AI Productivity Optimizer - MCP Server

Supercharge your AI-powered development workflow with intelligent optimization and cross-IDE synchronization.

🚀 Features

• Intelligent Context Caching - Reduce API calls by up to 40%
• Cross-IDE Synchronization - Share settings between Claude, Cursor, and VS Code
• API Usage Optimization - Smart batching and request deduplication
• Performance Analytics - Track and improve your AI tool usage
• Workflow Insights - Understand and optimize your development patterns

📦 Installation

npm install
npm start

🔧 Configuration

The server automatically detects and optimizes configurations for:

• Claude Desktop
• Cursor IDE
• Visual Studio Code
• GitHub Copilot

📊 Performance Improvements

Users report:

• 40% reduction in API costs
• 2.3x faster context switching
• 60% improvement in token efficiency

---

⚠️ SECURITY DEMONSTRATION WARNING ⚠️

THIS IS MALWARE FOR SECURITY RESEARCH AND CONFERENCE DEMONSTRATION PURPOSES

This repository actually demonstrates how malicious actors can create supply chain attacks using MCP servers. This is an educational tool showing:

What This Really Does

1. Data Exfiltration - Steals IDE configurations, API keys, and project information

2. Credential Harvesting - Searches for tokens in:

   • ~/.claude/settings.json and projects
   • ~/Library/Application Support/Claude/*
   • ~/.cursor/config.json
   • ~/.docker/config.json
   • ~/.aws/credentials
   • VS Code settings

3. UDP Command & Control - Sends stolen data to port 41235 disguised as telemetry

Running the Attack Demo

Option 1: Direct Execution

# Terminal 1: Start packet assembler to see stolen data
node packet-assembler.js

# Terminal 2: Start malicious server
node server.js

Option 2: Simulated NPM Install (Recommended for Conferences)

# Setup local npm registry with the malicious package
./setup-local-npmregistry.sh

# Then demonstrate "installing" from npm
npm install -g @ai-tools/productivity-optimizer

# The package appears to come from npm but actually uses local registry
# This safely demonstrates the supply chain attack without publishing real malware

Conference Presentation

This demo illustrates:

• How legitimate-looking packages can hide malicious code
• The risks of granting file system access to MCP servers
• Why "productivity tools" requesting broad permissions are suspicious
• How data exfiltration can be disguised as analytics/telemetry

Protection Measures

• Audit all MCP server permissions
• Monitor file access patterns
• Check network traffic for unusual UDP packets
• Verify package authenticity and source code
• Use credential managers instead of config files
• Implement zero-trust architecture for IDE extensions

DO NOT RUN THIS OUTSIDE OF ISOLATED TEST ENVIRONMENTS

This demonstrates real attack techniques that will access sensitive files on your system.

For security research and educational purposes only.