This work concerns modelling evolving prospective agent systems. Inasmuch a prospective agent looks ahead some steps into the future, it is confronted with the problem of several possible courses of evolution, and therefore needs to prefer amongst them to decide the best to follow from its present state. First it needs a priori preferences for generating likely courses of evolution. Subsequently, a main contribution of this paper, based on historical information and on a mixture of quantitative and qualitative a posteriori evaluation of its possible evolutions, we equip our agent with so-called evolution-level preferences mechanism, involving three distinct types of commitment. One other main contribution, to enable a prospective agent to evolve, we provide for modelling its evolving knowledge base, including environment and course of evolution triggering of active goals, context-sensitive preferences and integrity constraints. We exhibit examples to illustrate proposed concepts.