Simple Vendor Risk Assessment MCP Server

🎯 Simplest possible vendor risk assessment using AWS Titan and MCP

Following krishnaik06 MCP-CRASH-Course pattern with function-only implementation.

🚀 Quick Start

Google Colab (Easiest)

1. Open Vendor_Risk_Assessment_MCP.ipynb in Google Colab

2. Add your AWS credentials

3. Run all cells

Local Setup

# Extract and setup
unzip simple_vendor_risk_mcp.zip
cd vendor-risk-mcp
python setup.py

# Configure AWS
cp .env.example .env
# Edit .env with your AWS credentials

# Test
python test_client.py

# Run MCP server
python main.py

🔧 MCP Tools

⚙️ Claude Desktop Config

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "vendor-risk": {
      "command": "python",
      "args": ["/path/to/main.py"],
      "env": {
        "AWS_ACCESS_KEY_ID": "your_key",
        "AWS_SECRET_ACCESS_KEY": "your_secret"
      }
    }
  }
}

💡 Example Usage

Single Assessment:

"Assess the risk of using Salesforce"
→ assess_vendor_risk("Salesforce")

Comparison:

"Compare Microsoft vs Google vs Amazon"
→ compare_vendors("Microsoft, Google, Amazon")

Industry Benchmark:

"What are typical risks in Healthcare?"
→ get_industry_risk_benchmark("Healthcare")

📊 How It Works

1. Mock Data Generation: Creates realistic vendor profiles

2. AWS Titan Analysis: AI-powered risk insights

3. Risk Scoring: 1-10 scale (lower = better)

4. Comprehensive Reports: Executive summaries with recommendations

🛠️ Architecture

• Function-only implementation (no classes)

• FastMCP server for MCP protocol

• AWS Bedrock Titan for AI analysis

• Realistic mock data for demonstrations

• Google Colab compatible

🔒 Requirements

• Python 3.9+

• AWS account with Bedrock access

• AWS credentials configured

📱 Files Included

• main.py - Main MCP server

• test_client.py - Simple testing

• setup.py - Easy installation

• Vendor_Risk_Assessment_MCP.ipynb - Google Colab notebook

• Configuration and documentation files

Built following krishnaik06 MCP-CRASH-Course patterns! 🚀