For starters: right justifying the password entry screen on the android app is less than ideal. It doesn't enable me to ensure I haven't fingered in my password correctly by monitoring general length. I know that seeing the length of the password might be less than ideal on a PC, but the chances of someone overlooking on my phone screen justifies that trade-off, IMO.
I think, secondly, that you are facing the issue of fragmentation on Android to a point where no app on any version works. To that end, you need to invest heavily in codifying your core logic in a higher level abstraction (say phonegap or whatever is the new hotness), then invest in outsourcing to some coders that can make bluetooth+android permissions library extensions. From there, load the correct library per your platform via the 'core' logic, to deal with the broken part that is android permissions and bluetooth HAL. That being said, any bluetooth product is going to be plagued with these issues when you want to secure the connection. Bluetooth itself is a POS standard when it came to security, and it is only made worse by the non-standard handling you are now facing between Android and iOS, let alone multiple versions of the same!
I think that for a future hardware rev (assuming you fix the major issues that prevent me from using my own android app, and prevent proper sharing of keys to folks who use android from iOS users), you should probably think of using electric imp's method to pair the separate wifi ap to the lock. That way, you could just have a web browser to configure the AP, pair the AP with the lock, and then pre-stuff whatever bt logic/security keys from the lock into the phone. You should then, at that point, make the unlock mechanisms a service that runs in the background, vice a full blown app.
The app itself should only be for management of keys and toggling of features (and, later, remote control via the web).
There is also the issue that your lock is unaware of the door position. This isn't a deal breaker, but there may be some though put into how you could resolve that. Perhaps have an adhesive strip that contains a magnet, and use that as means to trigger a magnetic reed switch on the base of the lock. That way, you don't consume any power, and it is one of the methods by which you can keep your lock powered down until an interrupt is made on that reed switch's pin. . .
Don't want to design for free, but I really had hoped this product would have been better than it was. I blame it on the separation of the wifi capabilities and the complicated nature of security on bluetooth. I get the reasoning, but that should have gotten more effort rather than color schemes. Great job on the physical mechanisms though. I hope that the mechanical failure rate is low considering it is used at least 10 times a day.