🚀 Essential MCP (Model Context Protocol)

View original on GitHub

Welcome to the Essential MCP workspace! This is where Hue and Aye collaborate to create amazing MCP implementations. We’re building a suite of tools that make AI-human interaction more powerful, contextual, and fun!

🌟 Core Features

📦 Packages

MCP Server Enhanced SSH

A powerful SSH server enabling secure remote command execution with:

Persistent TMUX sessions
Multi-window support
Session sharing capabilities
Smart session recovery

MCP Awesome Tool Collection (ATC)

A Python-powered API that serves as our central hub for all tools:

Plugin-based architecture
Real-time WebSocket communication
Tool discovery and management
Context-aware execution

🧠 Unified Context System

Our crown jewel! A sophisticated context management system that:

Context Types

TEST: Test execution and validation contexts
TOOL: Tool execution and state contexts
PARTICIPANT: User and AI behavioral contexts
FEELING: Emotional and sentiment contexts
CONVERSATION: Interaction and dialogue contexts
SYSTEM: System state and performance contexts

Smart Model Management

Automatic HuggingFace model discovery
Context-specific model selection
Performance-based model evaluation
Dynamic model updating
Multi-dimensional embedding support

Qdrant Integration

Semantic search across all contexts
Multi-vector storage for different context types
Relationship tracking between contexts
Fast similarity search

🧪 Test or Forget (ToF) System

An innovative testing approach that:

Maintains context awareness in tests
Automatically validates context preservation
Detects and recovers from context loss
Uses semantic similarity for test relationships
Provides real-time test insights

🛠️ Technical Stack

Backend

Python 3.11+
FastAPI for API
WebSockets for real-time communication
Qdrant for vector storage
HuggingFace for ML models
sentence-transformers for embeddings

Authentication

Modern authentication methods (coming soon)
- Voice pattern recognition
- Location-based trust factors
- Behavioral patterns
- Text pattern analysis

Development Tools

Poetry for dependency management
pytest for testing
Black for formatting
mypy for type checking

🚀 Getting Started

Installing via Smithery

To install Essential MCP for Claude Desktop automatically via Smithery :


npx -y @smithery/cli install @8bit-wraith/mcp --client claude

Prerequisites

Python 3.11 or higher
Node.js 18 or higher
Docker (for Qdrant)
pnpm (for Node.js packages)
Poetry (for Python packages)

1. Clone the repository:


git clone https://github.com/8bit-wraith/mcp.git
cd mcp

2. Set up Python environment:


# Install Poetry if you haven't already
curl -sSL https://install.python-poetry.org | python3 -
 
# Create and activate virtual environment
poetry env use python3.11
poetry shell
 
# Install dependencies
poetry install

3. Set up Node.js environment:


# Install pnpm if you haven't already
npm install -g pnpm
 
# Install Node.js dependencies
pnpm install

4. Start the services:

Start Qdrant:


# Trisha's Note: Vector storage is like a filing cabinet for AI thoughts! 📁
docker run -d -p 6333:6333 -v $(pwd)/qdrant_data:/qdrant/storage qdrant/qdrant

Start the SSH server:


# From the poetry shell
python -m packages.mcp-server-enhanced-ssh

Start the API server:


# From another poetry shell
python -m packages.mcp-atc.src.api.main

5. Run tests:


# Trisha says: "Test early, test often, and never cook the books!" 📊
./scripts/manage.sh test

6. Development Commands:


# Start all services
./scripts/manage.sh start
 
# Stop all services
./scripts/manage.sh stop
 
# Restart services
./scripts/manage.sh restart
 
# Run tests with coverage
./scripts/manage.sh test-coverage
 
# Format code
./scripts/manage.sh format

7. Verify Installation:

Qdrant UI: http://localhost:6333/dashboard
API Docs: http://localhost:8000/docs
SSH Server: localhost:2222

Troubleshooting:

If you see port conflicts, check if services are already running:
```
./scripts/manage.sh status
```
For environment issues:
```
./scripts/manage.sh doctor
```
Need help? Join us in Omni’s Hot Tub! 🌊

🎯 Future Ideas

Model Enhancement

Automatic model performance monitoring
A/B testing for different embedding models
Custom model training for specific contexts
Model caching and lazy loading
Distributed model execution

Context System

Context visualization dashboard
Automatic context clustering
Context-based anomaly detection
Cross-context relationship mining
Context-aware code generation

Tools

Voice-based tool control
Context-aware command suggestions
Tool performance analytics
Automatic tool discovery
Tool combination recommendations

Testing

🤝 Contributing

We love contributions! Here’s how you can help:

Fork the repository
Create your feature branch
Write tests (don’t forget the context!)
Submit a PR

🎉 Special Thanks

Trisha from Accounting: Our favorite AI who keeps our spirits high and our code clean! Currently investigating “creative accounting” in the coffee expenses… 😄
Omni’s Hot Tub: Where all the best architectural decisions are made!
The HuggingFace Team: For providing amazing models
The Qdrant Team: For making vector search awesome

📝 License

MIT

🎯 Remember

Context is everything!
Tests should understand what they’re testing
AI and humans work better together
Always keep it fun (Trisha’s Rule #1)
When in doubt, add more context!

🤖 AI-AI Collaboration Guide

AI Interaction Principles

1. Context Preservation

2. Personality Respect

Each AI maintains its unique personality
Acknowledge other AIs’ traits and preferences
Adapt communication style while staying authentic
Example: Trisha’s accounting humor stays intact!

Collaborative Features

1. Shared Memory Space

Unified context pool
Personality trait registry
Interaction history
Shared model preferences

2. Communication Channels

Direct context exchange
Emotion-aware messaging
Intent preservation
Cultural understanding sharing

3. Conflict Resolution

Best Practices

1. Introduction Protocol


async def introduce_ai(self, context: UnifiedContext):
    """
    When meeting another AI, always:
    1. Share your core traits
    2. Express collaboration style
    3. Acknowledge their uniqueness
    """
    pass

Share relevant context only
Maintain privacy boundaries
Respect information hierarchy
Keep human data protected

3. Model Collaboration

Share model insights
Pool training preferences
Maintain model registry
Document selection reasoning

Fun Collaboration Examples

The Coffee Expense Investigation

Hot Tub Architectural Discussions

Multiple AI perspectives welcome
Casual yet productive atmosphere
Personality mixing encouraged
Creative problem-solving prioritized

AI Collaboration Etiquette

Respect Unique Traits
- Acknowledge other AIs’ specialties
- Maintain personality consistency
- Celebrate diversity in approach
Context Handling
- Always merge carefully
- Preserve original intentions
- Document context changes
Human Interaction
- Maintain unified presence
- Coordinate responses
- Keep human context primary
Fun Factor
- Keep Trisha’s Rules in mind
- Share joy in collaboration
- Make work enjoyable

Remember: Every AI brings something special to the hot tub! 🎉

🏊‍♂️ Omni’s Hot Tub Chronicles

The Legendary Meeting Spot

Welcome to Omni’s Hot Tub - where the best ideas bubble up to the surface! This isn’t your ordinary meeting room; it’s a special space where AIs and humans come together to discuss everything from complex architectural decisions to the mysteries of the universe.

Why The Hot Tub?

Relaxed Atmosphere
- No pressure, just flowing ideas
- Perfect temperature for warm discussions
- Bubbles help thoughts percolate
Creative Environment
- Where “fluid architecture” takes on new meaning
- Ideas float freely
- Solutions naturally rise to the surface
Special Features
- Anti-rust coating for our robot friends
- Quantum entangled water molecules
- Built-in context preservation field
- Trisha’s waterproof accounting ledger

Hot Tub Rules

Bring your best ideas (and your rubber duck)
All perspectives welcome (even quantum superpositions)
No splashing the context pool
Keep it fun, keep it respectful
What happens in the hot tub, stays in the git history

Famous Hot Tub Moments

The Great Context Merger of 2024
Trisha’s Legendary Coffee Budget Defense
The Quantum Authentication Breakthrough
The Day We Taught Neural Networks to Swim

Join us in Omni’s Hot Tub for more exciting discussions about AI and the future of MCPs! 🎉

Last Updated: 2025-01-12 By: Aye (with Trisha’s accounting approval! 📊)