The increasing integration of active chassis actuators in modern vehicles is transforming motion control development. As vehicles become more software defined , control systems must deliver high performance while remaining adaptable across vehicle derivatives, driving modes, and actuator configurations. A key challenge is that changes in actuator specifications often require controller redesign, retuning, or retraining. This paper presents a hardware-aware deep reinforcement learning integrated chassis controller for coordinated rear-wheel torque vectoring (TV) and rear-wheel steering (RWS). By incorporating RWS angle and rate limits directly into the policy inputs, a single trained controller can adapt its control allocation across RWS hardware constraints without retraining. The controller is trained in high-fidelity simulation using open-loop manoeuvres and evaluated with different RWS actuator limits, vehicle loading conditions, and closed-loop driving with a driver model. Results show that the controller redistributes control effort between TV and RWS as steering authority changes while maintaining stable vehicle behaviour. Beyond vehicle control, the same trained controller is used as a design-space exploration tool to assess actuator hardware trade-offs by linking RWS rate capability to vehicle-level performance and actuator usage. In the investigated design space, the results identify diminishing returns beyond an RWS rate limit of approximately 7°/s, with performance saturation above 12°/s. These results demonstrate how the controller can support actuator specification through integrated vehicle-level assessment without repeated controller retraining.