Stick with the task like a bulldog
This will sound trivial, but I see it all too often: when the participant goes off-topic mid-task the UXR forgets to steer them back to the original task, but they continue with the next task.
Participants will naturally go off on tangents. That is ok and often insightful. But when participants are done, we need to get them back to where they were. Otherwise, we might end up in the uncomfortable situation that is: gathered a bunch of insights, non of which answered our main question - whether users can complete the tasks or not.
Sometimes it feels uncomfortable steering participants back to where they were. I think this comes from the fact that in natural conversations people use tangents to intentionally avoid sensitive topics. And then a person with empathy (e.g. a Researcher) doesn't want to push that topic again.
In usability testing however, there is nothing sensitive. So do feel comfortable taking the participant back to where they were.
I would say: "Great. So, before we talked about how this fits your workflow, you were about to create a reminder in this calendar... First, you picked the week, then you clicked 'create'... Please continue from there."
This prompt has two parts. 1) reminds the participant of the task, 2) reminds them where they left off, so they continue from there and not start all over again.