Validates learned AI patterns through systematic benchmarking, quality scoring, and regression testing.
The ecc-cmd-learn-eval skill provides a robust framework for validating candidate learnings derived from agent tasks. It allows users to define specific acceptance metrics—such as pass rates and cycle times—to run controlled trials that compare baseline performance against new patterns. By prioritizing evidence-based execution, this skill ensures that only patterns demonstrating measurable improvements are adopted, while regressions are flagged immediately for review, making it essential for maintaining high-quality autonomous workflows.
