Cliff faces, crevasses, and impact craters can be used to unlock the geological history of rocky and ice bodies. Unfortunately, these areas generally have extreme terrain that makes them inaccessible to most wheeled rovers. Instead, we present an autonomous climbing system modeled after a "rope team" of alpine climbers. Near-realtime perception and path-planning algorithms are developed and demonstrated for this cooperative team of climbing robots robots. An end-to-end pipeline reconstructs the environment from noisy data and selects hopping points that form a path to the objective, minimizing risk. We show how the algorithm would successfully navigate on the asteroid Itokawa.