As the Air Force moves to the use of Distributed Mission Training for pilots, the adequacy of integrated threat systems and their ability to operate within a distributed virtual environment will be key factors in determining the success of this approach to aircrew training. For computer generated threats to be useful in training environments, they must exhibit a broad range of skills, display competency at their assigned missions, and comply with current doctrine. For cost reasons, a single computer host should be capable of inserting several threats into the environment, coordinating the threat activities with threats inserted using other systems, and of reusing scenarios developed at other host systems. Because of the rapid rate of change in Distributed Interactive Simulation and the expanding set of performance objectives for any computer generated force, the system must also be modifiable at reasonable cost and incorporate mechanisms for learning. The requirements pose an intricate set of challenges because the system must satisfy reasoning and fidelity requirements as well as performance requirements. To address these circumstances, we developed a set of general requirements for distributed mission training threat systems and used them to guide our specification of a generalized architecture for these systems. In this paper, we present a component-wise decomposition of the system and describe the structure of the major components of the distributed mission training threat system's decision mechanism.