EMBA-MCP
An MCP server for EMBA firmware analysis that exposes structured security findings and tools to LLMs. It enables users to programmatically query, reason over, and correlate firmware analysis results such as kernel details, SBOMs, and attack paths.
README
🧠 EMBA-MCP
Model Context Protocol (MCP) server for EMBA firmware analysis. EMBA-MCP exposes EMBA firmware analysis results as structured tools via MCP, allowing LLMs (Claude, ChatGPT, etc.) to query, reason, and correlate firmware security findings programmatically. It parses, normalizes, and reasons over EMBA output.
✨ Features
📦 Parse EMBA results (kernel, services, credentials, crypto, SBOM, binaries, PHP, etc.) 🔍 Filesystem-aware analysis (SUID, secrets, weak crypto, services) 🚨 High-risk correlation engine (multi-signal findings) 🧭 Attack-path explanation engine 🧠 MCP-native tools (plug into Claude / MCP clients) 🔁 Works with existing EMBA output (no re-scan required)
📋 Requirements System Linux (recommended: Ubuntu / Kali) Python 3.10+ EMBA Tool
🛠️ Installation
git clone https://github.com/0xbuz3R/EMBA-MCP.git
cd EMBA-MCP
Set Up a Virtual Environment (Recommended)
Bash
python -m venv venv
# On Windows:
.\venv\Scripts\activate
# On macOS:
source venv/bin/activate
3. Install Dependencies
Bash
pip install -e .
Make sure EMBA is installed in your local machine: https://github.com/e-m-b-a/emba
🧩 Claude Desktop MCP Configuration (Required)
To use EMBA-MCP inside Claude Desktop, you must register the MCP server in Claude’s config file. This step is mandatory and is the most common setup issue.
📍 1. Locate claude_config_desktop.json
Claude Desktop stores the MCP configuration in the following location:
Linux
~/.config/claude/claude_config_desktop.json
If the file does not exist, create it manually.
📍 2. Add EMBA-MCP Server Configuration
Add the following JSON under the mcpServers section.
⚠️ Do NOT copy paths blindly — replace them with paths valid on your system.
{
"mcpServers": {
"emba": {
"command": "<PATH_TO_PYTHON>",
"args": ["-m", "emba_mcp.mcp_server"],
"env": {
"EMBA_HOME": "<PATH_TO_EMBA_DIRECTORY>"
}
}
}
}
📌 3. How to Fill the Paths Correctly
🔹 PATH_TO_PYTHON
This must be the Python interpreter where EMBA-MCP is installed. Examples:
which python
which python3
Typical values:
/usr/bin/python3 /home/user/.venv/bin/python /home/user/.local/bin/python
🔹 PATH_TO_EMBA_DIRECTORY
This must be the root directory of EMBA, where the emba executable exists.
Example:
ls <EMBA_HOME>/emba
Example value:
/home/user/tools/emba
📍 4. Restart Claude Desktop
After saving the file:
- Fully close Claude Desktop
- Reopen it
Claude will now auto-load the EMBA MCP server
✅ 5. Verify MCP Is Loaded
Inside Claude, try:
List EMBA scans or Run EMBA scan on firmware
If configured correctly, Claude will respond without MCP errors.
📽️ Demo
🤝 Contributing & Feedback
I am still working on a few other data items to integrate; I'll update in the next release.
Meanwhile, please give it a try and provide feedback on how to improve! You can leave your suggestions in our https://github.com/0xbuz3R/EMBA-MCP/issues/1
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。