Fix mujoco agent bug, that incorrectly set the end effector points of the first time step of every sample for all but one of the conditions. (cleanest way to do this would be to omit joint angles and joint velocities from state, and only include ee point data types)
Add end-effector points to point mass example, for visualization purposes