Scrape once, spam forever.

I think it's pretty clear you need to use an anonymization scheme in the way commits are handled so that it links back to your github account and the email addresses are kept private.

Privacy centric companies like Apple do this for users offering hashed emails, on a per login basis.

I'm sure this would not work in a world of scraping, but having that kind of ability to figure out bad actors would be nice. You could require authenticated users for certain kinds of requests, and block user information from non-authenticated requests.