MCP CLI Command Server
Provides safe, controlled access to network and system diagnostic tools including ping, nmap, dig, traceroute, curl, and whois for troubleshooting connectivity, scanning ports, and performing DNS lookups through whitelisted commands with security validation.
README
MCP CLI Command Server
A Model Context Protocol (MCP) server that provides safe, controlled access to common network and system CLI commands. This server enables AI assistants to execute whitelisted commands like nmap, ping, dig, and curl programmatically.
What is MCP?
The Model Context Protocol (MCP) is an open protocol that enables AI assistants like Claude to securely interact with local tools and services. This server implements MCP to provide CLI command execution capabilities.
Features
Network Diagnostic Tools (4 tools)
- cli_ping - Ping hosts to check connectivity and measure latency
- cli_traceroute - Trace network route to destination hosts
- cli_mtr - Combined ping and traceroute diagnostic (MTR)
- cli_netcat_test - Test port connectivity using netcat
Network Scanning Tools (2 tools)
- cli_nmap_ping_scan - NMAP host discovery on network ranges
- cli_nmap_port_scan - NMAP port scanning with service detection
DNS Tools (3 tools)
- cli_dig - DNS lookups with multiple record type support
- cli_host - DNS host command for quick lookups
- cli_whois - WHOIS lookups for domains and IP addresses
HTTP Tools (1 tool)
- cli_curl_get - HTTP GET requests with curl
Quick Start
Using Docker (Recommended)
# Build and start
docker-compose up -d
# View logs
docker logs -f cli-mcp-server
# Stop
docker-compose down
Configuration
Edit docker-compose.yml to configure:
environment:
- PORT=3017 # HTTP server port
Using with Claude Desktop
Add to your Claude Desktop MCP configuration (~/.claude/mcp.json):
{
"mcpServers": {
"cli": {
"type": "http",
"url": "http://localhost:3017/mcp",
"description": "CLI command execution for network and system diagnostics"
}
}
}
Using with MCP Aggregator
The MCP Aggregator automatically discovers this server when configured:
# In mcp-aggregator docker-compose.yml
environment:
- MCP_SERVER_CLI=http://host.docker.internal:3017
Tools will be available as cli_* (e.g., cli_ping, cli_nmap_port_scan)
Testing
# Health check
curl http://localhost:3017/health
# List available tools
curl http://localhost:3017/mcp/list_tools | jq
# Test ping command
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_ping",
"arguments": {"host": "8.8.8.8", "count": 4}
}'
# Test nmap ping scan
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_nmap_ping_scan",
"arguments": {"target": "192.168.1.0/24"}
}'
# Test nmap port scan
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_nmap_port_scan",
"arguments": {
"host": "scanme.nmap.org",
"ports": "22,80,443",
"service_version": true
}
}'
# Test DNS lookup with dig
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_dig",
"arguments": {
"domain": "example.com",
"record_type": "A",
"short": true
}
}'
# Test HTTP request with curl
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_curl_get",
"arguments": {
"url": "https://api.github.com",
"headers_only": true
}
}'
# Test traceroute
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_traceroute",
"arguments": {"host": "google.com", "max_hops": 15}
}'
# Test WHOIS lookup
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_whois",
"arguments": {"target": "github.com"}
}'
# Test host lookup
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_host",
"arguments": {
"hostname": "cloudflare.com",
"record_type": "MX"
}
}'
# Test netcat port connectivity
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_netcat_test",
"arguments": {"host": "github.com", "port": 443, "timeout": 5}
}'
# Test MTR network diagnostic
curl -X POST http://localhost:3017/mcp/call_tool \
-H "Content-Type: application/json" \
-d '{
"name": "cli_mtr",
"arguments": {"host": "8.8.8.8", "cycles": 10}
}'
API Endpoints
GET /health- Health checkPOST /mcp- MCP HTTP transport (JSON-RPC 2.0) - recommendedGET /mcp/list_tools- List available tools (REST)POST /mcp/call_tool- Execute a tool (REST)
Security
Whitelist-Only Approach
- No Arbitrary Commands: Only predefined, whitelisted commands are allowed
- Argument Validation: All arguments validated with regex patterns
- No Shell Metacharacters: Blocks
;,&&,||,|,$, backticks, etc. - Timeout Enforcement: All commands have maximum timeout limits
Input Validation
- IP Addresses: Validated with regex (0-255 per octet)
- Hostnames: RFC-compliant hostname validation
- Ports: Range check (1-65535)
- URLs: Protocol and format validation
- CIDR Notation: IP and prefix validation
Resource Limits
- CPU: Limited to 0.5 CPU cores
- Memory: Limited to 512MB
- Timeouts: Command-specific (5-600 seconds)
- Non-Root User: Runs as unprivileged user (uid 1000)
Docker Security
- No New Privileges:
security_opt: no-new-privileges - Minimal Capabilities: Only NET_RAW and NET_ADMIN for network tools
- Network Isolation: Runs on isolated Docker network
- Health Checks: Automated container health monitoring
Audit Logging
- All commands logged with timestamps
- Includes tool name, arguments, and execution time
- Exit codes and errors logged for troubleshooting
Requirements
- Docker & Docker Compose
- Network connectivity for external scans
- Sufficient permissions for network tools (NET_RAW, NET_ADMIN capabilities)
Port
3017 - HTTP server for MCP protocol
Architecture
- Language: Python 3.11
- Framework: Flask + MCP Python SDK
- Transport: HTTP with JSON-RPC 2.0
- CLI Tools: nmap, ping, curl, dig, traceroute, whois, host, netcat, mtr
Installed CLI Tools
The Docker container includes these CLI tools:
- nmap - Network mapping and port scanning
- iputils-ping - ICMP ping utility
- curl - HTTP client
- dnsutils (dig, nslookup, host) - DNS utilities
- traceroute - Network path tracing
- whois - Domain/IP WHOIS lookup
- netcat-traditional - TCP/UDP connectivity testing
- mtr-tiny - Combined ping and traceroute
Troubleshooting
Container won't start
# Check if port 3017 is in use
lsof -i :3017
# View container logs
docker logs cli-mcp-server
# Check Docker network
docker network inspect mcp-network
Permission errors
# Ensure container has network capabilities
docker inspect cli-mcp-server | jq '.[0].HostConfig.CapAdd'
# Should show: ["NET_RAW", "NET_ADMIN"]
Commands fail with validation errors
- Ensure hostnames are valid (RFC-compliant)
- IP addresses must be in dotted-quad format (192.168.1.1)
- Ports must be 1-65535
- No shell metacharacters allowed in arguments
Timeout errors
- Increase timeout parameter in tool arguments
- Check network connectivity from container
- Some commands have maximum timeout limits (security)
Common Use Cases
Network Troubleshooting
- Check if hosts are reachable with ping
- Trace network routes to diagnose connectivity issues
- Test port connectivity with netcat
Security Scanning
- Discover active hosts on network with nmap
- Scan ports for open services
- Identify service versions
DNS Diagnostics
- Look up DNS records with dig
- Verify DNS configuration
- Check domain ownership with WHOIS
API Testing
- Test HTTP endpoints with curl
- Check response headers
- Verify SSL/TLS connectivity
Development
# Rebuild after code changes
docker-compose up -d --build
# View detailed logs
docker logs cli-mcp-server --tail=100 -f
# Test locally without Docker
PORT=3017 python src/http_server.py
Integration with Mattermost Bot
When integrated with the Mattermost Azure OpenAI bot via the MCP aggregator:
- Bot discovers tools automatically (5-minute refresh)
- Tools appear as
cli_*functions - Use natural language: "Can you ping 8.8.8.8?"
- Bot executes command and returns results in chat
Example conversation:
User: Can you check if github.com is reachable?
Bot: *calls cli_ping with host=github.com*
Bot: Yes, github.com is reachable! Average latency is 15ms.
User: Scan ports 22, 80, and 443 on scanme.nmap.org
Bot: *calls cli_nmap_port_scan*
Bot: Port scan results:
- Port 22 (SSH): Open - OpenSSH 7.9
- Port 80 (HTTP): Open - Apache 2.4
- Port 443 (HTTPS): Closed
License
MIT License - See LICENSE file for details
Contributing
Contributions welcome! Please:
- Fork the repository
- Create a feature branch
- Make your changes
- Submit a pull request
Support
For issues and questions:
- GitHub Issues: https://github.com/jbmurphy/mcp-cli/issues
- MCP Documentation: https://modelcontextprotocol.io
Security Notice
⚠️ Important: This server executes network and system commands. While security measures are in place (whitelisting, validation, timeouts), use caution when exposing to untrusted networks or users.
Best Practices:
- Run on isolated Docker network
- Monitor command execution logs
- Review firewall rules for outbound scanning
- Limit access to trusted users/systems
- Keep container updated with security patches
Recommended Servers
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.
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.
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.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
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.
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.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
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.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.