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.