Between Source Code And Cyanide

What the Symantec source-code leak really means
With all of the talk about source-code theft, extortion attempts by a shadowy probably-Anonymous-affiliated group, and most importantly the Giants winning the Super Bowl, I thought I’d spend a moment to reflect on what the release of source code for PCAnywhere and, in all likelihood, a depreciated version of Norton Antivirus could mean for the average persistent threat.

Much of the commentary on the topic, specifically that relating to PCAnywhere, has downplayed the release, owing to the age of the stolen code and number of active users. While it is true that installations of PCAnywhere certainly do not seem to be as widespread as they once were, it is certainly still out there and remains utilized by large private and government organizations. Although I don’t consider the release of PCAnywhere source to be particularly severe, I do question why Symantec chose to advise users to cease use of the product after the release and not before. And what makes it so sure that the product is now safe? In any case, the question of Norton Antivirus may be a little complex.

Targeting security products (whether that be an IDS, firewall, or AV product) is hot business these days, and vulnerabilities in antivirus engines can be extremely valuable to attackers if it means they are now able to slip an email attachment or drive-by download that would have otherwise been caught onto the target's system.

Anyone that has ever worked with, or had anything to do with, any kind of software product company will know that while names, logos, and even the interface for a product may change over time, the code behind it all will not necessarily follow suit. Even in circumstances where a “complete rewrite” has been done, they seldom ever are, and even in extreme cases we all know that a certain amount of CTRL-C/V action is going to go down somewhere along the way.

Note that although at the time that this was written the source code for Norton Antivirus does not appear to have been made public, we can safely assume that the stolen code has been shared privately, amongst a closed community associated with the individual responsible for the original heist.

So what does this all mean? Well, for the high-end adversary, probably not a whole lot as you’re likely to already have a copy of the source code. And it’s likely to be a much more recent version. On the other hand, folks who do not have a few hundred Gs laying around for bribing the employee of a software vendor so he’ll cut you a DVD full of source code are likely to see this as something of an opportunity. Through the use of not-uncommon analysis tools, figuring out which code segments are shared between the compromised source and possible modern derivatives thereof is a relatively trivial and inexpensive task.

While many groups who may do such a thing have probably put Symantec products under the microscope before, source-code analysis often opens up a whole, new world of subtle bugs in hard-to-reach regions of code that may have previously gone unnoticed. While the world's most well-funded and sophisticated actors are unlikely to find the release of source code particularly exciting, this may provide an excellent opportunity for less well-resourced groups involved in organized crime (such as botnet herders) and acts of industrial espionage to get one up on a product that has in the past spoiled the fun.

Tom Parker is Chief Technology Officer at FusionX.

Recommended Reading:
Editors' Choice
Kirsten Powell, Senior Manager for Security & Risk Management at Adobe
Joshua Goldfarb, Director of Product Management at F5