I would never limit access to a feature on my site based on badges someone has earned on another site. It's like slapping a big sign up that says, "In order to use this site, you also need to sign up for X, Y, Z, ..."
It's hard enough to get users to sign up for my stuff, let alone make them sign up for something else. Throwing in additional roadblocks is definitely not the way to go.
That's not to say I'd be against showing the badges they've earned elsewhere. That would be nice, I suppose.
Another problem: If things go very well for you, you'll end up with hundreds of partners all pushing badge info. On the other side, you'll have consuming partners that want to use this data, but they'll need to know exactly what badges to extract. Not every user will have the same badges from the same sites, so the consuming sites have to do the legwork of deciding which badges they want to display.
Further, if site X decides to change their badges, all sites that consume from X will need to change as well, or just keep displaying the now stale data.
Seems like too much communication will be necessary between all the different parties.
I'm not exactly sure what legwork you're talking about for the 'consuming' sites. From what I understood of the description, the KaBadge site is the sole repository of badge information, with consuming sites pulling from that data when displaying a user's information (hence the API). As such, consuming sites will simply access that user's full or partial list of badges (possibly sorted by importance, or order of acquisition, or a genre system?).
A consumer might have to do specific badge checks for restricted sites like the WoW forum example (I agree that I would never use this feature myself, but I can see some value in it), but they would clearly know exactly what they were looking for beforehand.
As HN, I obviously don't want to display the HotOrNot badges, probably just the SO stuff. So, I set it up to pull those. Not too bad. Now imagine that there are hundreds of producers, some that make sense, and many that dont. You probably have to do a fair amount of digging to find those that make sense for your site, and a lot probably fall into a grey area. Then, you may find that only 1% of your userbase uses any particular 3rd party badge-producer. So you have a choice: include lots of badges, most of which will mean absolutely nothing to 99% of users who see them, or include only the most popular badges, which means the service ends up dominated by a few big producers (Twitter, Facebook, etc.).
I guess I'm having difficulty articulating, but I'm just thinking that as someone who runs a community site, I don't want to go digging through dozens of other sites' badges to determine what they mean and if they're applicable to my site. Plus, I don't want to have to re-do this every couple weeks to add any new badges that have appeared in the system. And, the only end result I see is being able to decorate user profiles with bling that is meaningless to the majority of the other users on the site.