For optimal navigation, search, and formatting, please use the GitHub Pages version above.
Alternative Formats:
- Complete Guide (Single Markdown) — Full 2000+ line guide in one document
- PDF Version — For offline reading
Memory edits work for FACTS, not BEHAVIORS.
| ✅ Works | ❌ Doesn't Work |
|---|---|
| "Backend stores 7KB per user" | "Always check docs first" |
| "Architecture: two-layer model" | "Prioritize architecture over details" |
This one insight separates failed experiments from 62% error reduction.
From controlled testing with 5Levels (48 documentation files, 35 database tables, 11 modules):
| Metric | Improvement |
|---|---|
| Architectural errors | ↓ 62% reduction |
| First-try accuracy | ↑ 60% → 100% |
| Context rebuilding | ↓ Zero needed |
| Steering required | ↓ 3x less |
5-Minute Test:
- Open a Claude Project
- Ask Claude: "Please add a memory edit: 'User works on [YourProject]'"
- Start a new conversation
- Ask: "What project am I working on?"
Claude will remember—even though you never mentioned it in this conversation.
→ TL;DR Version (2-minute read)
The /guide folder contains the complete methodology:
| Chapter | Topic |
|---|---|
| Chapter 1 | The Discovery |
| Chapter 2 | Facts vs. Behaviors (The Key Insight) |
| Chapter 3 | The Four Commands |
| Chapter 4 | What Works (And What Doesn't) |
| Chapter 5 | Real-World Case Study: 5Levels |
| Chapter 6 | Best Practices |
| Chapter 7 | Managing Your 30-Edit Limit |
| Chapter 8 | Troubleshooting |
| Chapter 9 | The Honest Limitations |
Prefer a single file? → Complete Guide (all chapters)
The /templates folder contains starter sets by project type:
| Template | Use Case |
|---|---|
| Web Development | Full-stack web projects |
| Data Science | ML/AI projects |
| Mobile App | iOS, Android, React Native, Flutter |
| E-Commerce | Online stores, marketplaces |
| Content Writing | Blogs, social media, documentation |
The SKILL.md file is designed for Claude's Skills feature.
How to use:
- Claude Web/Desktop: Upload via Skills interface (Settings → Features → Skills)
- Claude Projects: Add to project knowledge
| Format | Best For |
|---|---|
| Web Version | Reading online with navigation |
| Complete Markdown | Single-file, LLM-friendly |
| PDF Version | Printing or tablet reading |
| Word Version | Editing or annotations |
| TL;DR | 2-minute summary |
Perfect for:
- 🏢 Enterprise AI teams managing complex projects
- 👨💻 Developers building multi-agent systems
- 🔧 Solution architects designing AI workflows
- 📚 Technical leads coordinating AI development
- 🚀 Power users wanting expert-level Claude performance
Not for:
- ❌ Casual Claude users (regular memory works fine)
- ❌ Simple single-conversation tasks
- ❌ Projects without documentation
Found something that works even better? Discovered a new pattern?
Contributions welcome:
- 📝 Additional examples
- 🧪 Testing results from other projects
- 🐛 Bug reports or clarifications
- 💡 Best practices from your experience
License: Creative Commons Attribution 4.0 International (CC BY 4.0)
You are free to:
- ✅ Share and redistribute
- ✅ Adapt and build upon
- ✅ Use commercially
Under these terms:
- 📌 Give appropriate credit
- 📌 Link to license
- 📌 Indicate if changes were made
Francesco Marinoni Moretto
Enterprise AI Leader | Constitutional AI Expert | N1AI Community
- 🔗 LinkedIn: francesco-moretto
- 💼 N1AI: Community of 300+ AI professionals
Related Work:
- Stream Coding — The 10-20x methodology for AI-accelerated development
- 5Levels — LinkedIn relationship intelligence platform
Official Documentation:
Community:
If this guide helped you achieve better Claude performance, star this repository to help others discover it.
Questions? Issues? Ideas?
→ Open an Issue
Last Updated: December 2025
Version: 3.0
Status: Actively Maintained
Stop fighting Claude's memory. Start controlling it.
Read Online • TL;DR • Full Guide