Many present and past gamma‐ray burst (GRB) detectors try to be not only a “scout”, discovering new GRBs, but also the “cavalry”, simultaneously optimizing on‐board science return. Recently, however, most GRB science return has moved out from the gamma‐ray energy bands where discovery usually occurs. Therefore a future gamma‐ray instrument that isonlya scout might best optimize future GRB science. Such a scout would specialize solely in the initial discovery of GRBs, determining only those properties that would allow an unambiguous handoff to waiting cavalry instruments. Preliminary general principles of scout design and cadence are discussed. Scouts could implement observing algorithms optimized for finding GRBs with specific attributes of duration, location, or energy. Scout sky‐scanning algorithms utilizing a return cadence near to desired durations of short GRBs are suggested as a method of discovering GRBs in the unexplored short duration part of the GRB duration distribution. © 2004 American Institute of Physics