marketagents-ai/MarketAgents — reverse-engineered prompt

Build me a Python project for running simple market simulations with AI agents. I want to create economic agents that can make decisions, talk to large language models, and be orchestrated together in a repeatable simulation. The project should support parallel model calls and prompt caching so experiments don’t waste time or money.

Please include a clean package structure, a basic example simulation, and a simple command I can run to start it. Add an environment file example for API keys and settings, plus a configuration file where I can choose models and simulation options. I also want clear setup instructions so someone can clone it, install it locally, fill in their keys, and run the demo without guessing.

Keep the first version practical and easy to extend. It doesn’t need a fancy UI, just a working framework, docs, and tests for the core pieces. Look up current docs online if you need to.