A New York-based web developer and his colleagues have built a web-based application for people to see if their LinkedIn password hash is among 6.5 million released on a Russian hacker forum.
The password breach, revealed yesterday, is significant due to the detailed personal data stored by LinkedIn and the chance for hackers to spear phish high-level executives or spread malicious links.
LinkedIn is telling some users to reset their passwords, but there is another way for users to see if their account was compromised.
LeakedIn then checks to see if the hash is on the list of breached passwords. Not all of the hashes in the list have been converted to original passwords yet, but it is likely hackers are working on it. Shiflett said: "I discovered that my password was not only one of the 6.5 million that had been leaked, it was also among those that had been cracked. I was a victim."
Password hashes can be converted to plain-text by using powerful graphics processors and free password cracking tools such as "John the Ripper," which can be used with a regular PC, and "oclHashcat." How long that process takes depends on the passwords' complexity.
Those cracking applications use word lists compiled from other password breaches in so-called dictionary attacks, which seek to match already computed hashes with those on the new list. Another method is a brute-force attack in which the programs rapidly try different password combinations in the hope of finding a matching hash. Brute-force attacks are more time consuming for longer passwords that contain a mix of upper- and lower-case letters, digits and symbols.
Robert David Graham, CEO of the security consultancy Errata Security, said that each letter of a password has 100 possible combinations composed of either upper or lower case, digits or symbols. A five-character password would have 10 billion possible combinations and could be cracked in 5 seconds using a top-of-the-line Radeon HD 7970 graphics processor.
A six-character password would take around 500 seconds, but a seven-character password would take 13 hours, Graham wrote. Eight characters pushes the time up to 57 days, with a nine-character password taking up to 15 years.
"In other words, if your password was seven letters, the hacker has already cracked it, but if it's nine letters, it's too difficult to crack with brute force," Graham said.
Many of the hashes in the dump have five zeros as the first five characters of the hash. Graham said that some people "think that this means that the hacker has already cracked any passwords that have been zeroed out this way."
LinkedIn did not "salt" its hashes, which involves inserting random characters into the hash that make it more difficult for people trying a brute-force attack. The company said it is now salting hashes.
Security vendor Sophos said it determined there were 5.8 million unique hashes out of the 6.5 million released after duplicates were eliminated. Of those 5.8 million, some 3.5 million hashes or about 60% had been successfully brute forced, said senior security advisor Chester Wisniewski.
Sophos compared the passwords used for LinkedIn with those used by the Conficker worm to spread through network drives. All but two of the simple passwords used by Conficker were also used by LinkedIn users, Wisniewski said.
LinkedIn uses a person's email address as part of its sign-in process, and it's not known if the hackers also have those addresses, which would make the breach even more severe since it would allow them to directly access a person's account. LinkedIn will have to release more information in order to restore the confidence of its users, said Cameron Camp, a security researcher with the security company ESET in San Diego.
"It will be very interesting to see in the next two to three days to see what LinkedIn says," Camp said.