If you asked a group of hiring managers what online resources they would choose to help evaluate a candidate for a job that would require a significant amount of programming, what do you suppose the prevailing answers would be?  In all likelihood, you'll get a mix of Github and StackExchange.  While extremely useful in certain ways, I would include another medium, even if its usage for evaluating teammates is mostly hypothetical.  I'd argue that viewing a candidate's behavior and displayed competence on IRC networks is more useful in ways that that Github and StackExchange possibly can't provide.  

For starters, IRC itself is a bit of a technical shibboleth in ways that differ from at least Github.  Depending upon the choice of IRC client, in order to actually join a channel you would have to figure out how to get a client to connect to a server, likely register yourself with a nickserv, and figure out how to list out and join channels.  Many clients hide this behind nice GUIs and menus, but it's still a bit of a ceremony when compared to more recent chat systems.

If you think that alone is silly criteria and that any monkey could connect to an IRC server, consider that in the past a former colleague and I have actually decided not to work with specific developers before in a professional setting because they literally could not figure out how connect to freenode and join a keyed channel.  They instead wanted to use email and Skype, which are both wretched for team communication.  Let that sink in for a moment – if a person's resume has all of the criteria that says they are a great technically, but that person lacks the persistence or savvy to even use to one of computing's most useful enduring communication protocols, especially given how easy it is in modern graphical IRC clients, then do you really want that person on your team?

Going even further, you can deduce even more about how technical and curious a person is based on their choice of tools for IRC.  A simple /ctcp $nick version would be quite insightful to the tastes and possibly technical floor of the user.  If the response of that version ends up being the name of a bouncer or a command line client, you can probably assume this person is going to be rather comfortable working out of a terminal shell, which is in of itself a shibboleth one should place much weight in.  Taking tooling to a logical extreme you may eventually find out the person maintains/develops an IRC bot or is making use of or writing plugins on their client – though getting this deep in to what tools a person uses for their IRC experience may be taking the technical aspects a bit far beyond what is useful. 

Truly though, that's not where IRC's strength in evaluating a future teammate is.  Github's technical shibboleth of requiring you to actually use git (which is itself a much larger undertaking) makes it better in this regard.  Where the fun really begins with IRC is watching a person interact on an irc channel:

* Do they know how to ask questions to actually get answers and what is
  the nature of their questions?
* How well do they help other people with their questions?  Do they
  spoon-feed answers or 'teach a man to fish'?
* When are they a combative prick and when are they nice?
* How do they handle others' trolling?  When do they themselves engage
  in trolling?
* Can they communicate well enough in text without aolbonics or
  sounding like a college-aged stoner?
* How well do they adapt to the culture and (unwritten) rules of a
  network or channel?
* How much do they influence the discourse of a channel?

If the above sounds like unnecessary and contrived personality analysis, substitute "workplace" for "channel" and then rethink the importance of knowing these things about a candidate.

The first three bullets are pretty damn insightful in to how one's brain works, how a person reasons about problems, and how they go about solving them.  A potential coworker is going to be dead weight on your team if they are the type to ask a question on IRC without first using a search engine, or trying to solve it themselves by asking their interpreter/compiler, or attempting to produce a SSCCE.  The lack of doing at least one of those three things shows a complete lack of respect for the time, attention, and mental energy of those around them and is indicative of selfishness.  Conversely the person who tends to ask very smart questions, jumps on the difficult to answer questions, and provides thorough and/or Socratic help is going to raise the tide for all around them.

The other points are more cultural and give insight in to personality and culture-fit.  

Being a prick: Being a prick is generally not a good thing, but it is conditionally necessary. When it's time for you as a coworker to sternly call something or someone out, then you need to be able to don your prick hat and do it.  If you can't be a prick on IRC then you can't be a prick anywhere, really.  On the flipside, you don't want someone who is even close to being a prick half of the time either.

Trolling: Trolling is a bit of an analogue to passive aggressiveness and/or sarcasm - both of which are poison.  It is a no-brainer to avoid people who are trolls, but what is equally important is to avoid people who waste time feeding the trolls and become, at best, distracted by trolling behavior.

English:  I've almost universally observed in reading the text of others that the ones who have to abbreviate "you" as "u", "question" as "q", "because" as "cuz", and such forth are either 1) slow typists 2) people who tend toward sloppy and often unnecessary shortcuts.   You want neither of those two things from your teammates. 

Adapting and Influencing:  This is mostly about culture fit and leadership quality.  Can a person shift just enough to make themselves fit into a culture and can a person actually influence the behavior of others? 

Sure you could possibly glean some of these things from Github, StackExchange, and others, but I'm of the opinion that such assessment is tarnished to the degree that I can't evaluate truly what kind of person I'd be working with by viewing their behavior on those platforms alone.  Those platforms are their own cultures with their own rules and are gamified to a large degree by the owners of each platform.  Further, those platforms are fully exposed in public, have a very low barrier of entry and acceptance, are professionally moderated, and nowhere near as anonymous as most of the rest of the internet which puts many folks on their best behavior.  By comparison IRC is the Wild West.  If you want to find out who exactly a person is, get them out of the public eye and reduce the number of rules they need to follow - you'll find out with much greater precision who and what you are dealing with.