- I noticed this in a PR recently, maybe the maintainer didn't think
they would have to run `brew man` for this? In that case there's an
extra step after the PR is raised, red CI checks on the PR (which is
demoralising) and potentially requires a reviewer to point this out.
- This runs `brew man` if the `README` file has a diff as part of this
command. Then the user has to `git add .` and their PR is (probably)
good.