I recently installed Cerberus in my [personal] Android device to get some protection against theft (I'm not really interested in protecting its data, but rather in increasing the odds the device will be eventually recovered). But I'm worried that using this App (or similar others) might expose me to a greater risk of losing control of my own device, either by accident (bugs in the Cerberus system) or mallicious activity (attackers taking control of Cerberus' website).
Clarification: the question is not about how to physically protect the phone, but rather about the digital risks I'm taking when employing a computerized solution to said problem (with the constraints that this scenario impose - attacker with physical access to device, no guarantee of internet connection, etc).
Please note that I'm not requesting a shopping recommendation, but rather trying to work out all the details I should be aware of when looking for (or maybe even developing) a solution. I know compromises will be made, but I'd rather not be more lax than necessary in the digital aspects while I try to harden the physical ones. And the issues are not necessarily about Android, it should apply to any mobile device.
Feature: allows remote control of the device by sending it SMS commands.
- Why it's necessary: it might be the only way to reach the phone if internet is disabled.
- Concern: the password is sent in the SMS body.
- Mitigation: change the password every time this feature is used (after the device recovery).
Feature: allows remote control of the device by using a web-based interface (the user can enable internet access by a SMS command if necessary).
- Why it's necessary: a server is needed to do the hard-lifting - communicate with the device, collect its info and do something with it (usually e-mail it to the legitimate owner). While simpler commands might be possible without it (ex.: enabling the GPS and sending coordinates back by SMS itself) others are not, due to the message size (ex.: taking a picture or recording a short video with the front camera while the thief is trying to operate the phone and e-mailing it to the owner).
- Concern: if the server has enough credentials to authenticate itself with the device (otherwise it couldn't control it), then if the server is compromised the device becomes vulnerable to several attacks (locking the legitimate owner out of his device, tracking it, or wiping all its data).
- Mitigation: a) using a server that you control, not a third-party; b) using some kind of OTP that the user must provide, so the server can't do anything if the user is not in charge; c) disabling features with damaging effects (i.e. enable finding the phone disregarding the privacy risk, disable wiping the data disregarding the data theft risk, etc).
AFAIK these two features are the only ones that could be problematic, but if anyone has more info about things I should be careful with, it will be welcome. As an additional clarification, I'm aware that someone with physical access to the device - and enough knowledge - might be able to get rid of the App (though this one in particular was designed to survive even a factory reset). But I'm assuming it's not a trivial task, if I'm mistaken - and trying to recover the stolen device is in fact pointless - I'd like to know it so I can abandon this idea altogether.