I remember reading somewhere that once You include some url in robots.txt archive.org even though can have it archived it will stop showing it to the public.
The first page you linked tells the story of how the Wayback Machine once deleted archived content if robots.txt directed it to, but was moving towards not doing that anymore. You can also see in GitHub’s robots.txt that URLs matching `/*/tree/` should already have been forbidden, yet the Wayback Machine at one point had https://github.com/github/dmca/tree/565ece486c7c1652754d7b6d... archived. These facts suggest that it is your second theory that is correct—GitHub contacted archive.org to get the archive taken down.
I remember reading somewhere that once You include some url in robots.txt archive.org even though can have it archived it will stop showing it to the public.
https://blog.archive.org/2017/04/17/robots-txt-meant-for-sea...
https://github.com/robots.txt
They have some interesting nuggets in their robots though like `/ExplodingStuff/` and `/account-login` both of which seem to be some accounts.
Or probably more possible - they just got in contact with the archive.org people.