tauricResearch/TradingAgents — reverse-engineered prompt

Build me a Python research tool for stock trading that acts like a small trading firm made of AI agents. I want a command line app where I can give it a ticker and date range, and it should have separate agents for fundamentals, sentiment, news, and technical analysis, then let bullish and bearish researchers debate, then have a trader, risk manager, and portfolio manager decide whether to buy, sell, or hold.

Keep it focused on research and simulated trading, not real brokerage execution or financial advice. It should produce clear reports showing what each agent thought, the final decision, confidence or rating, and a decision log I can review later. If possible, support backtesting and a simulated exchange flow so approved trades can be tracked.

Make it easy to run locally with a simple CLI, environment variables for model and data provider keys, and Docker as an option. Please wire it so different major LLM providers can be swapped in. Look up current docs online if you need to.