Track MCP LogoTrack MCP
Track MCP LogoTrack MCP

The world's largest repository of Model Context Protocol servers. Discover, explore, and submit MCP tools.

Product

  • Categories
  • Top MCP
  • New & Updated
  • Submit MCP

Company

  • About

Legal

  • Privacy Policy
  • Terms of Service
  • Cookie Policy

© 2026 TrackMCP. All rights reserved.

Built with ❤️ by Krishna Goyal

    Bugbounty Mcp Server

    Comprehensive MCP server for bug bounty hunting and web application penetration testing

    4 stars
    Python
    Updated Oct 15, 2025

    Table of Contents

    • 🚀 Features
    • 🔍 Reconnaissance (13 Tools)
    • 🌐 Scanning (15 Tools)
    • 🛡️ Vulnerability Assessment (15 Tools)
    • 🌍 Web Application (10 Tools)
    • 🔧 Network Security (10 Tools)
    • 🕵️ OSINT (10 Tools)
    • ⚔️ Exploitation (10 Tools)
    • 📊 Reporting (10 Tools)
    • 📋 Total: 92+ Security Testing Tools
    • 🛠️ Installation
    • Prerequisites
    • 🐳 Docker Installation (Recommended)
    • Quick Docker Start
    • Manual Docker Commands
    • Docker Environment Variables
    • What's Included in Docker Image
    • 🔧 Native Installation
    • Quick Start
    • 🎯 Usage
    • Starting the MCP Server
    • � Docker Usage (Recommended)
    • 🚀 Native Usage with run.sh
    • 📋 Command Line Interface
    • 🤖 MCP Server Integration with LLMs
    • 🔗 Supported LLM Clients
    • 🔧 Troubleshooting MCP Integration
    • 🐳 Docker Advanced Usage
    • Development with Docker
    • Performance Tuning
    • Backup and Persistence
    • Docker Management Script
    • Example Configuration
    • 🗣️ Natural Language Examples
    • Reconnaissance
    • Vulnerability Testing
    • Comprehensive Testing
    • OSINT Gathering
    • 🔧 Configuration
    • Environment Variables
    • Tool Paths
    • Safety Features
    • 📁 Project Structure
    • 🔒 Security Considerations
    • Responsible Usage
    • Safety Features
    • Legal Disclaimer
    • 🤝 Contributing
    • Development Setup
    • 📄 License
    • 🙏 Acknowledgments
    • 📚 Documentation
    • 📞 Support
    • 🚀 Roadmap

    Table of Contents

    • 🚀 Features
    • 🔍 Reconnaissance (13 Tools)
    • 🌐 Scanning (15 Tools)
    • 🛡️ Vulnerability Assessment (15 Tools)
    • 🌍 Web Application (10 Tools)
    • 🔧 Network Security (10 Tools)
    • 🕵️ OSINT (10 Tools)
    • ⚔️ Exploitation (10 Tools)
    • 📊 Reporting (10 Tools)
    • 📋 Total: 92+ Security Testing Tools
    • 🛠️ Installation
    • Prerequisites
    • 🐳 Docker Installation (Recommended)
    • Quick Docker Start
    • Manual Docker Commands
    • Docker Environment Variables
    • What's Included in Docker Image
    • 🔧 Native Installation
    • Quick Start
    • 🎯 Usage
    • Starting the MCP Server
    • � Docker Usage (Recommended)
    • 🚀 Native Usage with run.sh
    • 📋 Command Line Interface
    • 🤖 MCP Server Integration with LLMs
    • 🔗 Supported LLM Clients
    • 🔧 Troubleshooting MCP Integration
    • 🐳 Docker Advanced Usage
    • Development with Docker
    • Performance Tuning
    • Backup and Persistence
    • Docker Management Script
    • Example Configuration
    • 🗣️ Natural Language Examples
    • Reconnaissance
    • Vulnerability Testing
    • Comprehensive Testing
    • OSINT Gathering
    • 🔧 Configuration
    • Environment Variables
    • Tool Paths
    • Safety Features
    • 📁 Project Structure
    • 🔒 Security Considerations
    • Responsible Usage
    • Safety Features
    • Legal Disclaimer
    • 🤝 Contributing
    • Development Setup
    • 📄 License
    • 🙏 Acknowledgments
    • 📚 Documentation
    • 📞 Support
    • 🚀 Roadmap

    Documentation

    BugBounty MCP Server

    A comprehensive Model Context Protocol (MCP) server for bug bounty hunting and web application penetration testing. This tool allows you to perform extensive security testing through natural language conversations with an LLM.

    🚀 Features

    🔍 Reconnaissance (13 Tools)

    • Subdomain Enumeration: Passive and active subdomain discovery
    • DNS Enumeration: Comprehensive DNS record analysis
    • WHOIS Lookup: Domain registration and ownership information
    • Certificate Transparency: SSL certificate log analysis
    • Google Dorking: Automated search engine reconnaissance
    • Shodan/Censys Integration: IoT and service discovery
    • GitHub Reconnaissance: Code repository analysis
    • Archive.org Search: Historical website analysis
    • Technology Detection: Web stack fingerprinting
    • Social Media Search: OSINT across platforms
    • Email Enumeration: Email address discovery
    • Reverse DNS: IP to hostname resolution
    • WAF Detection: Web Application Firewall identification

    🌐 Scanning (15 Tools)

    • Port Scanning: Comprehensive network port analysis
    • Service Enumeration: Detailed service fingerprinting
    • Web Directory Scanning: Hidden file/directory discovery
    • Web Crawling: Automated website exploration
    • Parameter Discovery: Hidden parameter identification
    • Subdomain Takeover: Vulnerability detection
    • SSL/TLS Analysis: Certificate security assessment
    • CMS Scanning: WordPress/Drupal/Joomla analysis
    • JavaScript Analysis: Client-side security review
    • HTTP Methods Testing: Verb tampering detection
    • CORS Analysis: Cross-origin policy testing
    • Security Headers: HTTP header security analysis
    • Nuclei Integration: Vulnerability template scanning
    • Fuzzing: Input validation testing
    • API Endpoint Discovery: REST/GraphQL/SOAP analysis

    🛡️ Vulnerability Assessment (15 Tools)

    • SQL Injection Testing: Automated SQLi detection
    • XSS Testing: Cross-site scripting analysis
    • Command Injection: OS command execution testing
    • File Inclusion (LFI/RFI): Path traversal analysis
    • XXE Testing: XML external entity detection
    • SSRF Testing: Server-side request forgery
    • IDOR Testing: Insecure direct object reference
    • CSRF Testing: Cross-site request forgery
    • Authentication Bypass: Login mechanism testing
    • Privilege Escalation: Permission boundary testing
    • JWT Security: JSON Web Token analysis
    • Session Management: Session security assessment
    • Race Condition: Concurrency vulnerability testing
    • Business Logic: Workflow security analysis
    • Deserialization: Unsafe object handling detection

    🌍 Web Application (10 Tools)

    • Access Control Testing: Authorization boundary testing
    • Security Misconfiguration: Configuration weakness detection
    • Sensitive Data Exposure: Information leakage analysis
    • API Security Testing: REST/GraphQL security assessment
    • File Upload Security: Upload mechanism testing
    • Input Validation: Data sanitization analysis
    • Cookie Security: Session cookie analysis
    • WebSocket Security: Real-time communication testing
    • GraphQL Security: Query language vulnerability testing
    • Error Handling Analysis: Information disclosure via errors

    🔧 Network Security (10 Tools)

    • Network Discovery: Live host identification
    • Firewall Detection: Security device identification
    • Load Balancer Detection: Traffic distribution analysis
    • CDN Detection: Content delivery network analysis
    • Proxy Detection: Intermediary service identification
    • Routing Analysis: Network path examination
    • Bandwidth Testing: Network performance analysis
    • Wireless Security: WiFi network assessment
    • Network Sniffing: Packet capture and analysis
    • Lateral Movement: Internal network exploration

    🕵️ OSINT (10 Tools)

    • Person Investigation: Individual background research
    • Company Investigation: Corporate intelligence gathering
    • Dark Web Monitoring: Hidden service surveillance
    • Data Breach Checking: Credential exposure analysis
    • Social Media Investigation: Profile analysis across platforms
    • Paste Site Monitoring: Leaked information detection
    • Code Repository Search: Source code intelligence
    • Geolocation Investigation: Physical presence analysis
    • Threat Intelligence: IoC analysis and attribution
    • Metadata Extraction: Document forensics

    ⚔️ Exploitation (10 Tools)

    • Exploit Search: Vulnerability database queries
    • Payload Generation: Custom exploit creation
    • Privilege Escalation: System access expansion
    • Lateral Movement: Network propagation techniques
    • Persistence Mechanisms: Backdoor installation methods
    • Data Exfiltration: Information extraction techniques
    • Credential Dumping: Password harvesting methods
    • Anti-Forensics: Evidence elimination techniques
    • Evasion Techniques: Security control bypass
    • Social Engineering: Human factor exploitation

    📊 Reporting (10 Tools)

    • Vulnerability Reports: Comprehensive security assessments
    • Executive Summaries: Business-focused reporting
    • Finding Tracking: Vulnerability lifecycle management
    • Metrics Dashboard: Security KPI visualization
    • Data Export: Multi-format result export
    • Remediation Planning: Prioritized fix roadmaps
    • Compliance Mapping: Framework alignment analysis
    • Risk Assessment: Business impact evaluation
    • Scan Comparison: Historical trend analysis
    • Proof of Concept: Exploit documentation

    📋 Total: 92+ Security Testing Tools

    🛠️ Installation

    Prerequisites

    • Python 3.10 or higher (Python 3.11+ recommended)
    • Git
    • Docker (for containerized deployment)
    • macOS, Linux, or Windows with WSL

    🐳 Docker Installation (Recommended)

    The easiest way to get started is using Docker, which includes all dependencies and security tools pre-installed.

    Quick Docker Start

    1. Clone the repository:

    bash
    git clone https://github.com/gokulapap/bugbounty-mcp-server.git
       cd bugbounty-mcp-server

    2. Configure API keys (optional but recommended):

    bash
    # Copy environment template
       cp env.example .env
       
       # Edit .env with your API keys
       nano .env

    3. Build and run with Docker Compose:

    bash
    # Build and start the container
       docker-compose up --build -d
       
       # View logs
       docker-compose logs -f bugbounty-mcp
       
       # Stop the container
       docker-compose down

    Manual Docker Commands

    bash
    # Build the Docker image
    docker build -t bugbounty-mcp:latest .
    
    # Run the container
    docker run -d \
      --name bugbounty-mcp-server \
      -v $(pwd)/output:/app/output \
      -v $(pwd)/data:/app/data \
      -v $(pwd)/.env:/app/.env:ro \
      bugbounty-mcp:latest
    
    # View logs
    docker logs -f bugbounty-mcp-server
    
    # Access container shell for debugging
    docker exec -it bugbounty-mcp-server /bin/bash
    
    # Stop and remove container
    docker stop bugbounty-mcp-server
    docker rm bugbounty-mcp-server

    Docker Environment Variables

    You can pass API keys and configuration directly to Docker:

    bash
    docker run -d \
      --name bugbounty-mcp-server \
      -e SHODAN_API_KEY="your_shodan_key" \
      -e VIRUSTOTAL_API_KEY="your_vt_key" \
      -e GITHUB_TOKEN="your_github_token" \
      -e LOG_LEVEL="INFO" \
      -v $(pwd)/output:/app/output \
      bugbounty-mcp:latest

    What's Included in Docker Image

    The Docker image includes:

    • ✅ Python 3.11 with all required packages
    • ✅ All 20+ security tools (nmap, nuclei, subfinder, httpx, etc.)
    • ✅ Essential wordlists for scanning
    • ✅ Optimized for security and performance
    • ✅ Non-root user for enhanced security
    • ✅ Health checks and monitoring

    🔧 Native Installation

    Quick Start

    1. Clone the repository:

    bash
    git clone https://github.com/gokulapap/bugbounty-mcp-server.git
       cd bugbounty-mcp-server

    2. Run the automated installation:

    bash
    # Make the run script executable
       chmod +x run.sh
       
       # Install everything automatically
       ./install.sh

    OR for manual installation:

    3. Create virtual environment:

    bash
    python3 -m venv venv
       source venv/bin/activate  # On Windows: venv\Scripts\activate

    4. Install dependencies:

    bash
    pip install -r requirements.txt
       pip install -e .

    5. Install external security tools (optional but recommended):

    bash
    # On Ubuntu/Debian
       sudo apt update
       sudo apt install nmap masscan nikto dirb sqlmap
       
       # On macOS with Homebrew
       brew install nmap masscan nikto dirb sqlmap
       
       # Install Go-based tools
       go install github.com/projectdiscovery/nuclei/v2/cmd/nuclei@latest
       go install github.com/projectdiscovery/subfinder/v2/cmd/subfinder@latest
       go install github.com/projectdiscovery/httpx/cmd/httpx@latest
       go install github.com/OJ/gobuster/v3@latest
       go install github.com/ffuf/ffuf@latest

    6. Configure API keys (optional):

    bash
    # Copy environment template
       cp env.example .env
       
       # Edit .env file with your API keys
       nano .env

    7. Download wordlists:

    bash
    # Download all wordlists (recommended)
       ./run.sh download-wordlists
       
       # Or download specific types
       ./run.sh download-wordlists --type subdomains
       ./run.sh download-wordlists --type directories
       ./run.sh download-wordlists --type parameters
       ./run.sh download-wordlists --type files
       
       # See available options
       ./run.sh download-wordlists --help

    8. Validate configuration:

    bash
    ./run.sh validate-config

    🎯 Usage

    Starting the MCP Server

    � Docker Usage (Recommended)

    Using Docker Compose (easiest):

    bash
    # Start the server
    docker-compose up -d
    
    # View logs in real-time
    docker-compose logs -f bugbounty-mcp
    
    # Stop the server
    docker-compose down

    Using Docker directly:

    bash
    # Start the server
    docker run -d \
      --name bugbounty-mcp \
      -v $(pwd)/output:/app/output \
      -v $(pwd)/.env:/app/.env:ro \
      bugbounty-mcp:latest
    
    # Check server status
    docker exec bugbounty-mcp bugbounty-mcp validate-config
    
    # View available tools
    docker exec bugbounty-mcp bugbounty-mcp list-tools

    🚀 Native Usage with run.sh

    The easiest way to start the server natively is using the provided run.sh script:

    bash
    # Navigate to the project directory
    cd bugbounty-mcp-server
    
    # Start the MCP server
    ./run.sh serve

    The script will:

    • ✅ Automatically activate the virtual environment
    • ✅ Load environment variables from .env file
    • ✅ Display server status and available tools
    • ✅ Start the MCP server for LLM integration

    📋 Command Line Interface

    bash
    # List all available commands
    ./run.sh --help
    
    # Start the MCP server
    ./run.sh serve
    
    # List all 92+ available tools
    ./run.sh list-tools
    
    # Validate configuration and tool availability
    ./run.sh validate-config
    
    # Perform a quick security scan
    ./run.sh quick-scan --target example.com
    
    # Download security wordlists
    ./run.sh download-wordlists --type subdomains
    
    # Export configuration template
    ./run.sh export-config --format yaml
    bugbounty-mcp export-config --format yaml -o config.yaml

    🤖 MCP Server Integration with LLMs

    The BugBounty MCP Server implements the Model Context Protocol (MCP), enabling seamless integration with various LLM applications for natural language penetration testing.

    🔗 Supported LLM Clients

    ##### 1. Claude Desktop (Recommended)

    Add to your Claude Desktop configuration file:

    macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

    Windows: %APPDATA%\Claude\claude_desktop_config.json

    For Docker (recommended):

    json
    {
      "mcpServers": {
        "bugbounty-mcp": {
          "command": "docker",
          "args": ["exec", "-i", "bugbounty-mcp-server", "bugbounty-mcp", "serve"],
          "env": {
            "DOCKER_HOST": "unix:///var/run/docker.sock"
          }
        }
      }
    }

    For Native Installation:

    json
    {
      "mcpServers": {
        "bugbounty-mcp": {
          "command": "/Users/your-username/Documents/bugbounty-mcp-server/run.sh",
          "args": ["serve"],
          "env": {
            "PATH": "/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin"
          }
        }
      }
    }

    ##### 2. VS Code with GitHub Copilot Integration

    To use the BugBounty MCP Server with VS Code and GitHub Copilot:

    Prerequisites:

    • VS Code with GitHub Copilot extension enabled
    • MCP extension for VS Code (if available in marketplace)

    Configuration Steps:

    1. For Docker Deployment (Recommended):

    First, ensure your Docker container is running with port 3001 exposed:

    bash
    # Start the container with automatic MCP server startup
       docker-compose up --build -d
       
       # Verify the server is accessible on port 3001
       nc -z localhost 3001 && echo "MCP server is ready"

    Then configure VS Code MCP settings by opening VS Code settings (Cmd/Ctrl + ,) and adding:

    json
    {
         "mcp.servers": {
           "bugbounty-docker": {
             "command": "nc",
             "args": ["localhost", "3001"],
             "description": "BugBounty MCP Server running in Docker",
             "env": {
               "LOG_LEVEL": "info"
             }
           }
         }
       }

    Alternative Docker configuration using direct Docker exec:

    json
    {
         "mcp.servers": {
           "bugbounty-docker": {
             "command": "docker",
             "args": ["exec", "-i", "bugbounty-mcp-server", "bugbounty-mcp", "serve"],
             "description": "BugBounty MCP Server via Docker exec",
             "env": {
               "DOCKER_HOST": "unix:///var/run/docker.sock"
             }
           }
         }
       }

    2. For Native Installation:

    json
    {
         "mcp.servers": {
           "bugbounty-native": {
             "command": "/Users/your-username/Documents/bugbounty-mcp-server/run.sh",
             "args": ["serve"],
             "description": "BugBounty MCP Server native installation",
             "env": {
               "PATH": "/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin"
             }
           }
         }
       }

    3. Verify Connection:

    • Restart VS Code or reload the MCP extension
    • Open the MCP panel in VS Code (if available)
    • You should see the BugBounty server connected
    • Test by asking GitHub Copilot: "List available security tools from BugBounty MCP"

    4. Troubleshooting Docker Integration:

    If using the Docker network approach and experiencing issues:

    bash
    # Check if container is running and healthy
       docker-compose ps
       
       # Test network connectivity
       nc -z localhost 3001 || echo "Port 3001 not accessible"
       
       # Check container logs
       docker-compose logs -f bugbounty-mcp
       
       # Verify MCP server response
       echo '{"jsonrpc": "2.0", "method": "initialize", "params": {"protocolVersion": "2024-11-05", "capabilities": {}, "clientInfo": {"name": "test", "version": "1.0"}}, "id": 1}' | nc localhost 3001

    ##### 3. Custom MCP Clients

    python
    import asyncio
    from mcp.client.session import ClientSession
    from mcp.client.stdio import stdio_client
    
    async def use_bugbounty_mcp():
        async with stdio_client(["./run.sh", "serve"]) as (read, write):
            async with ClientSession(read, write) as session:
                # Initialize the session
                await session.initialize()
                
                # List available tools
                tools = await session.list_tools()
                print(f"Available tools: {len(tools)}")
                
                # Call a tool
                result = await session.call_tool(
                    "subdomain_enumeration",
                    {"domain": "example.com", "passive_only": True}
                )
                print(result)
    
    # Run the client
    asyncio.run(use_bugbounty_mcp())

    ##### 4. Integration Examples

    Start the server and test:

    bash
    # Terminal 1: Start the MCP server
    ./run.sh serve
    
    # Terminal 2: Test with any MCP client
    # The server will be listening on stdio for MCP protocol messages

    Example LLM conversation:

    code
    User: "Please perform a comprehensive security assessment of example.com"
    
    LLM: I'll help you conduct a comprehensive security assessment using the BugBounty MCP tools. Let me start by gathering information about the target.
    
    [The LLM will automatically use tools like:]
    - subdomain_enumeration to find subdomains
    - port_scanning to identify open services  
    - vulnerability_scanning to detect security issues
    - web_directory_scanning to find hidden files
    - And 90+ other security tools as needed

    🔧 Troubleshooting MCP Integration

    If the server doesn't start in Claude Desktop:

    For Docker deployment:

    1. Ensure Docker container is running:

    bash
    docker ps | grep bugbounty-mcp
       # Should show running container

    2. Check container logs:

    bash
    docker logs bugbounty-mcp-server

    3. Test Docker integration:

    bash
    docker exec bugbounty-mcp-server bugbounty-mcp --help
       # Should show help output

    4. Verify Docker socket access (macOS/Linux):

    bash
    ls -la /var/run/docker.sock
       # Should be accessible

    For Native deployment:

    1. Check the path in your config:

    bash
    # Get the absolute path
       pwd
       # Use this full path in claude_desktop_config.json

    2. Verify the run.sh script is executable:

    bash
    chmod +x run.sh

    3. Test the server manually:

    bash
    ./run.sh serve
       # Should show "BugBounty MCP Server started successfully"

    4. Check Claude Desktop logs:

    • macOS: ~/Library/Logs/Claude/
    • Windows: %LOCALAPPDATA%\Claude\logs\

    🐳 Docker Advanced Usage

    Development with Docker

    bash
    # Build development image with debugging tools
    docker build -t bugbounty-mcp:dev --target builder .
    
    # Run with volume mounts for live development
    docker run -it --rm \
      -v $(pwd):/app \
      -v $(pwd)/output:/app/output \
      bugbounty-mcp:dev bash
    
    # Run specific tools
    docker exec bugbounty-mcp nmap --version
    docker exec bugbounty-mcp nuclei -version
    docker exec bugbounty-mcp subfinder -version

    Performance Tuning

    bash
    # Run with increased resources
    docker run -d \
      --name bugbounty-mcp \
      --cpus="2.0" \
      --memory="4g" \
      -v $(pwd)/output:/app/output \
      bugbounty-mcp:latest
    
    # Monitor resource usage
    docker stats bugbounty-mcp

    Backup and Persistence

    bash
    # Create data volume backup
    docker run --rm \
      -v bugbounty-data:/data \
      -v $(pwd)/backup:/backup \
      alpine tar czf /backup/data-backup-$(date +%Y%m%d).tar.gz -C /data .
    
    # Restore data volume
    docker run --rm \
      -v bugbounty-data:/data \
      -v $(pwd)/backup:/backup \
      alpine tar xzf /backup/data-backup-XXXXXXXX.tar.gz -C /data

    Docker Management Script

    For easier Docker management, use the included docker.sh script:

    bash
    # Make executable (first time only)
    chmod +x docker.sh
    
    # Build and run in one command
    ./docker.sh build && ./docker.sh run --api-keys
    
    # Quick operations
    ./docker.sh logs --follow     # View live logs
    ./docker.sh shell            # Access container shell
    ./docker.sh validate         # Validate setup
    ./docker.sh restart --force  # Force restart
    ./docker.sh clean --force    # Clean everything
    
    # Data management
    ./docker.sh backup           # Backup container data
    ./docker.sh restore backup/file.tar.gz  # Restore data
    
    # Development
    ./docker.sh build --dev      # Build dev image
    ./docker.sh run --dev        # Run with source mounting

    Example Configuration

    yaml
    # bugbounty_mcp_config.yaml
    api_keys:
      shodan: "your_shodan_api_key"
      virustotal: "your_virustotal_api_key"
      github: "your_github_token"
    
    tools:
      nmap_path: "nmap"
      nuclei_path: "nuclei"
      max_concurrent_scans: 10
      default_timeout: 30
    
    scanning:
      default_ports: ["21", "22", "23", "25", "53", "80", "443", "8080", "8443"]
      max_crawl_depth: 3
      max_pages_to_crawl: 100
    
    output:
      output_dir: "output"
      report_format: "json"
      create_html_report: true
    
    safety:
      safe_mode: true
      allowed_targets: ["*.example.com", "192.168.1.0/24"]
      blocked_targets: ["*.gov", "*.mil"]

    🗣️ Natural Language Examples

    Once integrated with an LLM, you can perform security testing through conversation:

    Reconnaissance

    code
    "Perform subdomain enumeration for example.com using both passive and active methods"
    
    "Check if example.com uses a CDN and try to find the origin server"
    
    "Search GitHub for any repositories mentioning example.com that might contain sensitive information"

    Vulnerability Testing

    code
    "Test the login form at https://example.com/login for SQL injection vulnerabilities"
    
    "Scan https://example.com for XSS vulnerabilities in all input parameters"
    
    "Check if https://example.com has any CORS misconfigurations"

    Comprehensive Testing

    code
    "Perform a complete security assessment of example.com including:
    - Subdomain discovery
    - Port scanning
    - Web application testing
    - SSL/TLS analysis
    - Generate a detailed report"

    OSINT Gathering

    code
    "Investigate the company Example Corp for:
    - Employee information
    - Technology stack
    - Recent data breaches
    - Social media presence"

    🔧 Configuration

    Environment Variables

    VariableDescriptionRequired
    SHODAN_API_KEYShodan API key for device discoveryNo
    VIRUSTOTAL_API_KEYVirusTotal API key for threat intelligenceNo
    CENSYS_API_IDCensys API ID for certificate/host searchNo
    CENSYS_API_SECRETCensys API secretNo
    GITHUB_TOKENGitHub token for repository searchNo
    SECURITYTRAILS_API_KEYSecurityTrails API for DNS historyNo
    HUNTER_IO_API_KEYHunter.io API for email discoveryNo
    BINARYEDGE_API_KEYBinaryEdge API for internet scanningNo

    Tool Paths

    The server automatically detects tools in your PATH, but you can specify custom paths:

    yaml
    tools:
      nmap_path: "/usr/local/bin/nmap"
      masscan_path: "/opt/masscan/bin/masscan"
      nuclei_path: "/home/user/go/bin/nuclei"
      # ... other tools

    Safety Features

    yaml
    safety:
      safe_mode: true                    # Enable safety checks
      allowed_targets:                   # Whitelist of allowed targets
        - "*.example.com"
        - "192.168.1.0/24"
        - "10.0.0.0/8"
      blocked_targets:                   # Blacklist of forbidden targets
        - "*.gov"
        - "*.mil"
        - "*.edu"
      rate_limit_enabled: true          # Enable rate limiting
      requests_per_second: 10.0         # Request rate limit

    📁 Project Structure

    code
    bugbounty-mcp-server/
    ├── bugbounty_mcp_server/
    │   ├── __init__.py
    │   ├── server.py              # Main MCP server
    │   ├── config.py              # Configuration management
    │   ├── utils.py               # Utility functions
    │   ├── cli.py                 # Command-line interface
    │   └── tools/
    │       ├── __init__.py
    │       ├── base.py            # Base tool class
    │       ├── recon.py           # Reconnaissance tools
    │       ├── scanning.py        # Scanning tools
    │       ├── vulnerability.py   # Vulnerability assessment
    │       ├── webapp.py          # Web application tools
    │       ├── network.py         # Network security tools
    │       ├── osint.py           # OSINT tools
    │       ├── exploitation.py    # Exploitation tools
    │       └── reporting.py       # Reporting tools
    ├── wordlists/                 # Wordlists for scanning
    ├── output/                    # Scan results and reports
    ├── data/                      # Persistent data storage
    ├── pyproject.toml             # Project configuration
    ├── README.md                  # This file
    ├── LICENSE                    # MIT License
    └── SECURITY.md               # Security guidelines

    🔒 Security Considerations

    Responsible Usage

    This tool is designed for authorized security testing only. Users must:

    1. Obtain explicit permission before testing any systems

    2. Comply with local laws and regulations

    3. Respect rate limits and avoid DoS conditions

    4. Follow responsible disclosure for any vulnerabilities found

    Safety Features

    • Target Whitelisting: Configure allowed targets
    • Rate Limiting: Prevent overwhelming target systems
    • Safe Mode: Enable additional safety checks
    • Logging: Comprehensive audit trails

    Legal Disclaimer

    Users are solely responsible for ensuring their use of this tool complies with applicable laws and regulations. Gokul (apgokul008@gmail.com) is not responsible for any misuse or damage caused by this software.

    🤝 Contributing

    We welcome contributions! Please see our Contributing Guidelines for details.

    Development Setup

    1. Clone and install in development mode:

    bash
    git clone https://github.com/gokulapap/bugbounty-mcp-server.git
       cd bugbounty-mcp-server
       pip install -e ".[dev]"

    2. Install pre-commit hooks:

    bash
    pre-commit install

    3. Run tests:

    bash
    pytest

    📄 License

    This project is licensed under the MIT License - see the LICENSE file for details.

    🙏 Acknowledgments

    • OWASP for security testing methodologies
    • ProjectDiscovery for excellent security tools
    • SecLists for comprehensive wordlists
    • The bug bounty and security research community

    📚 Documentation

    • **RUN_SCRIPT.md** - Detailed run.sh script documentation
    • **USAGE.md** - Comprehensive usage examples and workflows
    • **SECURITY.md** - Security guidelines and best practices
    • **env.example** - Environment configuration template

    📞 Support

    • Issues: GitHub Issues
    • Discussions: GitHub Discussions
    • Security: See SECURITY.md for reporting security issues

    🚀 Roadmap

    • [ ] Web-based dashboard
    • [ ] Integration with popular bug bounty platforms
    • [ ] Machine learning-powered vulnerability detection
    • [ ] Collaborative testing features
    • [ ] Advanced evasion techniques
    • [ ] Mobile application testing tools
    • [ ] Cloud security assessment tools
    • [ ] Blockchain security testing

    ---

    ⚠️ Warning: This tool is for authorized security testing only. Unauthorized use against systems you don't own or have explicit permission to test is illegal and unethical.

    Similar MCP

    Based on tags & features

    • AD

      Adls Mcp Server

      Python·
      4
    • BO

      Books Mcp Server

      Python·
      5
    • CH

      Chuk Mcp Linkedin

      Python00
    • PU

      Pursuit Mcp

      Python00

    Trending MCP

    Most active this week

    • PL

      Playwright Mcp

      TypeScript·
      22.1k
    • SE

      Serena

      Python·
      14.5k
    • MC

      Mcp Playwright

      TypeScript·
      4.9k
    • MC

      Mcp Server Cloudflare

      TypeScript·
      3.0k
    View All MCP Servers

    Similar MCP

    Based on tags & features

    • AD

      Adls Mcp Server

      Python·
      4
    • BO

      Books Mcp Server

      Python·
      5
    • CH

      Chuk Mcp Linkedin

      Python00
    • PU

      Pursuit Mcp

      Python00

    Trending MCP

    Most active this week

    • PL

      Playwright Mcp

      TypeScript·
      22.1k
    • SE

      Serena

      Python·
      14.5k
    • MC

      Mcp Playwright

      TypeScript·
      4.9k
    • MC

      Mcp Server Cloudflare

      TypeScript·
      3.0k