Inputs are identified in the task by the syntax {{inputName}} for text and {{image:inputName}} for images