Deployed Wireless Sensor Networks (WSN) experience conditions in the the physical world that are difficult to predict and test for during development. Even simulation and testbed deployment often fail to reveal every bug. The iterative processes of programing a deployed WSN, detecting failures, collecting information to diagnose the failures, and then re-programming the network, can be slow and laborious. It is also hampered by the parsimonious energy and resource constraints typical to WSNs. Our work focuses on the development of new software and hardware tools that provide a high level of visibility into deployed WSNs, while working within typical energy and resource constraints.