Stable Inverse Reinforcement Learning: Policies from Control Lyapunov Landscapes