Existing Mobile Agent Systems (MASs) suffer from security problems that must be solved, if mobile code is to be used in the development of mission-critical, real-world applications. In this paper we propose a framework that provides automated support for verification and analysis of MASs and allows for identifying security issues before the MASs are placed into action. The proposed approach is centered around an abstract reference model and combines static and dynamic security analysis techniques explicitly tailored to mobile agent systems. Our preliminary work shows that, in the cases we studied, appropriate analysis techniques facilitate the identification of security vulnerabilities in MASs.