This paper is concerned with the optimal control of linear dynamical systems in the presence of a set of adversary constraints. One of the novelties is the type of constraints introduced in the receding horizon optimization problem. These constraints can be considered as adversary by their non convex characteristics which make the convergence of the systems' dynamics towards the "natural" equilibrium position an impossible task. In this case, the default equilibrium point has to be replaced by a set of equilibrium points or even to accept the existence of limit cycles. The present paper proposes a dual-mode control strategy which builds on an optimization based controller and a fixed constrained control law whenever the adversary constraints are activated. Furthermore, the method which exhibits effective performance builds on invariance concepts and proves to be related to the classical eigenstructure assignment problems. In order to illustrate the benefits of the proposed method, typical applications involving the control of Multi-Agent Systems are considered.