Found via https://www.zylstra.org/blog/2018/07/wrapping-my-head-around-webmentions/: http://diggingthedigital.com//Waar-te-beginnen-met-Webmentions/ has no explicitly marked up u-url and all content wrapped in tags except a single <a href="https://news.indieweb.org/nl" class="u-syndication"></a> to submit it to IndieNews. Since there is no explicit url property, implied property parsing triggers, and this situation matches the second rule for implied urls:

else if .h-x>a[href]:only-of-type:not[.h-*], then use that [href] for url

This clearly should not happen.

At first glance, it seems like a rule similar to the recently introduced for p-name would be sensible: do not imply URLs if there are other urls marked up?

In general, I repeat my question from elsewhere: is there documentation for why the implied rules are like they are? They keep causing various amounts of trouble, and I'd love to understand why they are necessary as they are when discussing them. They clearly cover more than the tiny minimal h-cards: what are they intended to cover?