How zombie LulzSec exposed privates' love lives with PHP hack
Single soldiers site swallowed surprise load
A dating website for US soldiers was hacked and its database leaked after it blindly trusted user-submitted files, according to an analysis by security firm Imperva. The report highlights the danger of handling documents uploaded to web apps.
"LulzSec Reborn" hacktivists attacked MilitarySingles.com and disclosed sensitive information on more than 170,000 lonely-heart privates in March this year. Hackers uploaded a PHP file that posed as a harmless text document and then commandeered the web server to cough up the contents of its user and a hashed password database.
Rob Rachwald, director of security strategy at Imperva, said the attack would have been blocked if MilitarySingles.com had filtered user-supplied content.
He added that a similar Remote File Inclusion-style vulnerabilities will exist in other sites that use PHP and actively solicit photos, video and so on.
Imperva reckons more than 90 per cent of the MilitarySingles.com passwords were cracked in nine hours thanks to extended dictionary-based rainbow lookup tables. MilitarySingles.com stored passwords as non-reversible hashes, rather than in plain text, however it did not salt the hashes, which would have made the process of recovering the passwords far more difficult. Insisting on hard-to-guess passwords isn't good enough unless developers pay attention to encryption best practices, said Rachwald.
The attack against MilitarySingles.com is the only notable assault by LulzSec Reborn. Imperva's analysis suggests the group has no more than six members, who set out to "embarrass the military". The crew is apparently "not as motivated" as the original LulzSec, according to Rachwald, adding that it has made little or no contribution to IRC chats and hacker forums.
MilitarySingles.com, which bills itself as the "dating website for single soldiers... and those interested in meeting them", is run by eSingles Inc.
Government and military personnel ought to have special policies regarding social networking to prevent their information from being easily accessed and manipulated. Rachwald told El Reg that an outright ban is likely to be flouted. Instead soldiers should be encouraged to use pseudonyms and particularly warned against disclosing their location, he said.
An analysis of the website breach was published in the May edition of Imperva's monthly Hacker Intelligence, which can be downloaded here. ®
Re: Und jetzt wie man es richtig macht
If only there were some sort of system which did this? NDS, AD, LDAP, etc.
It seems to be a trait in developers to develop solutions to problems which have already been solved, in a much better more functional way by specialists in the subject. I speak as someone who specifies code to be created by developers for a reporting product. It is incredible the amount of times that we get problems in the report and the developers' solution is to just hack some script together, rather than ask the subject matter expert for a proper programatic solution.
@The crew is apparently "not as motivated" as the original LulzSec, according to Rachwald, adding that it has made little or no contribution to IRC chats and hacker forums.
So in otherwords they are making an attempt not to get caught, unlike the other fools? Stay low, do your business, get out safely.
Und jetzt wie man es richtig macht
..some real advice on how to properly secure a password-based system:
A) Store all usernames and passwords on an entirely different machine and an entirely different database. The "credentials" server will only handle authentication requests through a well-defined (proper grammar), simple TCP interface. All other services including X11, ping and so on will be disabled on this server. The server will be completely firewalled except for that specific port.
Application code will query this server for authentication purposes via a TCP socket and will then proceed to do the usual SQL against the "app" database.
B) Store a "retry counter" along with the password hash and lock the account for half an hour after five bad attempts. Lock for a day after 15 wrong attempts.
Then even passwords such as "apple15" will be quite secure.