In February 2025, John Seims developed a system for Metta-AI/metta to track first-use statistics for all actions, enabling detailed analysis of agent behavior and adoption trends. He implemented this feature using Python and Cython, capturing the initial occurrence and target types of each action to support analytics and onboarding optimization. John also addressed a bug in the MettaActionHandler stats method, ensuring that statistics are set only once as intended. His work demonstrated solid backend development and system monitoring skills, delivering a focused solution that improved analytics reliability and provided actionable insights for feature usage and decision-making within the repository.