We are a team whose main work is technical but not software development, and we have been using Scrum as our main framework for a while (although it's likely that what we're doing is closer to ScrumBut in various ways).
With that said, we had a question come up in Sprint Planning today that I didn't know the answer to - sometimes, when planning out work, there is a temptation to just do some of the tasks without including them in the Sprint Backlog. Either they're small enough that it's quicker to just get them done rather than creating a Jira task, or the task is part of the planning process (e.g. doing a quick bit of research to help understand something that needs estimation) but doing it leads naturally into larger work.
Is there any guidance on recognising when something needs to be put on the board rather than just knocked over on the side? The best I could suggest was "If you need to think about it, you probably need to put it on the board".