This paper proposes a new equivalence theory for MAGs learned from a combination of observational and (soft) interventional data, and propose a modification of the FCI algorithm that learns the resulting classes. The reviewers found the paper well-structured and a good contribution to the causal discovery literature.