The author's example of a conventional commit is not correct anyway IMO, which is maybe why they think the "fix" part is redundant:
> fix: prevent foo from bar'ing
The whole idea of conventional commit is:
> fix: [problem]
so the correct conventional commit would be:
> fix: foo bar'ing
which is succinct and perfectly fine.