cognizant-ai-lab/neuro-san-studio ? reverse-engineered prompt

Build me a Python playground for designing and testing intelligent multi agent systems.

I want a project where I can define agent networks in simple config files, run ready made examples, and try out different workflows without having to write everything from scratch. It should feel like a hands on studio for experimenting with multiple AI agents that can delegate tasks to each other, use custom tools, and work with different model providers. Please include a few realistic demo scenarios, like a network designer example and some business style use cases, so I can see how it works right away.

Make it easy to run locally, easy to extend with my own tools or agents, and include clear logging and tracing so I can understand what each agent is doing. If you need to look up current docs online to fill in any gaps, go for it.