I did look at the source, and I see that's it's HTML. (I hate trying to enter code on HN, so here's a gist with the full source of that page: https://gist.github.com/1116746)

But it's only a partial page. There's no doctype, no head, no formal declaration of body. There's HTML and CSS for the browser to render, but it's guaranteed to be in quirks mode (unless I'm confused). That's why I said "full HTML page" above. (Edit: I see now where my first post was unclear. I edited it to make my complaint clearer. Thanks.)

I'm not quite sure where this is useful - I may not be understanding what it's for. I would hate the idea of lots and lots of sites serving this kind of partial page.

Only the doctype, title, and the loss of </link>, are required for it to validate as HTML5; <head> and <body> are implied.

I think the original `markdown` utility generated a HTML fragment rather than a standalone page. If so, perhaps the point of this is that adding the CSS link is the Simplest Thing That Can Possibly Work in order to improve the page appearance?

   adding the CSS link is the Simplest Thing That Can Possibly Work in order to improve the page appearance?

That's correct. It was useful for me, and I figured it might be useful for others as well. The initial use case was for a README for a repo.

It looks to be just a css file to make Markdown files pretty when straight-compiling to HTML. It's better than keeping them with the browser's default stylesheet, but I don't know anyone who serves Markdown files straight up after a conversion...

> but I don't know anyone who serves Markdown files straight up after a conversion...

Exactly. This is why I don't see the use-case for this.