Lanalyzer is an advanced Python static taint analysis tool designed to detect potential security vulnerabilities in Python projects. It identifies data flows from untrusted sources (Sources) to sensitive operations (Sinks) and provides detailed insights into potential risks.

• ✨ Features
• 🚀 Getting Started
  • Prerequisites
  • Installation
• 💻 Usage
  • Basic Analysis
  • Command-Line Options
  • Example
• 🧩 Model Context Protocol (MCP) Support
  • Installing MCP Dependencies
  • Starting the MCP Server
  • MCP Server Features
  • Integration with AI Tools
  • Using in Cursor
  • MCP Command-Line Options
  • Advanced MCP Usage
• 🤝 Contributing
• 📄 License
• 📞 Contact

• Taint Analysis: Tracks data flows from sources to sinks.
• Customizable Rules: Define your own sources, sinks, sanitizers, and taint propagation paths.
• Static Analysis: No need to execute the code.
• Extensibility: Easily add new rules for detecting vulnerabilities like SQL Injection, XSS, and more.
• Detailed Reports: Generate comprehensive analysis reports with vulnerability details and mitigation suggestions.
• Command-Line Interface: Run analyses directly from the terminal.

• Python 3.10 or higher
• uv (recommended for dependency management)

1. Clone the repository:

   git clone https://github.com/mxcrafts/lanalyzer.git
   cd lanalyzer

2. Create a virtual environment and install dependencies:

   uv venv
   uv pip sync pyproject.toml --all-extras

3. Activate the virtual environment:

   source .venv/bin/activate

Run a taint analysis on a Python file:

lanalyzer --target <target_file> --config <config_file> --pretty --output <output_file> --log-file <log_file> --debug

• --target: Path to the Python file or directory to analyze.
• --config: Path to the configuration file.
• --output: Path to save the analysis report.
• --log-file: Path to save the log file.
• --pretty: Pretty-print the output.
• --detailed: Show detailed analysis statistics.
• --debug: Enable debug mode with detailed logging.

lanalyzer --target example.py --config rules/sql_injection.json --pretty --output example_analysis.json --log-file example_analysis.log --debug

We welcome contributions! Please see the CONTRIBUTING.md file for guidelines on how to contribute to Lanalyzer.

This project is licensed under the GNU Affero General Public License v3.0. See the LICENSE file for details.

• Issues: GitHub Issues
• Email: [email protected]

Lanalyzer now supports the Model Context Protocol (MCP), allowing it to run as an MCP server that AI models and tools can use to access taint analysis functionality through a standard interface.

If you're using pip:

pip install "lanalyzer[mcp]"

If you're using uv:

uv pip install -e ".[mcp]"

There are multiple ways to start the MCP server:

1. Using Python Module:

# View help information
python -m lanalyzer.mcp --help

# Start the server
python -m lanalyzer.mcp run --host 0.0.0.0 --port 8000 --debug

2. Using the lanalyzer Command-Line Tool:

# View help information
lanalyzer mcp --help

# Start the server
lanalyzer mcp run --host 0.0.0.0 --port 8000 --debug

# Use FastMCP development mode (if applicable, verify this command)
# lanalyzer mcp dev

The MCP server provides the following core functionalities:

1. Code Analysis: Analyze Python code strings for security vulnerabilities
2. File Analysis: Analyze specific files for security vulnerabilities
3. Path Analysis: Analyze entire directories or projects for security vulnerabilities
4. Vulnerability Explanation: Provide detailed explanations of discovered vulnerabilities
5. Configuration Management: Get, validate, and create analysis configurations

The MCP server can be integrated with AI tools that support the MCP protocol:

# Using the FastMCP client
from fastmcp import FastMCPClient

# Create a client connected to the server
client = FastMCPClient("http://127.0.0.1:8000")

# Analyze code
result = client.call({
    "type": "analyze_code",
    "code": "user_input = input()\nquery = f\"SELECT * FROM users WHERE name = '{user_input}'\"",
    "file_path": "example.py",
    "config_path": "/path/to/config.json"
})

# Print analysis results
print(result)

If you're working in the Cursor editor, you can directly ask the AI to use Lanalyzer to analyze your code:

Please use lanalyzer to analyze the current file for security vulnerabilities and explain the potential risks.

The MCP server supports the following command-line options:

• --debug: Enable debug mode with detailed logging
• --host: Set the server listening address (default: 127.0.0.1)
• --port: Set the server listening port (default: 8000)

You can use the get_config, validate_config, and create_config tools to manage vulnerability detection configurations:

# Get the default configuration
config = client.call({
    "type": "get_config"
})

# Create a new configuration
result = client.call({
    "type": "create_config",
    "config_data": {...},  # Configuration data
    "config_path": "/path/to/save/config.json"  # Optional
})

Analyze an entire project or directory:

result = client.call({
    "type": "analyze_path",
    "target_path": "/path/to/project",
    "config_path": "/path/to/config.json",
    "output_path": "/path/to/output.json"  # Optional
})