I'm in the process of creating a toolkit and game:
www.CircumReality.com . It's still a ways from being done, but a working version is available from the site.
I have implimented the following usability testing hooks:
1) It's multiplayer and everything is logged.
2) Not only is everything logged, but it's easy to keep metrics, like the average time to complete a task.
3) An administrator can "spy" on a player and see what they're seeing on the screen. Unfortunately, you can't see a video of the player's facial expressions or read their minds.
