Actually successful enforcement of many human-in-the-loop doctrinaire tasks directly count on the operator¿s skill or a programmer¿s learning of the presumed environment in which the task will be complete. Computer mediation of human inputs can augment this process to: 1- permit easy and rapid incorporation of local sensory information to augment performance, 2-provide variable performance (precision- and power-) assist for output motions/forces, and 3- hierarchical distribution of control and graceful degradation. Such intercede control has massive potential to both reduce factor error and permit integration of greater autonomy into human/robot interaction.