Dear Asyn and others that follow this thread,
In the light of the recent attacks against CCleaner with redirection to controlled C2 servers by sophisticated state hackers, known as Group 72, we should also consider the following insights:
The recent actions againgst Asian C2 servers:
https://tweakers.net/nieuws/123911/interpol-en-beveiligingsbedrijven-identificeren-8800-c2-servers-in-zuidoost-azie.html (translate to English using Google translate).
Because of collision issues we can no longer profoundly trust MD5 or SHA1 hashes. NIST recently removed a weakened NSA-algorithm
and NSA has difficulty getting two new weakened distrusted algoritms approved:
http://www.reuters.com/article/us-cyber-standards-insight/distrustful-u-s-allies-force-spy-agency-to-back-down-in-encryption-fight-idUSKCN1BW0GVBut then after the Snowden reports, who can trust a "burglar that sells locks"?
Another issue: Dual EC DRBG is a "cryptographically secure pseudorandom number generator", something that generatess streeams of bits, that are quasi-random, and one cannot tell the difference with real randomness. As such a tool in that is not an encryption algorithm, but it should have a place inside the crytographer's toolchest. Well this one should be quarantained, as it does more wrong than it is worthless as such.
And despite of that RSA Security (the firm by that name*) has Dual EC DRBG installed as per default, while there are much better choices available.
Is not that a coincidence? Why anyone should ever now believe NIT anymore?
Wanna have a go at it: download LCPT_gcc.cc program from directory: wuala.com/FreemoveQuantumExchange/Aspects/Randomness/Theory/Berlekamp-Massey
source code is there as well.
When you start to test files s01.dat and s.02.dat using the LCPT_gcc.cc program, it appears complexity halts at 19937
and does not go further, which is the complexity of a Mersenne-Twister. Whenever using Mersenne
to be found inside mentioned directory generate pseudo-random files and test those you will find the compexity is 4*19937.
This is why per output (of 32bits) 4 bits are being sampled. In the same way one can test the output of the Microsoft PRNG,
see that same dir. One would find similar results.
Now we see why with CCleaner the 32-bit versions were compromised. We know the trick now that the l33t hacker(s) used.
Is it not kind of weird that security organizations and state agents wanna undermine everyone's security with this kind of nonsense/crap?
So you can create backdoors when you alone own the secret key. Sort of similar to a normal public key scheme.
polonus (volunteer website security analyst and website error-hunter).
P.S. It should be a concern that the Microsoft Windows certificate store (you find it inside the registry) identifies certifivcates 'uniquely" on basis of their SHA1 hash - collision can not be avoided under all circumstances.
SHA1 is unsafe
