The use of IP networks for telephony applications (VoIP) is becoming increasingly popular, mainly due to its advantages in terms of communication and multimedia services. This fact may also shift several problems from the Internet context, such as spam, which - in the VoIP case - has been identified as SPIT. In this paper, we propose an abstract model for describing SPIT attack strategies by incorporating the underlying threats and vulnerabilities of the VoIP technology regarding SPIT phenomenon. Our model is mainly focused on the signaling part of VoIP sessions (i.e. the SIP protocol), and it is based on the representation of attacks through attack graphs and attack trees. We also demonstrate how this model could be used for the development of a set of reusable attack scenarios (patterns), with an eye towards the development of a SPIT Detection System.