We explore the concept of life cycles of knowledge-based components in knowledge-engineering environments, and particularly in distributed, agent-based settings. We provide a formal account of a knowledge life cycle based on abstract model theory, and show its relationship to channel theory---a mathematical model of semantic information flow. This leads to a general calculus for life-cycle description, which, as we demonstrate, can be applied to large-scale problems. Since practical applications require tool support, we describe a tool which uses our life-cycle calculus to assist the user in the construction of application-specific knowledge-transformation procedures; these procedures can then be invoked and executed in asynchronous, distributed environments.