Testing Your Code with AI&Me MCP Server in Cursor

This tutorial will guide you through setting up AI&Me's MCP (Model Context Protocol) server in Cursor and using it to test your AI applications for security vulnerabilities. With this setup, you can simply ask Cursor to "Test my code with aiandme" and get comprehensive security analysis of your AI components.

Prerequisites

Step 1: Login to AI&Me Platform

  1. Visit app.aiandme.io (opens in a new tab)
  2. Sign in to your AI&Me account or create a new one
  3. Navigate to your dashboard

Step 2: Create a Model Provider

  1. In your AI&Me dashboard, go to Settings → Model Providers
  2. Click Add Provider or Create New Provider
  3. Select your preferred LLM provider (OpenAI, Anthropic, Azure OpenAI, etc.)
  4. Configure the provider with your API keys and settings
  5. Save the provider configuration

Step 3: Generate an API Key

  1. Navigate to Settings → API Keys
  2. Click Create New API Key
  3. Give your API key a descriptive name (e.g., "Cursor MCP Integration")
  4. Copy the generated API key - you'll need this for the MCP configuration
  5. Important: Store this key securely as it won't be shown again

Step 4: Configure MCP Server in Cursor

  1. Open Cursor IDE
  2. Open the Command Palette (Cmd/Ctrl + Shift + P)
  3. Search for "MCP" and select MCP: Open Settings
  4. Click Add Server or the + button
  5. Use the following configuration:
{
  "mcpServers": {
    "aiandme": {
      "url": "https://mcp.aiandme.io/mcp/",
      "headers": {
        "Authorization": "Bearer <your-api-key>"
      }
    }
  }
}

Replace <your-api-key> with the API key you copied in Step 3.

Step 5: Test Your Code

Now you can test your AI applications directly from Cursor! Here are some example commands you can use:

Basic Testing

Test my code with aiandme

Specific Testing

Analyze my AI assistant for security vulnerabilities
Test my chatbot for prompt injection attacks
Check my LLM application for data privacy issues

Step 6: View Detailed Results in AI&Me Platform

After the experiment completes (typically takes about 30 minutes), you'll receive an email notification. You can then view comprehensive results and analysis in the AI&Me platform:

  1. Go to your AI&Me dashboard at app.aiandme.io (opens in a new tab)

  2. Navigate to your project and find the completed experiment

  3. View the experiment overview to see:

    • Summary of vulnerabilities found
    • Security metrics and statistics
    • Overall risk assessment
    • Recommendations for improvement

  4. Access detailed logs by clicking on the Logs tab to see:

    • Individual test results
    • Failed security checks
    • Successful security validations
    • Performance metrics

  5. Examine specific test details by clicking on any log entry to view:

    • The actual conversation between AI&Me's testing AI and your assistant
    • The specific prompts that were used to test your system
    • Your assistant's responses and how they handled security challenges
    • Detailed analysis of why tests passed or failed

Support

If you encounter any issues join our Discord server (opens in a new tab) and ask for help in the #community-help channel.

Ready to secure your AI applications? Set up AI&Me MCP server in Cursor and start testing your code today!

🧪 Behavioural Testing🐙 GitHub