CrypTweet was released yesterday but already at least two security and privacy experts have concluded it has major security holes.
Mark Pesce, CrypTweet’s author, wrote on the application’s web site he was aware “a number” of people had examined its code and found security problems. “I am waiting for an analysis of these flaws – hopefully they can be remediated – but until then CrypTweet should be considered an attempt toward security that may not be secure,” he said.
Pesce posted the caveat after a mild admonishment on Twitter from security expert Moxie Marlinspike: “Whoa, just looked at the CrypTweet source. This is not a secure protocol, you should be careful about claiming it is.”
Security and privacy researcher Christopher Soghoian was less kind: “I hate to sound elitist. But if you have no experience with crypto code, please don’t release your first project to users. It is dangerous.”
In an interview, Pesce said he released the application in order to get feedback, and it wasn’t intended to be a complete end product.
“I really need to have people look over my code. I am a neophyte when it comes to cryptography,” said Pesce.
CrypTweet is composed of several Python applications which use RSA public-key cryptography in order to encrypt direct messages on Twitter, which are person-to-person private messages. Those messages typically are not encrypted, meaning the communications could be read by an attacker who intercepted the messages.
CrypTweet was built for the Windows 7, OSX and Linux desktop operating systems. A version for Android is under development, and Pesce wrote that it will run on a Nokia N9 running Meego Linux. It will also work on a jailbroken iPhone or iPad as long as the device is running a newer version of Python than in the Cydia package manager, Pesce wrote.