The comments on How to handle wealthy player characters as a GM? got a bit out of hand, but it gave me an idea...
The contention seems to be between these two perspectives on how to use system-agnostic:
[system-agnostic] should be for questions that are useful without knowing anything about the system being used, so it should, as reasonably as possible, apply to all systems possible.
[system-agnostic] is about avoiding system-specific answers, so it only needs to apply to a goodly range of systems, not all or nearly all.
Is that a good summary?
One comment stood out (emphasis original):
Your answer is only applicable to classical fantasy broadly based on the medieval period in systems that try to accurately simulate that period's economics. Consider D&D 3.5 [...] The wealth in 3.5 isn't meant to model medieval economics, it's meant to throttle the rate at which characters acquire more powerful magic items. The system is in fact a ludicrous economic model.
This has been my gut feeling about the [system-agnostic] tag in the past: that it should only apply if the question can truly be said to be answerable across every game.
Now, though, I think that's an unreasonably high bar to set. There's a lot of good use in questions and answers that are "unaware" (agnostic) of the exact system that will be used, while still being constrained to a particular kind or style of mechanics or play. There is very wide community support for using [system-agnostic] that way.
It's possible that we need some tags for play style–specific questions.
Rather than trying to make [system-agnostic] do duty as a catch-all for every question that is broader than one specific system, maybe we could use tags that are game-agnostic while still being limited to a particular kind of game. Maybe that would eliminate the arguments about [system-agnostic], since they seem to largely turn on the point that a particular question is still limited to a small set of games or a mode of play.
We already have sandbox and megadungeon, two well-recognised (because they're old) modes of campaign play. We have dungeons-and-dragons, which gets used as often for "the genre created by and around Dungeons and Dragons" as often as it gets used for "all D&D systems" questions. We have battle-map, which sometimes gets used in a system-agnostic way to indicate a common mode of roleplaying out combats. We have fantasy, science-fiction, and post-apocalyptic for questions about roleplaying within those literary genres. (Perhaps too broad, considering they're rarely used?) We have realm-management, which gets used for that play style as well as for mechanics for supporting it. gm-less covers a niche mode of play. I was surprised to find that we have simulation for that part of the GNS! Even optimization can be seen as a play style tag (it's certainly useful for avoiding questions that are irrelevant for my preferred play styles). We even have a tag to ask questions about different gaming-styles.
I don't suggest going on a tag-creating spree, even if more such tags would be useful. The question is, would play style tags be useful? If we decide that tagging for play style is useful, then simply deciding that is enough. That will put us individually on the look-out for opportunities for new, useful tags as questions come up.
If such tags wouldn't be useful though, then that's good to establish too. That would mean that [system-agnostic] is the right tag for questions that might be limited to a particular play style by the body of the question, but which shouldn't be answered with solutions relevant to only a particular RPG.