- Explore MCP Servers
- SELF2025
Self2025
What is Self2025
SELF2025 is a presentation focused on utilizing Large Language Models (LLMs) with Model Context Protocol (MCP) tools for enhanced system administration and diagnostics.
Use cases
Use cases include monitoring and repairing Docker containers, conducting network diagnostics, and retrieving system information using LLMs to facilitate automated management and troubleshooting.
How to use
To utilize the tools, set up the Docker containers using the provided docker-compose.yml
and access Python scripts to perform tasks such as container management and network diagnostics. The included example scripts illustrate LLMs interacting with the MCP servers for system administration tasks.
Key features
Key features include Docker container management functions (listing containers, checking status, retrieving logs), network tools for domain and IP lookups, and integration with LLMs for automated sysadmin functions. The setup is flexible and allows access to a variety of MCP resources.
Where to use
These tools can be used in environments that employ MCP technology for server management, network diagnostics, and system administration. Useful for developers and system admins looking to leverage LLMs for efficient infrastructure management.
Overview
What is Self2025
SELF2025 is a presentation focused on utilizing Large Language Models (LLMs) with Model Context Protocol (MCP) tools for enhanced system administration and diagnostics.
Use cases
Use cases include monitoring and repairing Docker containers, conducting network diagnostics, and retrieving system information using LLMs to facilitate automated management and troubleshooting.
How to use
To utilize the tools, set up the Docker containers using the provided docker-compose.yml
and access Python scripts to perform tasks such as container management and network diagnostics. The included example scripts illustrate LLMs interacting with the MCP servers for system administration tasks.
Key features
Key features include Docker container management functions (listing containers, checking status, retrieving logs), network tools for domain and IP lookups, and integration with LLMs for automated sysadmin functions. The setup is flexible and allows access to a variety of MCP resources.
Where to use
These tools can be used in environments that employ MCP technology for server management, network diagnostics, and system administration. Useful for developers and system admins looking to leverage LLMs for efficient infrastructure management.
Content
SELF2025
Example scripts and resources for my presentation on LLMs using MCP tools at SELF2025
File Descriptions
Python Files
If you use these python scripts yourself, make sure to replace the example server urls for certain extra lookup services that are used.
system_tools.py
A FastMCP server that provides Docker container management tools. Functions include:
- List containers
- Check if containers are running
- Retrieve container logs
- Get container stats
- Restart containers
- Access MCP Resources:
- Docker compose files
- Text files
- System IP addresses
- Date/time information
- System information
network_tools.py
A FastMCP server providing network-related tools. Functions include:
- Domain lookups
- IP lookups
- Ping tests
- HTTP GET testing
- DNS checking
- IP address listing
- Service monitoring
- Access MCP Resource
- System IP Addresses
smolagents_sysadmin_examples.py
Demonstrates how to use SmolAgents with MCPAdapt to interact with MCP servers. The example shows how to check and fix a service running in a Docker container, using LLMs to perform system administration and diagnostic tasks.
Example Run: https://asciinema.org/a/8YdOS2KoQXTGVNDzeVtWIrvyw
YAML Files
mcp-proxy-tools.yml
Docker Compose configuration for running the MCP server tools as containers:
mcp-proxy-network
: Runs the network_tools.py script, exposing port 8097mcp-proxy-system
: Runs the system_tools.py script, exposing port 8098, with access to Docker socket
Dockerfile
mcp-proxy_tools.Dockerfile
Builds the container image used by the MCP proxy services:
- Based on the mcp-proxy base image
- Installs necessary Python packages (psutil, mcp[cli], requests, docker)
- Sets up the environment for running MCP services
Example Usage
The repository includes example scripts showing how an LLM can be used to:
- Monitor and repair Docker containers (e.g., fixing the “homebox” service)
- Perform network diagnostics
- Retrieve system information
Additional Resources
Personal Informational MCP server
Code: https://github.com/Red5d/me-mcp
Live: https://mcp.red5d.dev/sse
MCP-capable client software
Where to find MCP servers to run/use:
- MCP Github: https://github.com/modelcontextprotocol/servers
- Docker Hub: https://hub.docker.com/catalogs/mcp
- Glama: https://glama.ai/mcp/servers
- Smithery: https://smithery.ai/
- Pipedream: https://mcp.pipedream.com/
- GitHub Awesome List: https://github.com/punkpeye/awesome-mcp-servers