How does Scikit-Learn Pipeline prevent data leakage?

Pipelines ensure transformations are applied only on training folds during CV.

Each fold applies fit only on training subset, preventing test data from influencing transformations.