This paper introduces Carnegie Mellon campus-wide CarnegiePLUG test-bed. This test-bed in a hardware-in-the-loop simulator that enables large-scale sensing, computation and actuation over a network of heterogeneous energy producers and consumers (prosuming) assets. CarnegiePLUG is a multi-agent framework that facilitates plug-and-play integration of a wide range of virtual and physical sensors and controllers through a multi-layer architecture. Virtual functionalities are accessed in software environments that interface with communication layer of our test-bed. This paper discusses CarnegiePLUG layers, provides implementation details and showcases potential applications of this test-bed through preliminary results.