Truecrypt or not truecrypt…

POSTED BY: Dave B (Staff Writer)

…that is the question.

Isn’t it?

Recently, I’ve noticed a surge of topics on reddit and the like that continue to perpetuate the rumor that Truecrypt (a popular volume encryption utility) has been compromised or is otherwise unsafe to use.

According to an audit conducted on the Windows version of Truecrypt 7.1a, Truecrypt does not contain any backdoors or serious flaws that would compromise confidentiality under normal circumstances.

Does this also rule out the Truecrypt site itself serving up bad executables to selected IPs/blocks of IPs? No, but using the Github mirror does (find it at: GRC) and verifying all signatures are the same as used in the Open Crypto Audit Project (here: OCAP). Alternatively, you could build it from source using code from the above listed locations.

Still feeling a little uneasy about Truecrypt? That’s OK – a little paranoia can be healthy!

Sadly, in the digital age, it is much easier to ruin a reputation than the decade it takes to build one. There have been instances where the government admitted they were unable to crack Truecrypt (as of 2010 at least: according to this article). That being said, one case stands out (here) where the FBI was able to gain access to a TC volume – although it is unclear whether this was accomplished with technical means or otherwise. Also, see A long list of instances where police weren’t able to decrypt volumes from various popular applications.

Linux is your friend

If TrueCrypt has lost your trust for good, then I’d suggest migrating to Linux and using a tool called “GnuPG.” It is based on the open-PGP standard and makes managing asymmeric keypairs simple.

Also, a special Linux distro called Ubuntu Privacy Remix is available as a .iso for use on a live CD. It creates encrypted TrueCrypt volumes that you can place your super-secret PGP ciphertexts in.

I’d recommend the following settings for UPR volumes:

  • Use the TWOFISH-SERPENT algorithms for the wrapper volume (the one you’ll place your asymmetric ciphertexts in)
  • Use Whirlpool for your hashing algorithm
  • Use a passphrase that is at least 30 chars long, and make it random
  • Use DSA-Elgamal 4098 bit keys for GnuPGP
  • Again, a long, random and secure passphrase is required here

This set up provides you with a margin of safety.

Since UPR saves nothing to the local disk, it is safe to work with files on its desktop. For instance, you could first encrypt the sensitive file with your public key. Next, mount the TrueCrypt volume and place the ciphertext (encrypted file) in it. TrueCrypt will never have access to the plaintext, so even if it is somehow secretly compromised, the ciphertext is safe.

The down-low?

Its simple. TrueCrypt is 99% likely to be safe. Its been verified several times and even passed a bonafide audit with a pretty decent score considering TrueCrypt was not created by a multi-million dollar corporation or state actors. The developers invested a pretty good chunk of their lives into this, and any future discovery of malicious intent withstanding, recieved a lot of flak and permanent damage to their reputations that was probably not deserved.

TrueCrypt is most likely perfectly safe to use, provided that you fully understand what it can and can not do. If your end-point is compromised: it won’t help. If you have the  volume mounted when an armed attacker busts into your home: it won’t help. If your powered down machine is stolen by a burglar: it will do its job. There are plenty of other limitations too, and those were just a few examples that people tend to forget. Overall? I’m still using it. But, then again, my threat model doesn’t include state actors (at least not from this country). But I’m sure others’ do.

One last thought: you’d better not even be obessing about this unless you have end-point security handled!

2 thoughts on “Truecrypt or not truecrypt…

  1. The fact is that encryption presents few challenges to the intelligence agencies, who will chose individual PC compromise over network level where crypto does its job (and if it is being used). Oh and at rest too, but as you said it doesnt protect it while running or if passwords are in RAM.

    Then again, the people worrying about the government in the first place are probably but not always hiding something significant lol


  2. JALEX88, congratulations on being our first reply!

    I agree with your first point, especially given the absolute insecurity of most end-user systems. The only concern in this regard is the cost/benefit aspect. Whereas existing infrastructure was already collecting communications, specific compromise must be achieved via physical attack or remote exploit – neither of which are as cheap as collection from infrastructure already in place.

    So, cryptography is only helpful when A) The secret key is known only to you; B) The sensitive information (=plaintext) is in the secure (=ciphertext) format; C) No mechanism is present on or in the general area of the system used for encryption to capture the secret key or plaintext. Encryption is literally the process of applying an algorithm to a secret key and generating some output unable to be reversed without the secret key or exhaustive search of the keyspace.

    Many people have legitimate concerns of privacy and need to protect proprietary or sensitive information from thieves and hackers who could profit from it. Personally, the Government of the US is not included in my threat model – however, foreign state level actors are (N. Korea, China, ISIL, etc)!

    To those who do not trust the US Government, I understand. In light of the secret SIGINT programs, many people feel that some trust between the people and Government has been damaged. While this is a natural outcome of such revelations, it is prudent to realize that there are many other ways they could achieve this task without weakening encryption through backdoor access. My point is that cryptographic systems are likely not going to be the weak point in ANY threat model, ever.

    For those of us focused on the “non-state yet advanced adversary,” TrueCrypt will certainly provide protection of data that is encrypted and unmounted who’s secret key is known only to the creator. Remember, a huge percentage of data theft involves private actors and not the Government of any nation. I feel that security research should be focus on the most critical and largest attack surfaces first. Then, as each threat is mitigated, continue down a list of other “pretty bad, but less likely” incidents. Rinse, repeat!

    Security is never a solid state. It is actually best to think of it as a number line with a constantly moving point. This literally changes hourly, as new exploits are discovered and then patched – a sort of tug-of-war between security specialists and malicious actors.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s