research-artifact
Open protocol for testing whether ML and AI-safety metrics add information beyond cheap baselines.
Python library for evidence sufficiency scoring in governance assessments under delayed ground truth, drift, and decision-readiness constraints.
