Thanks for making this PR. I didn't realize the linter would complain about this solution. Gonna tag a couple of people to get some thoughts on it.

@MaoShizhong @thatblindgeye So with this the situation is that we have a heading that's ### The each_with_index method. Our Kramdown code strips the underscores from the href name when creating the section links. So to have a working link with the heading as is, the link's href must also be stripped of the underscores. But doing this has mdlint complaining because it now thinks the link's href isn't a valid reference to any heading on the page.

Do either of you have any thoughts about the best way forward on this?

@JoshDevHub Let's accept this solution for now since we need it for the link to actually work.

Related is #28330 - markdownlint uses GitHub's section heading algorithm which removes /[^\p{Word}\- ]/u, i.e. underscores are not removed as they count as connector punctuation. I didn't realise there were even more differences between that algo and Rails' #parameterize method. This definitely throws a spanner in the works and my custom rule is no longer a sufficient fix IMHO.

Initial ideas which I'll discuss further with Eric and Kevin:

1. Disable MD051
2. Disable MD051 then write a custom version that uses Rails' #parameterize algo
3. Change our website to use the GitHub algo instead of #parameterize, though this will break a potential many links which we'd need to lint the whole curriculum for to fix any MD051 violations.