Consider the following setting: There is some simple unknown graph G=(V,E). One has access to G only via queries known as *cut queries*, in which one can specify some subset S of V and obtain back the number of edges that cross the cut defined by S in G. A natural question to ask is what is the cut-query complexity of finding the minimum cut in the input graph. This is a special case of the well-known problem of sub-modular function minimization. Throughout the talk, we'll use this problem as a vehicle for exemplifying the usefulness of a novel simple combinatorial procedure, styled Star-Contraction. Namely, this tool allows us to obtain a randomized cut query protocol for the problem with O(n) cut queries. I'll then discuss additional applications of Star-Contraction and open problems.