My point was mostly that there isn't a real precedent for removing "map bugs" as a whole, even if they have large implications for runs.
If we ignore precedent (since it is very contradictory and subjective), I think it makes sense to argue this from basics, and I think these are some good precepts:
1. Tempus should only intervene in gameplay if it's mandatory to make speedruns or basic gameplay work (e.g. patch out level select cheats, massive sequence breaks like 4starters, adding regen, removing softlocks).
2. For new maps, Tempus should hold maps (and mappers) to contemporary standards
3. For existing maps, Tempus should do as little as possible to preserve the integrity of existing times.
This skip is constrained to one level and doesn't involve leaving the bounds of the level as defined by the intended route. It also doesn't require a second player or garbage like resetting the map so this doesn't violate the first rule.
The map is ancient, so it shouldn't be held to the same standards as new maps being considered for adding.
It's disingenuous to pretend that this discovery is different from a Tempus change in any way other than origin given it's implications, so this violates the third rule. However, the party which is "wronged" by this is only those people who have invested time into speedrunning since it's a T3. A substantial majority of runners are in favor, so I think this can be an exception.
If Tempus admins believe that something is sufficiently broken to warrant debate (new skips, strats, etc), I think it's perfectly fine to preemptively stripper it and then poll interested parties (could be all completions if it makes a hard map easy, e.g. One Quick Trick To Beat Rickoconnel) before rendering a final decision. But if the argument against changes is that it invalidates previous runs, and the holders of previous runs are fine with it, it's overreach for admins to modify the map to prevent new strats.
More generally, I think this can be applied to "fixing" existing older maps as well. For example, many maps have "rng" telehops which require getting good tick-based spacing in order to avoid bonking the tele door and maintain momentum. As someone who has spent a good amount of time running such maps (e.g. bouncerific and arc), I'd be perfectly happy with Tempus strippering the tele out and back in much larger to make this consistent (or smaller to make it impossible!), despite the fact that this may make my times easier to beat. If a supermajority of relevant players desire such a change, I don't see why it shouldn't be done.
It's true that skipping "1 level" means half the map by time here. However, you can't ignore that the jump being skipped is pogoing in place for 30 seconds. Arguments for angle bhops on soldier usually focus on how they skip long downpogos; I think the same logic applies here. Likewise, the existence of the speedrun strat doesn't fully replace the "casual" strat.
Specifically regarding glow, I think that map's a great example. If glow were added now, it would get rejected for a multitude of reasons, so we'll pretend that the course skip was just found. If we follow the argument I outlined, the admins would stripper it to make the cheat impossible, leaving whatever newly "cheated" times were attained, and then poll top 10 time holders (both classes, using discretion if there are any other interested parties e.g. gato's offline -10 wr run). After some deliberation or even public discussion, either the stripper code would be removed, or the "cheated" times wiped.
I guess my question is: What group is hurt by allowing the green box skip? The only Tempus-motivated parties are tt/wr holders, who are largely in agreement. I don't think anybody has ever run the map offline, and it's not like this makes it easier. From a public image standpoint (if anybody cares...), the Tempus Records video looks a hell of a lot better with an instant tele than sitting in Green Box for 30 seconds.