LangGraph & Docker: Secure Code Execution for Agents?

[AndreiPopg]

I'm exploring LangGraph for a multi-agent workflow project, and I'm really impressed with its flexibility and async capabilities. However, one concern I have is around secure code execution.

In AutoGen, code-executing agents can run their actions within Docker containers, providing a layer of isolation and security. Is it currently possible to achieve something similar with LangGraph?

My requirements are:

• Containerize agents: Run each code-executing agent within its own Docker environment.
• Customizable environments: Specify the Docker image and configuration for each agent's container.

Are there any workarounds or recommended approaches?

Use Case:
I'm working on a project where multiple agents need to execute code securely and in isolation to prevent any potential interference or security issues.

Any insights, suggestions, or examples of how you've addressed similar challenges would be greatly appreciated!
Thanks in advance.