- Teams often require their members to coordinate with respect to specific aspects of their joint task. These coordination relations can often fail as a re-sult of intermittent failures of sensor readings, com-munication failures, etc. Detection of such failures, based upon agents' observations of each other, is of prime importance. Though different solutions have been presented, none has presented a compre-hensive, formal approach to solving this problem. Our research has produced a compact matrix-based representation of pre-defined agent coordination re-lations, that is used with a matrix-based represen-tation of a plan-recognition process. The result is a novel solution that is both generic and efficient for large-scale teams. Additionally, it facilitates easy design of coordination requirements, modu-larity and reuse of existing systems.