A small handful of scripts keep on changing their script_checksum and so ask to continually applied

We have our database set up with a change automation project using a shared development model - one single database where everyone makes changes, and then each person generates migrations from just the changes they made.

One annoyance we have is that almost every time we pull to get the latest version of the git repo, the add-on says "there are unapplied scripts" even though everything is up to date. So we click apply, it makes no real changes (it just alters stored procedures to the state they are already in), then do our work.

Digging into it, I believe it's because the script_checksum for some files keeps on changing for some reason. If I examine the __MigrationLog table in our dev environment, for one of the objects it looks like the checksum keep on flipping back and forth between two values, as it gets applied by the different users (and yes you can't see the whole checksum in this screenshot but I've double checked and there are only two distinct values)



I don't really know how that checksum is calculated, some MD5 hash I guess, but is there something I can do about our settings to stop this from happening? Is it because user "tre" maybe has some whitespace setting changed, or some weird Windows line ending thing, or something different about Git so the file encoding gets changed slightly? How can I narrow this down?