AI支援リバースエンジニアリングのためのモデルコンテキストプロトコル(MCP)サーバーを提供するGhhidra拡張

ReVa(Reverse Engineering Assistant)は、AI言語モデルがGhidraの強力なリバースエンジニアリング機能と連携できるようにするGhidra MCPサーバーです。ReVaの使用 コンテキストロットを抑え、可能にする最先端の技術 長編のリバースエンジニアリングタスク。

ReVaは、REタスクのためのAIアシスタントを構築する他の取り組みとは異なり、_tool主導のapproach_を用いています 効果的なLLM利用のためのツール設計について。 ReVaは、あなたのRE環境が提供するのと同じように、LLMに対して多様な小型ツールを提供することを目指しています あなたに。

LLMに与えられたツールは、LLMが使いやすく、さまざまな入力に耐えられるように構築されています そしてLLMによる幻覚を減らすために。これは、LLMにスキーマを提供しつつ、他の入力も許容することで実現します。 LLMを導く説明、修正可能な誤りをLLMに戻す、さらに追加情報も含めます LLMの次の意思決定を導くための出力です。

ReVaのツールは他のソリューションとは異なり、より小型のものを提供します。 重要断片と補強、他とのリンク

ReVa - Ghidra MCP Server for AI-Powered Reverse Engineering

A Ghidra extension that provides a Model Context Protocol (MCP) server for AI-assisted reverse engineering

ReVa (Reverse Engineering Assistant) is a Ghidra MCP server that enables AI language models to interact with Ghidra’s powerful reverse engineering capabilities. ReVa uses state of the art techniques to limit context rot and enable long form reverse engineering tasks.

ReVa is different from other efforts at building AI assistants for RE tasks because it uses a tool driven approach with a focus on designing tools for effective LLM use. ReVa aims to provide a variety of small tools to the LLM, just as your RE environment provides a set of small tools to you.

Each of the tools given to the LLM are constructed to be easy for the LLM to use and to tolerate a variety of inputs and to reduce hallucination by the LLM. We do this by providing the LLM with a schema but tolerating other input, including descriptions that guide the LLM,and redirecting correctable mistakes back to the LLM, and including extra output to guide the next decision by the LLM.

ReVa’s tools differ to other solutions, they provide smaller, critical fragments with reinforcement and links to other relevant information to reduce context usage and hallucination. This greatly improves performance, especially on long form reverse engineering tasks. This allows ReVa to handle large binaries and even entire firmware images.

To encourage exploration as a human would, we report additional context like the namespace and cross references along with the decompilation, this is a small nudge to make the LLM explore the binary in the same way a human would.

Using this technique you can ask general questions and get relevant answers. The model prioritises information from the tools, but when there is no information it can still respond to generic questions from its training.

As an MCP server, ReVa can be used alongside other MCP servers to enrich its analysis. For example you can use the GitHub MCP Server to allow ReVa access to source code on GitHub, or the Kagi MCP Server to allow ReVa to search the web.

You can ask questions like:

• Examine the programs in this project and explain the relationship between the main binary and the shared libraries.
• What are the interesting strings in this program?
• Does this program use encryption? Write a markdown report on the encryption and where it is used.
• Draw a class diagram using plantuml syntax.
• Start from main, examine the program in detail. Rename variables as you go and provide a summary of the program.
• Explain the purpose of the __mod_init segment.
• What does mmap return?
• What does the function at address 0x80000 do?
• This is a CTF problem. Write a pwntools script to get the flag.