The paper describes the evolution and implementation of a self-executing security examination (SENSE) for agents executing in non-trusted domains. It outlines the shortcomings of some of the existing agent security schemas. To meet these shortcomings, the paper proposes the implementation of SENSE as a reliable method for detecting a malicious action attempted on an executing agent at run-time. The incorporation of SENSE into its normal processing operation will make the agent self-reliant with respect to its security function and allow it to confidently execute in an alien environment, without having to depend on an external support for security. To build a strong case for its adoption, the paper elaborates the ease of operation, simplicity and efficiency of the schema.