This project connects AI agents like Claude, Copilot, and Cursor directly to Cheat Engine via the Model Context Protocol (MCP), enabling natural language interaction for automating reverse engineering, pointer scanning, and memory analysis. It targets users needing to accelerate tasks such as mod creation, trainer development, security auditing, and game botting, by transforming complex memory analysis into conversational queries.

How It Works

The system employs a client-server architecture. An AI agent communicates with a Python MCP server (mcp_cheatengine.py) using JSON-RPC over standard input/output. This server then interfaces with Cheat Engine through Windows-specific named pipes. Cheat Engine's Lua script (ce_mcp_bridge.lua) acts as the bridge, translating AI commands into Cheat Engine API calls. This allows the AI to read memory, follow pointer chains, analyze structures, disassemble functions, and set hardware breakpoints, leveraging Cheat Engine's advanced features like DBVM mode for deep system analysis.

Quick Start & Requirements

• Installation: Install dependencies via pip install -r MCP_Server/requirements.txt or manually with pip install mcp pywin32.
• Prerequisites: Windows operating system is mandatory due to the use of pywin32 for named pipes. Cheat Engine (version 7.6 or later recommended based on linked documentation) is required. DBVM must be enabled within Cheat Engine.
• Setup: Load the ce_mcp_bridge.lua script in Cheat Engine. Configure your MCP client (e.g., mcp_config.json) to include the Cheat Engine server details, pointing to the Python script. Restart your IDE to apply the configuration.
• Documentation: Command reference (AI_Context/MCP_Bridge_Command_Reference.md), Cheat Engine Lua documentation (AI_Context/CE_LUA_Documentation.md), and AI server implementation guide (AI_Context/AI_Guide_MCP_Server_Implementation.md) are available.

Highlighted Details

• Automated memory analysis and pointer chain resolution via natural language prompts.
• AI-driven identification of C++ object types using RTTI and automatic structure dissection.
• Support for invisible debugging through hardware breakpoints and Ring -1 hypervisor capabilities via DBVM.
• Comprehensive toolset for memory reading, AOB scanning, function analysis, and cross-reference finding.

Licensing & Compatibility

The specific open-source license is not explicitly stated in the README. The project is explicitly marked for "educational and research purposes only," and the README includes a disclaimer against its use for malicious hacking, cheating, or violating terms of service. Compatibility is restricted to Windows.

Limitations & Caveats

This project is strictly limited to Windows environments. A critical configuration step involves disabling Cheat Engine's "Query memory region routines" to prevent system instability (BSODs) when using DBVM with protected memory pages. The project is presented as a demonstration of software engineering automation and MCP capabilities, rather than a production-ready tool for all use cases.