I see you're getting at if there's a common group of "scope things" that we're dealing with here...
They're not very similar other than they're all introducing some form of scope in the traditional sense of the word, just by being a with... bounded piece of code.
Semantically they're a bit different though, so I'm not sure about stating they're the same group with just a different "kind" of the scope.
- cancellation shields: the primary semantic they introduce is a boundary (well, "shield", we named those nicely I think). As the only place where the effect matters is at the boundary between the "before-the-method" and the "inside-the-closure" -- we prevent cancellation propagating through this piece (and we prevent a deadline lookup up through it -- amendment here).
- cancellation scopes, on the other hand: don't really do anything at the enter-the-scope boundary. Their effect is confined entirely between the "inside-the-closure". There isn't any interaction about the outside/inside really. It does of course affect everything "inside the scope" if we cancel it, but there's no interaction with the outside.
While it's worth looking for some commonalities but I'm not sure I'd describe them as the same "scope" concept in any more formal meaning of the word. I do think the previous word "shield" conveys its semantics very well, and while technically one could say "a scope into which cancellation does not propagate from parent scope" I don't think this is more helpful than explaining it as it's own thing.