A post describing EvolveTrade says it treats a trading agent's system prompt as its policy. After each trading interval, a separate Policy Agent reads decision traces and realized returns, then rewrites the prompt without changing the underlying model. The post reports that across several market regimes and two backbone models, the evolved agent beat fixed-prompt baselines on Sharpe ratio and cumulative return in most settings. It also says the rewritten prompts led to more code-based analysis and signals suited to the current market regime.