Modelling MAS-Specific Security Features

In this paper, we pursue a modelling approach to address security requirements for multi-agent systems (MAS). This will allow developers to account for both the system and agent-specific security requirements of a MAS during the requirements phase and throughout the whole Software Development Lifecycle of the system. We focus on autonomy, mobility and cooperation of individual agents and how these create additional security vulnerabilities to the system. In proposing a set of generic modelling primitives for these engendered requirements in the analysis of the MAS, we extend our recently proposed MAS metamodel.