In January 2026, this developer enhanced the microsoft/agent-lightning repository by implementing dynamic prompt template configuration for the APO algorithm. Using Python and applying skills in algorithm design and software engineering, they enabled prompt templates to be configured through constructor arguments and allowed alternate templates to be loaded based on user settings. This approach increased the flexibility of the prompting strategy, supporting faster experimentation and per-customer customization. The work also introduced new parameters to support future features like gradient and apply edit prompt files. No bugs were reported, indicating a focused and well-executed feature delivery within a short timeframe.