Building Backdoors and Fixing Malfeasance

You might have seen the recent news this week that there is an exploitable backdoor in Zyxel hardware that has been discovered and is being exploited. The backdoor admin account with the clever name ‘zyfwp’ is not something that has been present in the devices forever. The account was put in during firmware version 4.60, which was released in Q4 2020.

Zyxel is rushing to patch the devices and remove the backdoor account. Users are being advised to disable remote administration until the accounts can be deactivated and proven to be removed. However, the bigger question in my mind relates to the addition of the user account in the first place. Why would you knowingly install a backdoor?

Hello, Joshua

Backdoors are nothing new in the computer world. I’d argue the most famous backdoor account in the history of computer hacking belongs to Joshua, the dormant login for the War Operations Programmed Response (WOPR) computer system in the 1983 movie Wargames. Joshua was an old login for the creator to access the system outside of the military chain of command. When the developer was removed from the project the account was forgotten about until a kid discovered it and kicked off the plot of the movie.

Joshua tells us a lot about developers and their desire to have access to the system. I’ll admit I’ve been in the same boat before. I’ve created my own logins to systems with elevated access to get tasks accomplished. I’ve also notified the users and administrators of those systems about my account and let them deal with it as needed. Most were okay with it being there. Some were hesitant and required it to be disabled after my work was done. Either way, I was up front about what was going on.

Joshua and zyfwp are examples of what happens when those systems are installed outside of the knowledge of the operators. What would have happened if the team in the Netherlands hand’t found the account? What if Zyxel devices were getting hacked and networks breached without anyone knowing the vector? I’m sure the account showed up in all the admin dashboards, right?

Easter Egg Hunts

Do you remember the Windows 3.1 Bear? It was a hidden reference in the credits to the development team’s mascot. You had to jump through a hoop to find it by holding down a keystroke combination and clicking a specific square in the Windows logo. People loved finding those little nuggets in the software all the way up to Windows 98.

What changed? Turns out, as part of Microsoft’s Trustworth Computing Initiative in 2002 they removed all undocumented features and code that could cause these kinds of things. It also might have had something to do with the antitrust investigations into Microsoft in the 1990s and how undocumented features in Windows and Office might have given the company a competitive advantage. Whatever the reason, Microsoft has committed to removing undocumented code.

Easter eggs are fun to find but represent the bright side of the dark issue above. What happens when the easter egg in question isn’t a credit roll but an undocumented account? What if the keystroke doesn’t bring up a teddy bear but instead gives the current user account full admin access? You scoff at the possibility but there’s nothing stopping a developer from making that happen.

These issues are part of the reason why all code and features need to be documented. We need to know what’s going on in the program and how it could impact us. This means no backdoors. If there is a way to access the system aside from the controls built in already it needs to be known and be able to be disabled if necessary. If it can’t be disabled then the users need to be aware of that fact and make the choice to not use the software because of security issues.

If you’re following along closely, you should have picked up on the fact that this same logic applies to backdoors that have been mandated by the government too. The current slate of US Senators seem to believe that we need to create keys that allow end-to-end encryption to be weakened and readable by law enforcement. However, as stated by companies like Apple for years, if you create a key for a lock that should only ever be opened under special circumstances you have still created a weakness that can be unlocked. We’ve seen the tools used by intelligence agencies stolen and used to create malware unlike anything we’ve ever seen before. What do you think might happen if they get the backdoor keys to go through encrypted messaging systems?


Tom’s Take

I don’t run Zyxel equipment in my home or anywhere I used to work. But if I did there would be a pile of it in the dumpster after this mess. Having a backdoor is one thing. Purposely making one is another. And having that backdoor discovered and exploited by the Internet is an entirely differently conversation. The only way to be sure that you’ve fixed your backdoor problem is to not have one in the first place. Joshua and zyfwp are what we need to get away from, not what we need to work toward. Malfeasance only stops when you don’t do it in the first place.