ml-governance
Lineage layer for ML teams: every published model traces back to the exact commit, dataset and run that produced it, verified by content hash. A thin CLI over git, DVC and MLflow; no pipeline DSL, train however you want.
