## Problem/motivation I was testing in prep for beta1 and noticed this behavior which I think is okay, but I want to talk with the UX team about it. 1. Context item is scoped to use case (writing words) and entity bundle (article) 2. Agent scope subscriptions are use case (writing words), tag (foobar), site section (foobar), and entity bundle (blog) In this case, the context item is injected into the agent's prompt. This is because they both have use case = writing words. Let's discuss if scope should be handled like this. I think it makes sense. ## Next steps - Discuss - If necessary, make follow-up to tweak logic - Either way, make follow-up to make sure documentation is clear and update if not, and consider adding a note to the UI