Space manipulator robots are employed in a wide range of tasks. The diversity and complexity of the required operations often exceed human operators’ capabilities in manual control, necessitating increased autonomy. This article addresses the enhancement of autonomy in space manipulator systems through the theory and methodology of “skills.” The primary focus is on robotic assembly, which entails a sequence of complex actions—grasping objects, transporting them in space, and integrating them into a target structure. Within the “skills” framework, these actions are formalized as elementary, structured primitives encompassing coordination, control, and task execution. As a representative case, we consider the simplest assembly task: inserting one plastic cup into another. The required skill set includes automatic estimation of the objects’ linear and angular poses, trajectory planning, execution of planned motions, object grasping, and precise insertion. We present the results of experimental validation on a semi-physical testbed comprising a Kawasaki FS-03N industrial robot, a Kinect-based vision system, a six-axis force–torque sensor, and a scalable computing infrastructure built on the Robot Operating System (ROS).