Slashdot Mirror
Snapchat Account Registration CAPTCHA Defeated
hypnosec writes "Snapchat's security troubles continue as a security researcher has managed to hack its account registration CAPTCHA system with a program of less than 100 lines that took 30 minutes to develop. Steve Hickson, a computer engineer by education, wrote a small computer program with very little effort that identifies Snapchat's ghost from the given set of images. Hickson equates Snapchat's ghost very particular and calls it a template that can be matched easily using a computer program. Hickson used a combination of Open Source Computer Vision Library (OpenCV), SURF points and FLANN matching "with a uniqueness test to determine that multiple keypoints in the training image weren't being singularly matched in the testing image.""
I need this code because half the time I can't figure out what the capture characters are myself.
So is there a way you could randomly seed an algorithm to generate a ghost with some noise in its drawing to throw off the vision processing? I realize the ghost is their logo but distorting it randomly could help thwart such an attack. Or am I missing something?
A site with barely-broadcastable body pictures that end up disappearing from it and yet still end up preserved on other parts of the web?
I say it's already like MySpace.
You can hold down the "B" button for continuous firing.
uses 3 well developed source libraries
...Mr. Hickson disappeared after 10 seconds.
never bring a twinkie to a food fight.
There are two problems with higher-order processing CAPTCHAs like that. One is the small problem set. A human at the website has to actually think of those connections between plugs and sockets, or umbrellas and rainstorms, or pizza and ovens, or hair and shampoo, etc. So the problem space is small. Then, blindly guessing answers still yields a decent success rate. Your particular example can be guessed with a success rate of 1 in 256.
Blurring a pair of words from a dictionary onto each other automatically generates millions of possible challenges, and random guessing won't work as well- at least some image analysis is needed.
My own idea for a CAPTCHA is to use images from Google Street View. Show random street view images of a bunch of houses, and ask, "what's the house number"? That would probably take a while to crack, long enough for me to dump my startup site's shares before all the porn gets leaked- if not for those assholes at Google interfering.
If the captcha is easy enough for humans, 1 in 256 random chance is fine for many applications. I've designed several very successful captcha systems used on thousands of sites. There are two reasons I say 1 / 256 is often fine.
First, let's consider one typical use case - blog spam. The spammer has a choice. He can spend this evening posting to 1,000 blogs with captchas, or the same amount of time post to 256,000 blogs without captchas. Which would you choose if you were a spammer? You choose the unprotected sites, of course. Sites without captchas get hundreds of times as much spam. Bad guys are by definition lazy, so they go after the low hanging fruit. Don't be low hanging fruit.
In other use cases, there may not be direct competition. Still, there's a cost / benefit analysis. Let's say it costs 1 penny of resources to register and use a snapchat account in a way the generates 12 cents in revenue. Multiply the cost by 256 and it's no longer profitable to abuse the service.
For most of our customers, the captcha is one part of a defense against brute force on the login screen. Assume that due to the other components of the system, you need 10,000 proxies to successfully brute force the login, because IPs banned after a dozen failed attempts. The captcha multiplies that by 256, so you now need over 2.5 MILLION proxies. I suspect that nobody has 2.5 million proxies to use. We have one of the largest lists of open proxies in the world, and even we don't have quite that many.