frida-mcp

frida-mcp

An MCP server that enables AI models to perform Android dynamic analysis using Frida, including spawning and attaching to applications, listing apps, and injecting scripts.

Category
Visit Server

README

Frida MCP Server | Frida MCP 服务器

English | 中文

English

A Model Context Protocol (MCP) server that enables AI models to perform Android dynamic analysis using Frida.

Project Structure

frida-mcp/
├── frida_mcp.py          # Core MCP server implementation
├── pyproject.toml        # Project dependencies and configuration  
├── requirements.txt      # Alternative dependency file
├── config.json           # Optional Frida server configuration
├── README.md            # Documentation
└── .gitignore           # Git ignore rules

Core Files

  • frida_mcp.py: Main MCP server with Frida integration
  • pyproject.toml: Modern Python project configuration (recommended)
  • requirements.txt: Traditional dependency file
  • config.json: Optional configuration for frida-server settings

Installation

# Clone repository
git clone http://git.dev.sh.ctripcorp.com/octopus/frida-mcp.git
cd frida-mcp

# Install dependencies (choose one method)
# Method 1: Using pip + requirements.txt
pip install -r requirements.txt

# Method 2: Using pip + pyproject.toml (recommended)
pip install -e .

# Setup frida-server on Android device
adb push frida-server /data/local/tmp/
adb shell "chmod 755 /data/local/tmp/frida-server"
adb shell "su -c /data/local/tmp/frida-server &"

MCP Configuration

Add to your MCP client configuration (e.g., Claude Desktop config file):

Windows (%APPDATA%\Claude\claude_desktop_config.json):

{
  "mcpServers": {
    "frida": {
      "command": "python",
      "args": ["C:\\Users\\YourName\\frida-mcp\\frida_mcp.py"],
      "transport": "stdio"
    }
  }
}

macOS/Linux (~/.config/claude/claude_desktop_config.json):

{
  "mcpServers": {
    "frida": {
      "command": "python",
      "args": ["/home/username/frida-mcp/frida_mcp.py"],
      "transport": "stdio"
    }
  }
}

Configuration (Optional)

The config.json file contains optional Frida server configuration:

{
  "server_path": "/data/local/myfr",
  "server_name": "aaabbb", 
  "server_port": 27042,
  "device_id": null,
  "adb_path": "adb"
}
  • server_path: Custom path for frida-server on Android device
  • server_name: Custom frida-server binary name
  • server_port: Port for Frida server communication
  • device_id: Specific device ID (null for auto-detection)
  • adb_path: Path to ADB executable

Available Tools

spawn(package_name, initial_script?, wait_seconds?, max_output_messages?)

Start an Android application with optional script injection.

attach(target, initial_script?, wait_seconds?, max_output_messages?)

Attach to a running process with optional script injection.

get_frontmost_application()

Get the currently active application.

list_applications()

List all installed applications.

Example Usage

// Hook HashMap operations
Java.perform(function() {
    var HashMap = Java.use("java.util.HashMap");
    HashMap.put.implementation = function(key, value) {
        console.log("HashMap.put:", key, value);
        return this.put(key, value);
    };
});

中文

一个 Model Context Protocol (MCP) 服务器,使 AI 模型能够使用 Frida 进行 Android 动态分析。

项目结构

frida-mcp/
├── frida_mcp.py          # MCP 服务器核心实现
├── pyproject.toml        # 项目依赖和配置文件
├── requirements.txt      # 传统依赖文件
├── config.json           # 可选的 Frida 服务器配置
├── README.md            # 文档说明
└── .gitignore           # Git 忽略规则

核心文件

  • frida_mcp.py: 集成 Frida 的主要 MCP 服务器
  • pyproject.toml: 现代 Python 项目配置(推荐使用)
  • requirements.txt: 传统依赖文件
  • config.json: Frida 服务器设置的可选配置文件

安装

# 克隆仓库
git clone http://git.dev.sh.ctripcorp.com/octopus/frida-mcp.git
cd frida-mcp

# 安装依赖(选择一种方法)
# 方法1:使用 pip + requirements.txt
pip install -r requirements.txt

# 方法2:使用 pip + pyproject.toml(推荐)
pip install -e .

# 在 Android 设备上设置 frida-server
adb push frida-server /data/local/tmp/
adb shell "chmod 755 /data/local/tmp/frida-server"
adb shell "su -c /data/local/tmp/frida-server &"

MCP 配置

添加到您的 MCP 客户端配置(如 Claude Desktop 配置文件):

Windows (%APPDATA%\Claude\claude_desktop_config.json):

{
  "mcpServers": {
    "frida": {
      "command": "python",
      "args": ["C:\\Users\\你的用户名\\frida-mcp\\frida_mcp.py"],
      "transport": "stdio"
    }
  }
}

注意:将 你的用户名 替换为实际的 Windows 用户名,路径使用双反斜杠 \\

macOS/Linux (~/.config/claude/claude_desktop_config.json):

{
  "mcpServers": {
    "frida": {
      "command": "python",
      "args": ["/home/用户名/frida-mcp/frida_mcp.py"],
      "transport": "stdio"
    }
  }
}

注意:将 用户名 替换为实际的系统用户名


### 配置说明(可选)

`config.json` 文件包含可选的 Frida 服务器配置:

```json
{
  "server_path": "/data/local/myfr",
  "server_name": "aaabbb", 
  "server_port": 27042,
  "device_id": null,
  "adb_path": "adb"
}
  • server_path: Android 设备上 frida-server 的自定义路径
  • server_name: frida-server 二进制文件的自定义名称
  • server_port: Frida 服务器通信端口
  • device_id: 指定设备 ID(null 为自动检测)
  • adb_path: ADB 可执行文件路径

可用工具

spawn(package_name, initial_script?, wait_seconds?, max_output_messages?)

启动 Android 应用程序,可选注入脚本。

  • package_name: 应用包名
  • initial_script: 可选的 JavaScript 脚本
  • wait_seconds: 等待输出的时间(默认 1.5 秒)
  • max_output_messages: 最大输出消息数(默认 100)

attach(target, initial_script?, wait_seconds?, max_output_messages?)

附加到运行中的进程,可选注入脚本。

  • target: 进程名或 PID
  • initial_script: 可选的 JavaScript 脚本
  • wait_seconds: 等待输出的时间(默认 1.0 秒)
  • max_output_messages: 最大输出消息数(默认 100)

get_frontmost_application()

获取当前活跃的应用程序。

list_applications()

列出所有已安装的应用程序。

使用示例

// Hook HashMap 操作
Java.perform(function() {
    var HashMap = Java.use("java.util.HashMap");
    HashMap.put.implementation = function(key, value) {
        console.log("HashMap.put:", key, value);
        return this.put(key, value);
    };
});

特性

  • 🚀 自动 Gson 对象序列化
  • 🔍 console.log 自动重定向
  • 📱 自动设备连接管理
  • 🤖 为 AI 交互优化

常见问题

Q: 应用崩溃怎么办? A: 减少 hook 频率,避免复杂序列化操作。

Q: 没有输出? A: 确认方法被调用,spawn 时脚本在应用启动前注入。

Q: 连接失败? A: 检查 frida-server 是否运行:adb shell ps | grep frida

Requirements

  • Python 3.8+
  • Android 设备 (root)
  • 查看 requirements.txt

License

MIT

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured