cyberedu-mcp
This is the Oficial Model Context Protocol (MCP) server for the CyberEdu CTF platform (cyber-edu.co / cyberedu.ro)
Documentation
CyberEdu MCP Server
This is the Oficial Model Context Protocol (MCP) server for the CyberEdu CTF platform (https://cyber-edu.co / https://cyberedu.ro). This server automatically discovers and exposes all methods from the `CyberEduClient` as MCP tools, making it easy to interact with the CyberEdu platform through MCP-compatible clients.
Features
- Dynamic Tool Discovery: Automatically discovers all public methods from `CyberEduClient` and exposes them as MCP tools
- Zero Configuration for New Methods: When new methods are added to `CyberEduClient`, they automatically become available as MCP tools without any code changes
- Type-Safe: Automatically generates JSON schemas from method signatures and type hints
- Error Handling: Comprehensive error handling with detailed error messages
CyberEDU Platform Overview
CyberEDU is a cybersecurity training platform that provides hands-on labs, realistic simulations, and competitive environments. It’s designed for enterprise security teams, academic institutions, government agencies, and
individual learners.
Core Description
CyberEDU is a comprehensive cybersecurity training platform that provides hands-on labs, realistic simulations, and competitive environments. It's designed for enterprise security teams, academic institutions, government
agencies, and individual learners who want to develop practical cybersecurity skills through real-world scenarios.
Key Differentiators
- Hands-on approach: Interactive cyber ranges where users attack and defend real infrastructure (not just videos or theory)
- MITRE ATT&CK mapping: Scenarios mapped to MITRE ATT&CK, using real malware samples (safely contained)
- Better retention: 3.5x better skill retention compared to passive learning
- Real-world scenarios: Simulates actual adversary techniques and attack patterns
Platform Components
1. Cyber Range — Enterprise-scale cyber warfare simulation with complex network topologies
2. Cyber Labs — 650+ hands-on labs mapped to MITRE ATT&CK, browser-based and auto-graded
3. Tournament Suite — Gamified competitions (CTFs, Red vs Blue, war games)
Key Statistics
- 30,000+ active users worldwide
- 650+ hands-on labs
- 1,400+ simulation profiles
- 500+ events hosted
- 45+ countries served
- 250+ hours of training content
Target Audiences
- Students: Career-focused training with CTF challenges and leaderboards
- Academia: Curriculum with LMS integration and auto-grading
- Enterprise: Technical hiring assessments, team training, compliance mapping
- Government: Air-gapped deployments, OT/SCADA simulation, critical infrastructure defense
Deployment Options
- Cloud-hosted SaaS (browser-based, no installation)
- On-premise (VMware, Proxmox, bare-metal)
- Air-gapped deployments for classified environments
Installation
Clone the Repository
The `cyberedu-client` is included as a git submodule. Clone with `--recursive` to get everything:
git clone --recursive https://github.com/CyberEDU-Cyber-Range/cyberedu-mcp.git
cd cyberedu-mcpIf you already cloned without `--recursive`, initialize the submodule:
git submodule update --init --recursiveInstall Packages
Install both packages (client and MCP server):
macOS/Linux:
python3 -m venv venv
source venv/bin/activate
pip install -e ".[local]" # Installs with local cyberedu-client submodule
# Or for development:
# pip install -e ".[local,dev]"Windows (PowerShell):
python -m venv venv
.\venv\Scripts\Activate.ps1
pip install -e ".[local]" # Installs with local cyberedu-client submoduleWindows (Command Prompt):
python -m venv venv
venv\Scripts\activate.bat
pip install -e ".[local]"> Alternative: Install packages separately:
> ```bash
> pip install -e ./cyberedu-client
> pip install -e .
> ```
Configuration
Session Persistence (Recommended)
The MCP server automatically persists session credentials to disk. This means:
- Set your cookie once using the `cyberedu_set_session_cookie` tool, and it will be remembered across MCP sessions
- No environment variables needed after the first authentication
- Tenant selection is preserved when you switch tenants
Session file location:
- macOS/Linux: `~/.cyberedu-mcp/session.json`
- Windows: `%USERPROFILE%\.cyberedu-mcp\session.json` (e.g., `C:\Users\YourName\.cyberedu-mcp\session.json`)
The file has restricted permissions (owner read/write only) for security on Unix systems.
Environment Variables (Alternative)
You can also use environment variables. The server loads credentials in this priority order:
1. Persisted credentials from disk (highest priority)
2. Environment variables
3. Defaults
Environment variables:
- `CYBEREDU_SESSION_COOKIE`: Your CyberEdu session cookie
- Get this from your browser's developer tools after logging in to https://app.cyber-edu.co
- Look for the `cyberedu_session` cookie value
- `CYBEREDU_TENANT`: Your tenant identifier (optional, defaults to "cyberedu")
Getting Your Session Cookie
Chrome/Edge:
1. Open Developer Tools (F12)
2. Go to Application/Storage tab
3. Navigate to Cookies → `https://app.cyber-edu.co`
4. Find `cyberedu_session` and copy its value
Firefox:
1. Open Developer Tools (F12)
2. Go to Storage tab
3. Navigate to Cookies → `https://app.cyber-edu.co`
4. Find `cyberedu_session` and copy its value
Usage
Running the MCP Server
The server can be run directly (for testing):
macOS/Linux:
python3 -m venv venv
source venv/bin/activate
python -m cyberedu_mcpWindows:
python -m venv venv
.\venv\Scripts\Activate.ps1
python -m cyberedu_mcpMCP Client Configuration
To use this server with an MCP client (Cursor IDE or Claude Desktop), add it to your MCP configuration.
Important: Use the full path to the Python executable in your venv. MCP clients run servers externally and won't have access to an activated virtual environment.
macOS/Linux Examples
Cursor IDE (`~/.cursor/mcp.json`):
{
"mcpServers": {
"cyberedu": {
"command": "/path/to/cyberedu-mcp/venv/bin/python3",
"args": ["-m", "cyberedu_mcp"]
}
}
}Claude Desktop (`~/Library/Application Support/Claude/claude_desktop_config.json`):
{
"mcpServers": {
"cyberedu": {
"command": "/path/to/cyberedu-mcp/venv/bin/python3",
"args": ["-m", "cyberedu_mcp"]
}
}
}Windows Examples
Cursor IDE (`%APPDATA%\Cursor\User\mcp.json` or `C:\Users\YourName\.cursor\mcp.json`):
{
"mcpServers": {
"cyberedu": {
"command": "C:\\path\\to\\cyberedu-mcp\\venv\\Scripts\\python.exe",
"args": ["-m", "cyberedu_mcp"]
}
}
}Claude Desktop (`%APPDATA%\Claude\claude_desktop_config.json`):
{
"mcpServers": {
"cyberedu": {
"command": "C:\\path\\to\\cyberedu-mcp\\venv\\Scripts\\python.exe",
"args": ["-m", "cyberedu_mcp"]
}
}
}Cross-Platform Examples
VS Code (`.vscode/mcp.json` in your workspace):
{
"servers": {
"cyberedu": {
"type": "stdio",
"command": "/path/to/cyberedu-mcp/venv/bin/python3",
"args": ["-m", "cyberedu_mcp"]
}
}
}*Windows: Use `C:\\path\\to\\cyberedu-mcp\\venv\\Scripts\\python.exe`*
Antigravity / Windsurf (`mcp_config.json` - access via MCP store → Manage MCP Servers → View raw config):
{
"mcpServers": {
"cyberedu": {
"command": "/path/to/cyberedu-mcp/venv/bin/python3",
"args": ["-m", "cyberedu_mcp"],
"env": {}
}
}
}*Windows: Use `C:\\path\\to\\cyberedu-mcp\\venv\\Scripts\\python.exe`*
Note: Session credentials are persisted to `~/.cyberedu-mcp/session.json`, so no environment variables are needed after the first authentication via the `cyberedu_set_session_cookie` tool.
Available Tools
The server automatically exposes all public methods from CyberEduClient as MCP tools. Tools are prefixed with `cyberedu_` to avoid naming conflicts.
Session Management Tools
These tools allow you to manage authentication and tenant switching without restarting the MCP server. Credentials are automatically persisted to `~/.cyberedu-mcp/session.json`:
- `cyberedu_get_session_status` - Check if authenticated, which tenant is selected, and if credentials are persisted
- `cyberedu_set_session_cookie` - Set/update the session cookie for authentication (persists to disk)
- `cyberedu_switch_tenant` - Switch to a different tenant/organization (persists to disk)
- `cyberedu_clear_session` - Clear stored credentials from disk and memory
Example usage:
1. Check status: "What's my CyberEdu session status?"
2. Set cookie: "Set my CyberEdu session cookie to `eyJ...`" (only needed once, persists!)
3. Switch tenant: "Switch to tenant `myorg`"
4. Clear credentials: "Clear my CyberEdu session"
Authentication & User Tools
- `cyberedu_check_auth` - Verify authentication and get user info
- `cyberedu_get_user_info` - Get full user information
- `cyberedu_list_tenants` - List all available tenants
- `cyberedu_get_current_tenant_info` - Get current tenant information
- `cyberedu_get_user` - Get user information by ID
Challenge Tools (Archive)
- `cyberedu_list_challenges` - List all challenges (with optional filters)
- `cyberedu_get_challenge` - Get challenge details
- `cyberedu_get_challenge_difficulties` - Get available difficulty levels
- `cyberedu_get_challenge_tags` - Get available challenge tags
- `cyberedu_subscribe_to_challenge` - Subscribe to a challenge
Flag & Submission Tools (Archive)
- `cyberedu_get_flag` - Get flag/question information
- `cyberedu_submit_flag` - Submit a flag/answer
File Tools (Archive)
- `cyberedu_download_file` - Download a challenge file (use optional `save_path` param to save directly to disk)
Service Tools (Archive)
- `cyberedu_start_service` - Start a challenge service
- `cyberedu_get_service_status` - Get service status
- `cyberedu_extend_service` - Extend service time
- `cyberedu_restart_service` - Restart service
Contest Tools
- `cyberedu_list_contests` - List all available contests
- `cyberedu_get_contest` - Get contest details
- `cyberedu_get_contest_ranks` - Get contest leaderboard
- `cyberedu_get_contest_challenge` - Get challenge details within a contest
- `cyberedu_subscribe_to_contest_challenge` - Subscribe to a contest challenge
Contest Flag & Submission Tools
- `cyberedu_get_contest_flag` - Get flag information within a contest
- `cyberedu_submit_contest_flag` - Submit a flag within a contest
Contest File Tools
- `cyberedu_download_contest_file` - Download a file from a contest challenge (use optional `save_path` param to save directly to disk)
Contest Service Tools
- `cyberedu_start_contest_service` - Start a service within a contest
- `cyberedu_get_contest_service_status` - Get service status within a contest
- `cyberedu_extend_contest_service` - Extend service time within a contest
- `cyberedu_restart_contest_service` - Restart service within a contest
Usage Examples & Prompts
Example prompts for interacting with the CyberEdu MCP server:
Session & Authentication
"Check my CyberEdu session status"
"Set my CyberEdu session cookie to eyJpdiI6Ik..."
"Switch to tenant 'mycompany'"Challenges (Archive)
"List all web security challenges"
"Show me the easiest challenges from tenant unbreakable/rocsc"
"Show me hard difficulty forensics challenges"
"Get details for challenge abc123"
"Subscribe me to this challenge and start the service"
"Download challenge files to ./downloads/"
"Submit flag 'CTF{i-like-web-security-ctf-challenges}' for this challenge"Contests
"List available CTF contests"
"Show leaderboard for contest 'defcamp ctf quals 2025'"
"Get challenge abc123 from contest 'rocsc26-quals'"
"Start service for this contest challenge"
"Submit flag 'FLAG{solved}' for contest challenge"Workflow Example
1. "List easy web challenges from tenant rocsc"
2. "Subscribe to 'why-xor' and start the service"
3. "Download the challenge files"
4. [Solve...]
5. "Submit flag 'CTF{xor-is-not-safe}'"Architecture
Dynamic Tool Discovery
The server uses Python's `inspect` module to automatically discover all public methods from the CyberEduClient class. For each method:
1. Method Discovery: Scans the class for public methods (excluding private methods and helpers)
2. Schema Generation: Automatically generates JSON schema from method signatures and type hints
3. Tool Registration: Registers each method as an MCP tool with appropriate metadata
Tool Registry
The `ToolRegistry` class provides a flexible system for managing tools:
- Automatic Discovery: Discovers methods from classes using introspection
- Manual Registration: Allows manual registration of custom methods
- Category Organization: Automatically categorizes methods (auth, challenges, contests, services, etc.)
Extensibility
To add new functionality:
1. Add methods to CyberEduClient: Simply add new public methods to the CyberEduClient class
2. Automatic Exposure: The MCP server will automatically discover and expose the new methods
3. No MCP Code Changes: No changes needed to the MCP server code
For custom tools that don't map directly to client methods:
from cyberedu_mcp.tool_registry import ToolRegistry
registry = ToolRegistry()
def custom_tool(param1: str, param2: int) -> dict:
"""Custom tool description."""
return {"result": f"{param1}: {param2}"}
registry.register_method(
name="custom_tool",
method=custom_tool,
description="A custom tool",
category="custom"
)Error Handling
The server provides comprehensive error handling:
- HTTP Errors: Returns detailed HTTP error information including status codes and response bodies
- Validation Errors: Returns clear error messages for missing or invalid parameters
- Client Errors: Returns structured error information for all exceptions
Documentation
Additional documentation is available in the `docs/` folder:
- **docs/index.md** - Documentation overview and quick reference
- **docs/architecture.md** - Server design, tool discovery, and session management
- **docs/extending.md** - How to add custom tools and modify behavior
License
MIT
Frequently asked questions
What is cyberedu-mcp?
cyberedu-mcp is This is the Oficial Model Context Protocol (MCP) server for the CyberEdu CTF platform (cyber-edu.co / cyberedu.ro)
How do I install cyberedu-mcp?
Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.
Is cyberedu-mcp open source?
Yes — it is hosted on GitHub at https://github.com/CyberEDU-Cyber-Range/cyberedu-mcp and has 8 stars.
Related MCP tools
Cognee is the open-source AI memory platform for agents. Give your AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.
Automate browser based workflows with AI
Hindsight: Agent Memory That Learns
A privacy-first app that strips AI watermarks from content you own.
Agent framework and applications built upon Qwen>=3.0, featuring Function Calling, MCP, Code Interpreter, RAG, Chrome extension, etc.
The power of Claude Code / GeminiCLI / CodexCLI + [Gemini / OpenAI / OpenRouter / Azure / Grok / Ollama / Custom Model / All Of The Above] working as one.
Run your own MCP server? See who uses it and what to fix.
Measure it with TrackMCP