That's not what I'm seeing. If I revert the bugfix I made in this PR and run the tests, the "should not open accordion via keyboard navigation when accordion is readonly" test fails, and no other tests fail. This is what we expect, since this is the one case this PR addresses. The other tests I added are for behavior that was already correct, but didn't have test coverage.

If the keyboard & mouse tests are redundant with each other, I would expect either both to pass or both to fail in any situation.

Put another way: if I removed the keyboard tests (which I think is what you're suggesting), all the tests would pass without my bugfix being applied.