We present results on the Swarm Counter-Swarm (SCS) task, in which a swarm of Unmanned Aerial Vehicles (UAVs) is deployed to counter a swarm of malicious UAVs attempting to reach a protected area of interest. To this end, we introduce \texttt{Defend}, a novel open Multi-Objective Multi-Agent Reinforcement Learning (MOMARL) environment on which a set of joint policies is learned to control the defending UAV swarm. The multi-objective formulation of the reward function enables user-specified preferences over competing objectives, allowing fine-grained behavioural control of the swarm at deployment time. Our preliminary results demonstrate that the chosen joint policy could either effectively mitigate the entire enemy swarm, or reduce sub-optimal behaviours, such as collateral damage and energy spent.