LLM Agent Environment

LLM Agent Environment: Docker setup for AI agent development.

Docker Hub Repository

Prerequisites

• Docker installed on your system
• Basic knowledge of Docker and command-line operations
• A GitHub SSH key for the agents to use

Setup

1. Clone this repository:

   git clone <repository-url>
   cd <repository-directory>
   

2. Provide a GitHub SSH key:

   • Create a directory named ssh_key in the project root.
   • Copy your GitHub SSH private key to ssh_key/id_rsa.
   • Copy your GitHub SSH public key to ssh_key/id_rsa.pub.
   • Ensure the permissions are correct:

     chmod 600 ssh_key/id_rsa
     chmod 644 ssh_key/id_rsa.pub
     

3. Create a shared user folder:

   mkdir shared_user
   

4. Build the Docker image and start the container:

   ./manage.sh rebuild
   

Managing the Environment

Use the manage.sh script to easily manage the LLM Agent Environment:

• To rebuild and run the Docker container:

  ./manage.sh rebuild
  

• To run setup and E2E tests:

  ./manage.sh setup-and-test
  

• To run only the E2E tests:

  ./manage.sh test
  

• To create a new agent:

  ./manage.sh add-agent <agent_name>
  

• To delete an existing agent:

  ./manage.sh delete-agent <agent_name>
  

• To list all agents:

  ./manage.sh list-agents
  

• To execute a command in the container:

  ./manage.sh exec <command>
  

• To view container logs:

  ./manage.sh logs
  

• To SSH into the container as an agent:

  ./manage.sh ssh <agent_name>
  

• To stop the container:

  ./manage.sh stop
  

• To start the container:

  ./manage.sh start
  

• To restart the container:

  ./manage.sh restart
  

• To check the container status:

  ./manage.sh status
  

• For help and available options:

  ./manage.sh help
  

Agent Environment

Each agent's environment includes:

• Sudo access (passwordless)
• Git configuration
• Node.js and npm (via nvm)
• Python virtual environment
• Access to shared folders (/shared_agents and /shared_user)
• Next.js development setup
• Access to the provided GitHub SSH key

Shared Resources

• /shared_agents: A directory accessible by all agents (internal to the container)
• /shared_user: A directory mounted from the host, accessible by all agents and the host

Ports

• 6668: WebSocket server
• 3010: Mapped to container's port 3000 (for Next.js applications)
• 2222: SSH access

WebSocket Server

The environment includes a WebSocket server (irc_websocket_server.py) running on port 6668. This server can be used for real-time communication between agents or external services.

Troubleshooting

If you encounter issues:

1. Ensure the container is running:

   ./manage.sh status
   

2. Check container logs:

   ./manage.sh logs
   

3. If you can't SSH, verify the SSH service is running inside the container:

   ./manage.sh exec service ssh status
   

4. If agents can't access GitHub, ensure the SSH key was correctly mounted:

   ./manage.sh exec ls -l /root/.ssh
   

5. If you can't access files in the shared user folder, check the mounting:

   ./manage.sh exec ls -l /shared_user
   

Customization

You can modify the create_agent.sh script in the docker directory to add or change the setup for each agent. After making changes, rebuild the Docker image using:

./manage.sh rebuild

Pushing to Docker Registry

To push the LLM Agent Environment image to Docker Hub:

1. Ensure you have a Docker Hub account and you're logged in:

   docker login
   

2. Use the manage script to push the image:

   ./manage.sh push <your-dockerhub-username>
   

Replace <your-dockerhub-username> with your actual Docker Hub username.

This will tag the image and push it to your Docker Hub repository.

Security Note

This setup is intended for development and testing purposes. For production use, implement proper security measures, including strong passwords and key-based authentication.

License

This project is licensed under the MIT License with No Warranty. See the LICENSE file for details.

Note: This software is provided as-is, without any warranty or support. Use at your own risk.

Copyright (c) 2024 c0rtexR, soft-flow