Slashdot Mirror
Encrypted Messaging Startup Wickr Offers $100K Bug Bounty
alphadogg writes "Two-year-old startup Wickr is offering a reward of up to $100,000 to anyone who can find a serious vulnerability in its mobile encrypted messaging application, which is designed to thwart spying by hackers and governments. The reward puts the small company in the same league as Google, Facebook and Microsoft, all of which offer substantial payouts to security researchers for finding dangerous bugs that could compromise their users' data. Wickr has already closely vetted its application so the challenge could be tough. Veracode, an application security testing company, and Stroz Friedberg, a computer forensics firm, have reviewed the software, in addition to independent security researchers."
You'll get better regulation from this than from anything that could possibly be concocted by government bureaucrats.
Note: This requires the real threat of economic loss, so an organization that can demand payment regardless of its performance—i.e., the government—cannot implement something similar.
That's a nice publicity stunt though
Wouldn't it be funny if the NSA came forward and claimed the prize money many times until the company went under? Because surely they have backdoors all over the place to walk right through these guys' security measures.
"A door is what a dog is perpetually on the wrong side of" - Ogden Nash
I'd bet its susceptible to:
The phone you run it on is tracked, and the company that does so shares that data.
Timing attacks: if you send data at some time, and someone else gets a message then, that implies you communicated with them.
Visual surveillance. Camera sees you type, camera sees your message.
They claim "sender-based control over who can read messages, where and for how long". This is impossible. If the receiver can see the message, they can record it.
Boarder patrol requesting access.
Torturing you as an "enemy combatant"
And some likely others:
How do they handle key distribution? If you setup communication with someone via email, text or whatever, that can be compromised before you even start.
Looking through the tech they claim to be using, it seems like they lack defenses against Rubber-hose cryptanalysis. Is there any effort in the area of deniable encryption, or maintaining plausible deniability about having messages or particular contacts? I suspect not.
Its rather impractically expensive to provide sufficient random cover traffic on a phone to blind against timing correlation attacks on video messages. Given that we know the cell networks are heavily watched, even if the messages were routed through Tor that wouldn't be enough to reliably disassociate sender and receiver (You would want the ageing options planned for I2P for that). Then just get a warrant, and compel them to disclose the contacts and any pending messages. There are [partial] defenses that can be employed here (like TrueCrypt does with hidden volumes for example), its not unsolvable, just often ignored.
Security is hard. Security against a large scale threat such as governments is very hard. Securing the message contents is easy, securing that there was a message is the real challenge.
All that said, it looks like they likely do a pretty good job of making end to end encryption accessible. While thats not all one might want, its more than most of us get, so its still a good thing. Its progress, not a solution.
A serious vulnerability? The people using it of course, always the most serious vulnerability
"goodbye and hello, as always" ~Prince Corwin, from Zelazny's Amber series
It's 2014, after all.
What other vulnerability do you need ?
If they wanted to impress, they would have figured out how to implement this without any servers. Think big people, Skynet did.
I support the sentiment of these guys but your code is going to be running on a platform that is largely exploitable by most English speaking foreign governments and possibly well funded crooks.
What this means is that no matter how good your software is it will be ultimately rendered useless by going after the host platform and memory.
Also anything that uses a public key exchange is only secure because certain reversals of transformation are 'hard'. There is no universality to hard, what is hard for me may not be hard for you.. Globally governments and crooks seek out and employ people who are good at working with hard.
Then there are all the other sources of issue, like suitable entropy, which is not to be scoffed if something is 20% less random than is should be then that is a huge advantage.
However most of the above is a bit unfair though because they will not be in a position to do much about it but it does need to be considered by the users though.
Reading ../en/myapp.php in the URL of the official website (plus things like "military-grade encryption" etc.) makes me think this might be a worthwhile challenge.. But maybe I'm just again too prejudicial
Would the 100k cover lawyers expenses if you used this method?
XKCD:Security
I only look human.
My mother is a halfling and my dad is an ogre, so that makes me an Ogreling
... 5K please.
I guess it wouldn't count to run their app on a rooted phone that presents compromised APIs to the apps? Or crack it open, inject logging code, repackage and resign it then submit it to a third-party marketplace? That is to say, the standard security problems all apps face as opposed to a flaw specific to the Wickr app?
why not hire some QA to do stuff like that full time and they also get some in house beta testing that is not just the coders testing there own code.
Yeah, because the average QA is a master of cryptography. You need to hire security specialists for this....and they did.
Now, after all of that, they want to make sure nothing slipped.
Usually this sort of thing is a SCAM, offering crappy security and a contest that lacks key parts necessary for fair testing. For example, I've seen many a contest where only an encrypted message is provided. You have to provide the original plaintext. Of course, there are a great many possibly valid plaintexts, not infinite but close enough for practical purposes, and no way to distinguish which is the correct one.
Now this project is talking about security software. Software, not Hardware. Running over an unsecured data channel on emulated hardware if you so wish. Clearly it can be compromised for any individual system. It may require running secondary software to spy on the original software, or compromising the device (or kernel). But it can be done. Whether it can be done wholesale is another question.
It doesn't actually say that the keys are generated on the device.
1. American company
2. Software is not open source
3. Reliance on the company's servers, not peer to peer
4. I can't run my own server
5. Susceptible to traffic analysis
6. Runs on a mobile platform I cannot fully control
What more do you want?
It is no point to discuss details like buffer overflows when the whole premise is, quoting Doge: such flaw, much hot air.
The whole thing revolves around an illusion of trust. The company wants to create an image of trustworthiness so that you trust their ability to offer privacy. That's not how it works, guys. Privacy has to be provided via proven mathematical methods and good technical practises, not through a fluffy soft idea like trust.
A competing app did just that, and a guy from Russia won the $100k. Now they're offering $200k.
Still no article on Telegram in /.