Systems of autonomous agents providing automated services over the Web are fast becoming a reality. Often these agent systems are constructed using procedural architectures that provide a framework for connecting agent components that perform speci.c tasks. The agent designer codes the tasks necessary to perform a service and uses the framework to connect the tasks into an integrated agent structure. This bottom up approach does not provide an easy mechanism for con.rming global properties of constructed agent systems. In this paper we propose a declarative methodology based on logic programming for modeling such procedurally constructed agents and specifying their global properties as temporal logic formulas. This methodology allows us to bring to bear a body of work for using logic programming based model checking to verify certain global properties of procedurally constructed Multi-Agent Systems.