MCP Standards
# MCP Standards - Personal Memory for Claude
> **โ ๏ธ ARCHIVED PROJECT**: This repository is archived and provided for reference only. The project was experimental and served as a proof-of-concept for automatic preference learning with AgentDB vector memory. See [ARCHIVE.md](ARCHIVE.md) for complete details.
**Make Claude remember YOUR preferences automatically. Zero config, zero manual steps.**
[](LICENSE)
[](https://www.python.org/downloads/)
[](ARCHIVE.md)
---
## ๐ฏ What This Does
Stop repeating yourself to Claude. This MCP server learns your preferences automatically:
```
You: "Install pytest"
Claude: pip install pytest
You: "Actually, use uv not pip"
Claude: โ Remembered
Next session:
You: "Install requests"
Claude: uv pip install requests [automatic]
```
**One correction. Forever remembered.**
---
## โก Quick Start (5 Minutes)
### 1. Install
```bash
git clone https://github.com/airmcp-com/mcp-standards.git
cd mcp-standards
# Install dependencies
npm install
# Setup AgentDB
npm run setup
```
### 2. Configure Claude Desktop
Add to `~/Library/Application Support/Claude/claude_desktop_config.json`:
```json
{
"mcpServers": {
"mcp-standards": {
"command": "uv",
"args": [
"run",
"--directory",
"/ABSOLUTE/PATH/TO/mcp-standards",
"python",
"-m",
"mcp_standards.server_simple"
]
}
}
}
```
**โ ๏ธ Replace `/ABSOLUTE/PATH/TO/` with your actual path!**
### 3. Restart Claude Desktop
Quit and relaunch Claude Desktop.
### 4. Test It!
```
You: "Remember: use uv not pip"
Claude: โ Remembered: 'use uv not pip' (python)
You: "What do you remember?"
Claude: I remember you prefer:
- Use uv not pip for Python projects
```
**That's it! You're done.** ๐
---
## ๐ง How It Works
### Automatic Learning
Just correct Claude naturally - it learns automatically:
```
Session 1:
You: "Use uv not pip"
โ Auto-detected and stored in AgentDB
Session 2+:
You: "Install anything"
โ Claude uses uv automatically
```
### What Gets Remembered
โ
**Tool preferences**: "use uv not pip", "prefer yarn over npm"
โ
**Workflow patterns**: "run tests before commit"
โ
**Code style**: "use TypeScript for new files"
โ
**Project conventions**: "follow PEP 8"
### Categories (Auto-Detected)
- `python` - Python/pip/uv preferences
- `javascript` - npm/yarn/pnpm preferences
- `git` - Git workflow preferences
- `docker` - Docker/container preferences
- `testing` - Test framework preferences
- `general` - Everything else
---
## ๐ Features
| Feature | Status |
|---------|--------|
| **Auto-detection** | โ
Detects "use X not Y" automatically |
| **Semantic search** | โ
<1ms with AgentDB (150x faster than SQLite) |
| **Cross-session** | โ
Preferences persist forever |
| **Zero config** | โ
Works out of the box |
| **100% local** | โ
No cloud, all private |
| **Simple** | โ
5-minute setup |
---
## ๐ Documentation
- **[Quick Start Guide](docs/QUICKSTART_SIMPLE.md)** - Detailed setup instructions
- **[Validation Checklist](docs/VALIDATION_CHECKLIST.md)** - Testing & troubleshooting
- **[Implementation Plan](docs/SIMPLE_V2_PLAN.md)** - Technical details
- **[Skills Guide](.claude/skills/remember-preferences.md)** - How to use in Claude
---
## ๐ ๏ธ MCP Tools Available
### Personal Memory (Simple Version)
```javascript
// Store preference
remember({
content: "use uv not pip",
category: "python"
})
// Search preferences
recall({
query: "package manager",
top_k: 5
})
// List all categories
list_categories()
// Get statistics
memory_stats()
```
### Config Standards (Bonus)
```javascript
// Generate minimal CLAUDE.md from project config files
generate_ai_standards({
project_path: ".",
formats: ["claude"]
})
```
---
## ๐ Architecture
### Simple & Fast
```
User corrects Claude
โ
Auto-detection hook triggers
โ
Stores in AgentDB (semantic vector memory)
โ
Next session: Claude queries AgentDB automatically
โ
Uses remembered preference
```
### Technologies
- **AgentDB** - Ultra-fast vector memory (<1ms search)
- **Python** - MCP server (async)
- **SQLite** - Fallback storage
- **MCP Protocol** - Claude Desktop integration
- **100% Local** - No cloud dependencies
---
## ๐ Project Structure
```
mcp-standards/
โโโ src/mcp_standards/
โ โโโ agentdb_client.py # AgentDB wrapper
โ โโโ hooks/auto_memory.py # Auto-detection
โ โโโ server_simple.py # Simple MCP server
โโโ tests/
โ โโโ test_simple_setup.py # Validation tests
โโโ docs/
โ โโโ QUICKSTART_SIMPLE.md # Setup guide
โ โโโ VALIDATION_CHECKLIST.md # Testing guide
โ โโโ SIMPLE_V2_PLAN.md # Technical details
โโโ scripts/
โ โโโ setup-agentdb.js # Setup script
โโโ .claude/skills/
โ โโโ remember-preferences.md # Claude skill
โโโ README.md # This file
```
**Clean. Simple. Works.**
---
## ๐งช Testing
Run automated validation:
```bash
python3 tests/test_simple_setup.py
```
**Expected output**:
```
โ PASS: Directory Structure
โ PASS: Required Files
โ PASS: Module Imports
โ PASS: AgentDB Client Init
โ PASS: Auto Memory Patterns
Results: 5/5 tests passed
Status: Ready for dev testing ๐
```
---
## ๐ Troubleshooting
### Setup fails
```bash
# Check Node.js version
node --version # Need v18+
# Install AgentDB manually
npm install -g agentdb
npx agentdb --version
```
### Claude Desktop doesn't connect
```bash
# Check logs
tail -f ~/Library/Logs/Claude/mcp*.log
# Look for initialization messages
# Should see: "MCP Standards (Simple) initialized"
```
### Preferences not remembered
Check that:
1. Server is running (check Claude Desktop MCP status)
2. Corrections use clear phrases ("use X not Y")
3. AgentDB path exists: `~/.mcp-standards/agentdb`
**More help**: See [Validation Checklist](docs/VALIDATION_CHECKLIST.md)
---
## ๐ฏ Performance
| Metric | Value |
|--------|-------|
| **Setup time** | <5 minutes |
| **Server startup** | <2 seconds |
| **Search speed** | <1ms (AgentDB HNSW) |
| **Detection** | Real-time (async) |
| **Storage** | <10ms |
| **Memory usage** | ~50MB (embedding model) |
**150x faster than SQLite. Zero lag.**
---
## ๐ Privacy
- โ
**100% local** - Everything stored in `~/.mcp-standards/`
- โ
**No cloud** - No external API calls
- โ
**No telemetry** - No data collection
- โ
**Your data** - You control everything
---
## ๐ What Changed (v2 Simple)
We removed all the complexity:
| v1 (Old) | v2 Simple (New) |
|----------|-----------------|
| Manual MCP calls (4-5 steps) | โ
Automatic (zero steps) |
| SQLite keyword search (50ms+) | โ
AgentDB vector search (<1ms) |
| No semantic matching | โ
Semantic understanding |
| Complex setup | โ
5-minute setup |
| 6,000+ LOC | โ
~950 LOC |
**Result**: 80% less code, 100x better UX
---
## ๐ค Contributing
This is a personal side project. If you want to contribute:
1. Try it yourself first
2. Open an issue describing what you want to add
3. Wait for feedback before writing code
**Please don't**: Submit large PRs without discussion first.
---
## ๐ License
MIT License - See [LICENSE](LICENSE) file
---
## ๐ Credits
Built with inspiration from:
- **[AgentDB](https://agentdb.ruv.io)** - Ultra-fast vector memory
- **[Context Engineering Guide](https://github.com/coleam00/context-engineering-intro)** - Minimal CLAUDE.md principles
- **Gunnar's approach** - Simple side projects that solve personal problems
---
## ๐ฌ Questions?
**Q: Why not just use v1?**
A: v1 requires 4-5 manual MCP calls per correction. v2 is zero-touch.
**Q: Do I need AgentDB?**
A: Yes, but it's installed automatically via `npm run setup`.
**Q: Is my data private?**
A: 100% local. Everything stored in `~/.mcp-standards/`. No cloud.
**Q: What if I want the old version?**
A: Use `src/mcp_standards/server.py` instead of `server_simple.py`.
---
## ๐ Next Steps
```bash
# Try it now
git clone https://github.com/airmcp-com/mcp-standards.git
cd mcp-standards
npm run setup
# See: docs/QUICKSTART_SIMPLE.md
```
**Stop repeating yourself. Start remembering automatically.** ๐ฏ
---
## ๐ฆ Archive Status
**This project is archived.** See [ARCHIVE.md](docs/ARCHIVE.md) for:
- Complete project status and achievements
- Security audit results
- Full documentation index
- Lessons learned and technical insights
- How to use this repository as reference
---
**Made with โค๏ธ by keeping it simple**
TDQS
Scored across 9 tools
Most tools have distinct purposes, but some potential overlap exists between 'suggest_claudemd_update' and 'update_claudemd' where an agent might need to understand the difference between suggestion and application. The core memory operations (add_episode, list_recent, search_episodes) are clearly differentiated, and export/learning tools serve separate functions.
All tools follow a consistent verb_noun naming pattern with snake_case throughout. The naming convention is predictable and readable, with clear action-object relationships (e.g., add_episode, search_episodes, update_claudemd). No mixed conventions or style deviations are present.
With 9 tools, the count is well-scoped for a knowledge/standards management server. Each tool appears to serve a specific purpose in the workflow (memory management, export, learning, standards generation), and none seem redundant or unnecessary given the domain.
The tool set covers core workflows for knowledge management (add, list, search), learning from corrections, and standards generation/updating. A minor gap exists in direct memory modification beyond adding episodes (e.g., no update/delete episode tools), but agents can likely work around this through the learning system. The export and auto-generation tools provide good coverage for the stated purpose.