If we want to reach the goal of being able to customize context bars etc on a site-wide basis by 4.6, we can't shove it on all the package maintainers and expect the "perfect" solution in two weeks. Many packages aren't really scheduled for active work for 4.6 and we're resource poor.
So I think Jeff's interim plan is fine. We can always work on further cleaning up individual packages later. Jeff's approach gets us site-wide customization in a very short period of time, which is a real win IMO.