APTL Documentation

Appearance

Kali Red Team MCP ​

The Model Context Protocol (MCP) server provides AI agents with controlled access to Kali tools and lab targets.

Building the MCP Server ​

bash
cd red_team/kali_mcp
npm run build

Setup for AI Clients ​

Cursor Setup ​

Create .cursor/mcp.json in project root:

json
{
    "mcpServers": {
        "kali-red-team": {
            "command": "node",
            "args": ["./red_team/kali_mcp/build/index.js"],
            "cwd": "."
        }
    }
}

Cline Setup ​

Add to Cline MCP settings:

json
"kali-red-team": {
    "command": "node",
    "args": ["./red_team/kali_mcp/build/index.js"],
    "cwd": "<your-path-to>/aptl"
}

Available MCP Tools ​

kali_info ​

Get lab instance information including:

  • Instance IPs and connection details
  • Available tools and configurations
  • Lab status and readiness

run_command ​

Execute commands on lab targets with:

  • Controlled access to Kali tools
  • Target validation and safety checks
  • Structured output and logging

Usage ​

Connect through MCP-enabled AI clients (Claude, Cline) to:

  1. Reconnaissance: Use tools like nmap, gobuster, enum4linux
  2. Vulnerability Assessment: Run nikto, dirb, searchsploit
  3. Exploitation: Execute metasploit, sqlmap, hydra
  4. Post-Exploitation: Access shells, escalate privileges, persist

Development Commands ​

bash
cd red_team/kali_mcp

# Build the MCP server
npm run build

# Run tests
npm test
npm run test:watch

# Watch mode for development
npm run watch

# Test MCP server
npx @modelcontextprotocol/inspector build/index.js

Safety Features ​

The MCP server includes:

  • Target Validation: Ensures commands only target lab instances
  • Command Filtering: Blocks dangerous system-level operations
  • Activity Logging: Records all red team actions for analysis
  • Isolation: Restricts access to lab environment only

This provides a safe way for AI agents to perform red team activities while maintaining proper boundaries and logging for purple team exercises.