Yes, that's all I've seen and yes I've reviewed it.  I've not had a lot of time to go into minute detail and step through the code as it would potentially operate, but I see nothing nefarious in it. I informed them that whatever they sent me is what I would be compiling and using on my own hardware that they have not had access to prior to the test... so what I've reviewed is what I'll be using to conduct the test(s).  

Most of the code is actually already out in the Bitcoin community and well vetted, from what I can tell they've only made some minor changes to to accommodate the box as opposed to a GPU.  There's actually some optimization that can be done, I think, to speed things up.  If what they have now lives up to the claims, with code optimization, I think you could see an increase in hashing power.  But, that is purely speculation on my part.

I will inquire and I don't see any reason they would be opposed to releasing the code to other, more qualified individuals to review the code as well.  I will send them the same code I use to test, as opposed to BFL sending them code, assuming they (BFL) approves and someone is both qualified and willing to review.