No second repo needed: the same remote can do both halves. Clone with --depth 1 (2.4 s, timed above), then run git fetch --unshallow in the background. I just timed that against gulkily/v3 and it pulled the rest of history in about 7 s. A release repo would be one more thing to keep in step with main.
The other option is git clone --filter=blob:none: 3.2 s and 22 MB, and git log shows all 5,026 commits right away. Old file contents only download when something asks for them, like git log -p.