BinjaLattice MCP Server

BinjaLattice

BinjaLattice is a secure communication protocol for Binary Ninja that enables interaction with external Model Context Protocol (MCP) servers and tools. It provides a structured way to acquire information from Binary Ninja and the ability to modify an active Binary Ninja database over HTTP with a REST API.

Demo

Features

• Secure Authentication: Token-based authentication system
• Encrypted Communication: Optional SSL/TLS encryption
• Binary Analysis Context: Export pseudocode, disassembly, variable names, binary information etc.
• Binary Modification: Update function names, add comments, rename variables
• Token Management: Automatic expiration and renewal of authentication tokens

Installation

Windows (Automated)

Run the PowerShell installer for a one-shot setup:

.\scripts\install_windows.ps1

This will:

• Install the plugin to %APPDATA%\Binary Ninja\plugins\
• Create a Python virtual environment (.venv)
• Install all dependencies
• Output a ready-to-use MCP configuration

Manual Installation (All Platforms)

1. Copy plugin/lattice_server_plugin.py to your Binary Ninja plugins directory:

   • Linux: ~/.binaryninja/plugins/
   • macOS: ~/Library/Application Support/Binary Ninja/plugins/
   • Windows: %APPDATA%\Binary Ninja\plugins\

2. Create a virtual environment:

   python -m venv .venv
   

3. Activate and install dependencies:

   # Linux/macOS
   source .venv/bin/activate
   
   # Windows
   .venv\Scripts\activate
   
   pip install -r requirements.txt
   

Usage

Starting the Server in Binary Ninja

1. Open Binary Ninja and load a binary file
2. Go to Plugins > Start Lattice Protocol Server
3. The server will start and display the API key in the log console
4. Set the API key as the BNJLAT environment variable in your MCP configuration

Example MCP configuration (mcp.json):

{
    "mcpServers": {
      "binja-lattice-mcp": {
        "command": "/path/to/BinjaLattice/.venv/bin/python",
        "args": ["/path/to/BinjaLattice/mcp_server.py"],
        "env": {
            "BNJLAT": "your_api_key_here"
        }
      }
    }
}

On Windows, use backslashes:

{
    "mcpServers": {
      "binja-lattice-mcp": {
        "command": "C:\\path\\to\\BinjaLattice\\.venv\\Scripts\\python.exe",
        "args": ["C:\\path\\to\\BinjaLattice\\mcp_server.py"],
        "env": {
            "BNJLAT": "your_api_key_here"
        }
      }
    }
}

Tip: The Windows installer outputs a ready-to-paste configuration with the correct paths.

Available MCP Tools

The following tools are available through the MCP server:

Binary Information

• get_binary_info: Get metadata about the binary (filename, architecture, entry point, segments, sections, function count)
• get_all_function_names: List all function names in the binary
• get_strings: Get strings with optional min_length and substring filter
• get_imports: List imported functions with addresses and source libraries
• get_exports: List exported functions with addresses
• get_analysis_progress: Get Binary Ninja analysis status and progress percentage

Function Analysis

• get_function_disassembly: Get assembly instructions for function by name
• get_function_pseudocode: Get decompiled C-like pseudocode for function
• get_function_variables: Get parameters, local variables, and global variables
• get_cross_references_to_function: List functions that call the specified function
• get_call_graph: Get callers and callees of function with configurable depth
• get_global_variable_data: Read data from global variable referenced in function

Data Access

• get_data_at_address: Read bytes at address with optional type interpretation
• search_bytes: Search for hex byte pattern with wildcard support (e.g., '48 89 ?? 24')

Type Management

• get_types: List defined types (structs, enums, typedefs) with optional filter
• create_struct: Create a new struct type with JSON member definitions
• update_struct: Update an existing struct type

Annotations

• update_function_name: Rename a function
• update_variable_name: Rename a variable in a function
• set_variable_type: Set variable type annotation (C-style like 'uint32_t')
• set_function_signature: Set function prototype (C-style like 'int foo(char* arg1)')
• add_comment_to_address: Add comment at address
• add_comment_to_function: Add comment to function
• create_tag: Create tag at address with type and optional description
• get_tags: List all tags with optional type filter

Client Library Usage

The Lattice client library provides a Python interface for interacting with the BinjaLattice server:

from lib.lattice import Lattice

# Initialize c