The author claims the commit hash is bad links to an article... but that claim seems entirely unsubstantiated? Its not the hash that is bad, its the files that are bad. If you manage to hack git to get the commit without the files (clever trick by him), you don't hit any legal issues.
The hash is probably an excellent heuristic for spotting commits from the original YouTube-DL Git log, and so GitHub probably scans for that when searching for rehosted copies.
Agreed, but no one has actually stated that GH is scanning for these commits, let alone make the claim that the hashes are illegal. If you're going to start making legal claims they have to hold up to scrutiny, which this does not.
Am I missing something?