In the last few years, Wireless Sensor Networks (WSNs) have emerged as a disrupting technology for myriad military and civilian applications. They demand an accurate location of the event detected and is done by using a mobile beacon node to provide accurate location and assume a benign environment. However, in a hostile military environment, such a node can be easily tampered by an adversary. In this paper, we propose a distributed cluster based anomaly detection scheme by assigning few randomly chosen cluster heads a critical task of estimating the reliability of the mobile beacon node. As localization of remaining nodes is cautiously performed only after verifying the authenticity of the mobile beacon node, a considerable overhead is saved in the incorrect localization of the entire network. We perform extensive simulation for different attacks and observe our scheme to have a high detection rate of 99% and a very less false positive rate of 20%.