GhydraMCP

GhydraMCP is a bridge between Ghidra and AI assistants that enables AI-assisted reverse engineering through the Model Context Protocol (MCP).

Overview

GhydraMCP consists of:

1. Ghidra Plugin: Exposes Ghidra's powerful reverse engineering capabilities through a REST API
2. MCP Bridge: A Python script that translates MCP requests into API calls
3. Multi-instance Support: Connect multiple Ghidra instances to analyze different binaries simultaneously

This allows AI assistants like Claude to directly:

• Decompile functions and analyze binary code
• Understand program structure, function relationships, and data types
• Perform binary analysis tasks (identify cross-references, data flow, etc.)
• Make meaningful changes to the analysis (rename functions, add comments, etc.)

GhydraMCP is based on GhidraMCP by Laurie Wired with added multi-instance support and numerous enhancements.

Features

GhydraMCP combines a Ghidra plugin with an MCP server to provide a comprehensive set of reverse engineering capabilities to AI assistants:

Program Analysis

• Decompilation: Convert binary functions to readable C code
• Static Analysis:
  • Cross-reference analysis (find who calls what)
  • Data flow analysis
  • Type propagation and reconstruction
• Symbol Management:
  • View and analyze imports and exports
  • Identify library functions and dependencies

Interactive Reverse Engineering

• Code Understanding:
  • Explore function code and relationships
  • Analyze data structures and types
• Annotation:
  • Rename functions, variables, and data
  • Add comments and documentation
  • Create and modify data types

Multi-instance Support

• Run multiple Ghidra instances simultaneously
• Analyze different binaries in parallel
• Connect to specific instances using port numbers

Program Navigation

• List and search functions, classes, and namespaces
• View memory segments and layout
• Search by name, pattern, or signature

Installation

Prerequisites

• Install Ghidra
• Python3
• MCP SDK

Ghidra

First, download the latest release from this repository. The "Complete" artifact contains the zipped Ghidra plugin and the Python MCP bridge. Unpack the outer archive, then, add the plugin to Ghidra:

1. Run Ghidra
2. Select File -> Install Extensions
3. Click the + button
4. Select the GhydraMCP-1.1.zip (or your chosen version) from the downloaded release
5. Restart Ghidra
6. Make sure the GhydraMCPPlugin is enabled in File -> Configure -> Developer

Note: By default, the first CodeBrowser opened in Ghidra gets port 8192, the second gets 8193, and so on. You can check which ports are being used by looking at the Console in the Ghidra main (project) window - click the computer icon in the bottom right to "Open Console". Look for log entries like:

(HydraMCPPlugin) Plugin loaded on port 8193
(HydraMCPPlugin) HydraMCP HTTP server started on port 8193

Video Installation Guide:

https://github.com/user-attachments/assets/75f0c176-6da1-48dc-ad96-c182eb4648c3

MCP Clients

Theoretically, any MCP client should work with GhydraMCP. Two examples are given below.

API Reference

Available Tools

Program Analysis:

• list_methods: List all functions (params: offset, limit)
• list_classes: List all classes/namespaces (params: offset, limit)
• decompile_function: Get decompiled C code (params: name)
• rename_function: Rename a function (params: old_name, new_name)
• rename_data: Rename data at address (params: address, new_name)
• list_segments: View memory segments (params: offset, limit)
• list_imports: List imported symbols (params: offset, limit)
• list_exports: List exported functions (params: offset, limit)
• list_namespaces: Show namespaces (params: offset, limit)
• list_data_items: View data labels (params: offset, limit)
• search_functions_by_name: Find functions (params: query, offset, limit)

Instance Management:

• list_instances: List active Ghidra instances (no params)
• register_instance: Register new instance (params: port, url)
• unregister_instance: Remove instance (params: port)

Example Usage:

# Program analysis
client.use_tool("ghydra", "decompile_function", {"name": "main"})

# Instance management  
client.use_tool("ghydra", "register_instance", {"port": 8192, "url": "http://localhost:8192/"})
client.use_tool("ghydra", "register_instance", {"port": 8193})

Client Setup

Claude Desktop Configuration

{
  "mcpServers": {
    "ghydra": {
      "command": "uv",
      "args": [
        "run",
        "/ABSOLUTE_PATH_TO/bridge_mcp_hydra.py"
      ],
      "env": {
        "GHIDRA_HYDRA_HOST": "localhost"  // Optional - defaults to localhost
      }
    }
  }
}

Note: You can also use python instead of uv run, but then you'll need to manually install the requirements first with pip install mcp requests.

5ire Configuration

1. Tool Key: ghydra
2. Name: GhydraMCP
3. Command: uv run /ABSOLUTE_PATH_TO/bridge_mcp_hydra.py

Example Session

Below is an example of how an AI assistant might interact with multiple Ghidra instances using GhydraMCP:

User:

Register the following instances with the ghydra mcp tool:
port 8192 = application.exe
port 8193 = firmware.bin

Assistant:

I'll help you analyze these binaries. Let me register the instances with the ghydra mcp tool as requested.

View result from register_instance from ghydra (local){
  `port`: 8192
}
Registered instance on port 8192 at http://localhost:8192

View result from register_instance from ghydra (local){
  `port`: 8193
}
Registered instance on port 8193 at http://localhost:8193

Now that I've registered both instances, I'm ready to analyze the decompiled code.

User:

List the functions through ghydra for both instances and identify any related functionality.

Assistant:

I'll list the functions from both instances and analyze them.

View result from list_functions from ghydra (local){
  `port`: 8192
}
[Function list for application.exe...]

View result from list_functions from ghydra (local){
  `port`: 8193
}
[Function list for firmware.bin...]

I notice some interesting functions in both binaries. Let me decompile a few to understand their behavior.

View result from get_function from ghydra (local){
  `port`: 8192,
  `name`: "process_command"
}
[Decompiled code for process_command...]

View result from get_function from ghydra (local){
  `port`: 8193,
  `name`: "handle_message"
}
[Decompiled code for handle_message...]

Based on this analysis, I can see these binaries communicate using a simple protocol where...

Building from Source

You can build different artifacts with Maven:

Build Everything (Default)

Build both the Ghidra plugin and the complete package:

mvn clean package

This creates:

• target/GhydraMCP-[version].zip - The Ghidra plugin only
• target/GhydraMCP-Complete-[version].zip - Complete package with plugin and bridge script

Build Ghidra Plugin Only

If you only need the Ghidra plugin:

mvn clean package -P plugin-only

Build Complete Package Only

If you only need the combined package:

mvn clean package -P complete-only

The Ghidra plugin includes these files required for Ghidra to recognize the extension:

• lib/GhydraMCP.jar
• extension.properties
• Module.manifest