In the course of looking up something else, I found this headscratcher.
Over the past 6 months or so, the YandexBot--the real one afaik, not an impostor--has been asking for nonexistent URLs with name in the form
/\h{8}-\h{4}-\h{4}-\h{4}-\h{12}
for example
/374c902b-7a1f-4998-b7e0-878ca7b04e62
where \h is short for [a-f0-9] i.e. any hexadecimal character. (SubEthaEdit permits this useful shorthand; sadly I don't think anyone else does.) No extension.
Multiple sites, always a completely different string, but always the same pattern. Does anyone recognize the format? Is it, for example, something used by a major CMS (and, if so, would it be equivalent to google's routine soft-404 requests)? It's got no resemblance to the site-verification file* they use in their own WMT.
As noted in other threads, yandex does have a habit of doing weird and inexplicable things. But at least it doesn't prompt the kind of jaw-clenching terror you get when Google does something new.
* /yandex_\h{16}\.txt
