When is storing configuration in Windows directory good idea?
Well, I know this is quite controversial, but in some cases (specially scripting frameworks) I think it is good idea to use Windows directory to store some settings\configuration...
Why that? Generally you can always divide such solutions to push or pull (and mixed of course) - that means that (if we talk about servers), sometimes you push some data to server (from your workstation\laptop) and sometimes you want server to retrieve information from some central storage\share.
Windows directory is suddenly only directory that is easily accessible both locally and remotely - from remote system you can use \\Server\Admin$ and from server itself you can simply use %WinDir% variable.
I know most people would say that you can use \\Server\C$ and %SystemDrive%, but you never know if C: drive really exists (this is happening in SBC environment when you want to have client's C: drive mapped at C letter).
So for these reasons I like to use Windows directory - and then I make symlink\junction to some other location (so for example C:\Windows\CentralStorage\MyFramework --> C:\MyFramework). That way all required files are easily accessible for scripts (both push and pull) and also for administrators or people working with such environment.
There are of course another ways how to achieve that - create new shares etc, however this method is most reliable and can be easily accessible for all scripts\tools and also for all administrators.