[guardian-dev] learning from the latest leak

[Hans-Christoph Steiner]

Nathan of Guardian:
> 
> 
> On Mon, Feb 23, 2015, at 08:31 AM, Hans-Christoph Steiner wrote:
>>
>> I highly recommend reading this article about the latest from the Snowden
>> leaks:
>> 	
>> https://firstlook.org/theintercept/2015/02/19/great-sim-heist/
> 
> Just so I am thinking about this correctly, this still requires the
> attacker to intercept the over the air radio signals between the phone
> and the tower, and not somewhere upstream on the net, right? Not that it
> wouldn't be hard for a well-funded adversary to do this, or for this to
> be done, say using a stingray device mounted on an airplane...

Over-the-air intercepts would be decryptable with the SIM card keys.  And like
Matej says, they could then _passively_ intercept the radio communications
then decrypt, making it extremely difficult to detect when its happening.
Imagine trying to detect whether someone has a radio that is listening to a
given FM channel, that's the same idea.

But I imagine that cell towers are not responsible for decrypting the messages
and that happens deeper in the telecoms' networks.  So then, the anyone who
can tap the telecom's network can get all the encrypted messages.  This is
something that we have seen many governments around the world setup.  For
example, Libya's telecom network infrastructure was built entirely around the
idea that all traffic had to go through the central intercept and control
point.  Another key example here is that most of Ukraine's telecom traffic
goes through Russia somehow.


>> What is directly relevant for anyone who is working with well funded
>> state
>> actors in mind, is the description of the whole process of profiling
>> organizations in order to find out how they can be infiltrated.  For the
>> most
>> part, it sounds like Gemalto's security was terrible (i.e. plain text FTP
>> for
>> sending secret keys), but it sounds like they improved it a lot, and
>> those
>> targeting Gemalto had to find new approaches for getting key material.
> 
> Back to our Google Play phishing attack... thinking how someone without
> 2-factor auth enabled, and some sort of locked down policy on app
> updating might have gotten their account compromised. This would have
> resulted in, say, a new APK being published, but with a different signed
> key. Still a good attack, as many users would simply uninstall and
> reinstall to get the upgrade, thinking it was their fault the update had
> a problem.

Yes, for existing apps, the attacker could not transparently update existing
apps without having the signing key.  But this attacker could also be working
on getting private key material.  For example, by using
browser/javascript/flash/etc exploits in the phishing link that then runs code
to upload all key ring files it could find.  This exploit could also contain a
keylogger that looks for the key password.  These kinds of exploits are all
documented in the materials on Finfisher, for example, or from various pwn2own
competitions.

One takeaway here: developers should NEVER copy or use their APK signing keys
on a machine that they also use to read email and browse the web.

.hc


-- 
PGP fingerprint: 5E61 C878 0F86 295C E17D  8677 9F0F E587 374B BE81
https://pgp.mit.edu/pks/lookup?op=vindex&search=0x9F0FE587374BBE81