I don't think you need to cite it. I think you say "We developed an in-house script to ..." and then if the journal needs it you upload to GitHub and reference however.
If you really want to though, I do know that you can publish to a GitHub repo and get a DOI so your program is citable, see this link.
Pretty handy IMO. My question is whether or not a piece of software with a DOI counts as a publication on ERAS. I'm going with yes, because it's a piece of work that others can contribute to, reference, etc., but I would probably list it under the "non-published" category or whatever. Also it goes without saying, but in the chance you get a reviewer who knows his software engineering, that you may be asked to make the code more clear, comment it out, etc. Wouldn't want crappy code to be the reason a paper doesn't get pub'ed. And while I do think that all code contributions that are engineered well are worthwhile for the community, if the "short script" is a wrapper for something or is hard-coded or is trivial, I don't think that is necessarily is need of being "published" (again, outside of the journal asking for it).
