Turn a memory into a vivid scene without inventing anything. It asks the questions that pull out detail first, then drafts the scene in first person with two ways to reflect on it.