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.