We’ve talked in depth about garage doors and their vulnerabilities on ITS, but today we wanted to highlight a discovery made by Samy Kamkar using a children’s toy and some common materials. The device he’s created, dubbed the OpenSesame, can open a garage door using a brute force attack in less than 10 seconds. Before we get into the details of the vulnerability, let’s explain a bit about how automatic garage door openers developed. The first electric garage door opener was introduced in 1926, but didn’t gain in popularity until after World War II. These openers usually involved a wired switch being run from the door motor to a keypad or button that could be pressed from inside the vehicle. As technology improved, the wireless remote was created and used radio signals to transmit a code from the remote control to the opener itself. Once the code was transmitted, the opener would receive it and run the motor to draw the door up or down. In the 1960’s, as automatic openers were more widely adopted, it was discovered that the doors all used the same code.
Since the doors used the same signal and code, any remote could open any door. Thieves quickly discovered that by purchasing a few remotes from different manufacturers, they could open nearly any door. This led the garage door industry to introduce new openers that featured changeable codes. Manufacturers wanted to design an opener with a code that could be set by the owner in case they needed to replace or add new remotes. These new programmable openers featured a series of 8-12 dip switches that could be set in the up or down position to create a unique code. This meant that for an 8 switch remote there were 28 or 256 possible codes and on a 12 switch remote there were 212 or 4,096 possible codes. This greatly decreased the chances of the door’s code being matched by a thief armed with a standard remote, as they would need to sit outside and physically set each dip switch and test the new code. This might sound like a large number of possible codes but in binary terms, it’s not much.
In fact, using a two character alphanumeric password would be more secure than this and would provide more combinations. The method that Samy Kamkar used to attack these type of garage doors was a brute force attack, in which he sent every possible code (4,096) to the door until it would open. The device he used to perform this brute force attack was a slightly modified IM-ME children’s toy. He found that if he transmitted each code five times with a wait period behind the code, he could transmit every possible code to the door in about 29 minutes. That’s a pretty scary fact as theoretically, a thief could sit in a car outside your home for that long without being noticed. Unfortunately, that’s not the worst of it as Samy discovered that he could dramatically shorten the length of time it took to open the door using math. Samy found that by transmitting each code only once rather than five times, the time was reduced to 6 minutes. Many times, remotes transmit the code multiple times in case of interference, but there’s rarely ever any, so transmitting it once seems to work just fine.
Not only that, he also discovered that removing the wait times between the codes took the time down to 3 minutes. So if there wasn’t a wait time between codes, how could the opener know when one code stopped and another began? The answer lies in the fact that the openers use what’s called a Bit Shift Register, where it’s only looking for a part of the code to match up with the actual code. This is a very insecure method to check the code and results in a dramatic decrease in the overall time it takes to send all the codes to the opener since the opener is checking all possible permutations of a code and doesn’t have a defined start and end point. Knowing that the opener used a Bit Shift Register, Kamkar was able to apply The De Brujin Sequence to transmit all possible codes much quicker and managed to take the total time down to just 8.7 seconds. Luckily, most automatic openers today aren’t affected by this exploit since they use “Rolling” codes, where the code is changed every time the door is opened.
However, there are some manufacturers that are still producing openers with dip switches. These manufacturers include Nortek / Linear / Multi-Code and NSCD/North Shore Commercial Door. Older models from vendors such as Chamberlain and Liftmaster can also be vulnerable so you should double check to ensure that your door does not feature this technology. The best method to check your opener is to check the remote. If you open the remote and find dip switches, you should upgrade the opening system immediately. Look for a system that offers Rolling Codes, Hopping Codes, Intellicode or Security Plus. The remote above is a good example of a Rolling Code remote because it lacks dip switches. These technologies don’t rely on a fixed code and are harder to hack. Keep in mind though that nothing is foolproof, so you should design your security around that. Top 10 Garage Door Security Tips to Prevent Break-Ins Safety Announcement: Protecting Against Garage Door Break-Ins How to Escape from Zip Ties
Use These Tips to Protect You and Your Family Against Home Invasions Beat The Clock: Most Burglaries Take Place in Minutes What To Do in the Crucial Three Hours after a BurglaryAmericans’ garages, those sacred suburban havens of automobiles and expensive tools, are probably more important to us than many of our online accounts. But some garages are only protected by a code whose security is equivalent to a two-character password. And security researcher Samy Kamkar can crack that laughable safeguard in seconds, with little more than a hacked child’s toy. On Thursday, Kamkar revealed a new tool he’s created called OpenSesame, which he says can open any garage door that uses an insecure “fixed code” system for its wireless communication with a remote. Built from a discontinued Mattel toy called the IM-ME, altered with a cheap antennae and an open source hardware attachment, Kamkar’s less-than-$100 device can try every possible combination for these garage doors and open them in seconds.
“It’s a huge joke,” says Kamkar, a serial hacker who works as an independent developer and consultant. “The worst case scenario is that if someone wants to break into your garage, they can use a device you wouldn’t even notice in their pocket, and within seconds the garage door is open.” Before barricading or booby-trapping your garage against OpenSesame intruders, it’s important to note Kamkar’s exploit doesn’t work against just any garage door—only ones that respond to a “fixed code” wirelessly transmitted by a remote instead of a more secure “rolling code” that changes with every button press. And it’s not clear just how many garage doors actually use that fixed code system. Kamkar found that his own garage door, in a newly built Los Angeles condo, was vulnerable to his attack, though he couldn’t identify device’s manufacturer; the receiver in his building was hidden. When he checked the attack against two friends’ garage door openers—both made by a company called Linear owned by the parent company Nortek—it worked both times.
Nortek didn’t immediately respond to WIRED’s request for comment. Another major brand of garage door opener, Genie, didn’t respond to to a request for comment either, but says on its website that its devices use rolling codes. A spokesperson for Chamberlain, the owner of the Liftmaster brand and one of the biggest sellers of garage door openers, initially told WIRED the company hasn’t sold fixed code doors since 1992. But when Kamkar dug up a 2007 manual for a Liftmaster device that seemed to use fixed codes, Chamberlain marketing executive Corey Sorice added that the company has supported and serviced older garage door openers until much more recently. “To the extent there are still operators in the market begin serviced by replacement parts, part of the objective is to get to safer and more secure products,” he said in a phone interview, using the industry term “operator” to mean a garage door opener. “We’d love to see people check the safety and security of their operators and move forward.”
Kamkar has posted his own video to help people determine if their garage door is vulnerable or not. To attack fixed code garage door openers, criminals have for years used “code grabbers” that capture the code from a user’s garage door button press and replay it later to open the door. But for these vulnerable systems, Kamkar has reduced the time necessary so that it’s become practical try every possible wireless code. That means someone could walk or drive through a neighborhood, going door-to-door and trying the device until one of the vulnerable garages opens. “For code grabbers, you have to sit there and wait for the person to hit the button,” says Kamkar. “For this, [the victim] never even has to be there.” To perform his brute-force attack, Kamkar used a pre-smartphone toy called a Radica IM-ME. That chunky pink handheld device for wireless text messaging, once sold by Mattel, has been adopted by radio hackers because it’s capable of broadcasting and receiving at a broad range of frequencies.
Kamkar added his own antenna to the IM-ME and used GoodFET, a tool built by well-known radio hacker Travis Goodspeed, to reprogram the IM-ME with his cracking program. The fixed-code garage door remotes Kamkar tested use at most 12 bit codes—that’s 4,096 possibilities. In modern computer security terms, that’s a trivial level of security: Kamkar calculates that a password with just two characters offers at least 5,184 possibilities. “Imagine if your bank only let you have a two character password,” Kamkar says. Using a straightforward cracking technique, it still would have taken Kamkar’s program 29 minutes to try every possible code. But Kamkar improved his attack by taking out wait periods between code guesses, removing redundant transmissions, and finally using a clever optimization that transmitted overlapped codes, what’s known as a De Bruijn sequence. With all those tweaks, he was able to reduce the attack time from 1,771 seconds to a mere eight seconds.
Even so, that eight-second attack only works for a single frequency; Kamkar says he’s found four frequencies different for vulnerable garage doors he’s tested, and OpenSesame can cycle through its brute-force attack on all four frequencies in less than a minute. Kamkar has detailed OpenSesame’s attack on his website, and also published the tool’s code. But he intends it to serve as a warning, not a how-to manual. In fact, he says he’s even disabled the code so that criminals can’t use it, and wouldn’t comment on exactly how he’s crippled his exploit. That’s a rare move for Kamkar, and one that demonstrates how dangerous he believes his garage attack may be. OpenSesame is just the latest in a long string of high-profile hacks from Kamkar, who gained fame in 2007 when he launched a MySpace worm—what came to be known as the Samy worm—that added more than a million friends to his account in an hour. He’s also built a drone designed to seek out and wirelessly hijack other drones, and a 3-D printed robot that can crack Masterlock combination locks in seconds.