The diffusion of interconnected computational entities is rapidly increasing, offering new opportunities to programmers and system developers thinking in terms of distributed and concurrent systems.
But along with them, also complexity is increasing, introducing new problems that can't be solved by traditional computer science and software engineering.
This is leading to a change in the design and development of software systems, because of the need for higher abstractions and technologies capable of dealing with such complex systems. 

MASs (Multi Agent Systems) and related metaphors and technologies are the most discussed abstractions for interpreting and engineering the increasing complexity in modern software systems, and nowadays they are considered the next step after the object-oriented paradigm. 

However, since MAS abstractions are not yet universally accepted and formalised, many tools to support MAS engineering (such as methodologies and frameworks) are still missing or in an early stage of development.
The use of good tools is a mandatory requirement to deploy quality software systems, therefore their introduction in MASs is needed for building such systems efficiently and systematically.
In this work we focus on agent interactions: firstly, we choose a specific meta-model to interpret MASs (the A&A meta-model), and then we present a game-based programming methodology for building mediated, multi-agent interactions in MASs.