In order to develop and use shared libraries of problem-solving methods, it is of paramount importance to provide adequate descriptions of their capabilities and competence. Methods must be indexed and organized based on their capabilities so that they can be retrieved when their capability is adequate for the task at hand. This paper describes the approach taken in EXPECT for representing method capabilities and argues that it has important features that should be used for describing methods in shared libraries.EXPECT's capability representation is tightly coupled with the domain ontologies in the knowledge base, can express task-related parameters explicitly, and is based on case grammars. This representation allows the system to reason about the capability descriptions through class subsumption and reformulation. The benefits of this approach include self-organizing method libraries, reuse, and support for explanation. The representation has already been used extensively within EXPECT to express a wide range of method capabilities, ranging from abstract to specific, small to large, and domain-dependent to general-purpose methods.The paper also discusses some of the additional features that we anticipate will be useful to structure shared method libraries.