In this talk I will discuss the application of Reo to multi-agent systems. Reo supports a specific notion of compositionality for multi-agent systems which enables the composition and coordination of both individual agents as well as multi-agent systems. Accordingly, a multi-agent system consists of a set of individual and/or multi-agent systems whose collective behavior is coordinated by a Reo expression. This coordination language should be modified to specify and implement the organization of multi-agent systems and their dynamic reconfiguration during system run.