I'm new to this stack so my opinion may not be worth much but here it is if you are interested.
While I fully understand that questions about commercial software can be considered "too localized" at times, I do not think it is fair to group all of them together. In other words, saying questions about commercial software are not permitted or are all permitted (although I don't think anyone is arguing for the latter) isn't a proper classification. Instead, I think mods need to use judgement in how "localized" a question really is. Especially as some of these commercial frameworks gain traction, more and more developers are using them and may become less localized.
In my opinion, a question where a dev is using the wrong hook or missed a semicolon and posts the error he/she is getting and a couple specific lines of code is far more localized than a general question about a commercial framework. Is anyone else really going to benefit from the answer to the first question? Maybe in a few one-off cases but probably not very often. However, as it stands now I believe the former question would remain, probably get answered in about 2 minutes, and live on while the latter would be closed and eventually disappear.
This may not align with the StackExchange mission statement but for me most of these stacks serve more as a reference point than a support system. I'm sure we can all relate to googling for some obscure combination of problems and finding that the exact question had been asked and answered 9 months earlier. Hell, I've even seen questions that were years old that didn't really have any activity until several years after they were asked, but that activity ends up helping.
It goes back to what they always told us in school, if you have a question someone else probably does too. I know that realistically we can't have this approach and keep every question but I think before closing/deleting a question a mod should seriously consider whether or not it is possible that anyone else down the road may wonder the same thing.