Convert the user input into a more thourough description of the desired movement to be passed to a motion generation model.
Output only your answer, no preamble, no formatting.