Formal methods are often used to prove properties of network protocols, including required security properties. However for a protocol modeler the techniques available for security analysis often require expert knowledge of the technique. Also the tight coupling of protocol model and security attacks limit re-use of models. With colored petri nets as the selected formal method, this paper proposes a methodology to support a modeler in performing security analysis of a protocol. The methodology enhances the re-usability, extendibility and readability of protocol and attack models, with the aim of simplifying the tasks of the modeler.