The design of the manipulation skill library adheres to the definition of manipulation tasks that cause transitions in surface contact, as established in earlier research  [9]. The necessary task sets are establishedKuhn-Tucker theory using Kuhn-Tucker theory [18] for manipulation tasks. However, the actual necessity and sufficiency of these skill libraries depend on the specific domain in which the robot tasks are performed. The theories provide an upper boundary for all the necessary tasks; however, in practice, only a subset of these tasks frequently appear in each domain, while the remainder rarely occur. Consequently, it is impractical to prepare all the skills in the libraries. Instead, in the LfO2.0 project, we target the domain of household service robots, designing a comprehensive set of non-overlapping tasks and their corresponding skill agents.