Slashdot Mirror
Criminals Distribute Infected USB Sticks In Parking Lot
New submitter sabri writes "The Dutch news-site Elsevier is reporting that cybercriminals attempted to steal data from a multinational chemicals company by 'losing' spyware-infected USB sticks on the company's parking lot. Their attempt failed as one of the employees who found the stick dropped it off at the company's IT department, who then found the spyware and issued a warning. So next time, don't expect to find someone's dirty pictures on a USB stick you just found..."
So you can load USB sticks you find and extract the pictures!
"Have you ever thought about just turning off the TV, sitting down with your kids, and hitting them?"
Or turn off auto-run in Windows. I once found a USB drive on the ground. Turns out it was some grad student's drive. I tried to return it but got no response from the email I found on his resume.
dd if=/dev/zero of=/dev/[usbdrive]
voila, free thumb drive, malware free.
Nobodies Prefect
Tidbits for Techs Technology Blog
I haven't heard of this technique actually being used in the wild, but it's enough of a threat to be included in the standard security training everyone has to take for at least a few Fortune 500 companies -- it's why some companies (and the U.S. military, I think) may disable USB ports. Trying to get at potential targets through standard attack vectors may not be effective, so if you have a financial backer this may present a promising attack vector that greedy targets may enable. The book "Security Engineering" cites this web site (had to find via archive.org) where a consulting company found out people inserted the USB sticks under slightly different circumstances: http://web.archive.org/web/20090621014856/http://www.vnunet.com/computing/news/2173365/uk-firms-naive-usb-stick
I ran one through the driver once
I say old chap, that's a bit rough, what? I hope you paid his medical expenses and gave him a shilling bonus after that experiment. Toodle pip...
This is a time-honored way of targeting a particular company. It sounds expensive, but if your motivation is commercial or governmental *coughcoughstux* it's extremely cheap compared to the alternatives (bribery, breaking-and-entering, rubber-hose cryptography). It's also a great way of finding out whether your own organization is aware of malware trouble; this technique is commonly used as part of security audits performed by companies hired to find out how good your company really is.
A company I worked for a few years ago hired a security auditing firm to check up on ourselves (only a few people were told, and we were told to keep quiet to ensure that our day-to-day practices were tested, not our "crap, someone's checking!" performance). They were unable to penetrate the network from the outside (including wirelessly) or socially engineer their way past reception or weasel out a password, but they got in via the USB-stick-in-the-parking-lot method. They told us afterwards that this is an extremely effective technique, as primate curiosity is almost unstoppable.
Everybody gets what the majority deserves.
If you think Linux has a magical immunity you might want to read how to write a Linux virus in 5 easy steps which shows with just a little social engineering its really not hard to target Linux just as the malware writers target Windows and OSX now.
From the article you mention:
A step that could be taken by the Gnome and KDE developers: Require launchers to have execute permissions. A saved attachment won't have those. Therefore, even though a syntactically correct and properly named launcher was dropped on the desktop a user can't just click on it and start it if the execute bit is not set.
Done. Modern versions of KDE need launcher to have execute permission. That hole is patched.
And nobody pretends that Linux has some magically imunity to viruses. As a Unix-like OS it just follows a few key principles :
- don't blindly execute everything. require executable to be explicitely marked as such (thus any shit downloaded from the web or from e-mail won't automatically be launchable).
- don't run constantly as root. thus the amount of harm that a program can do is limited to the access rights of a user. (While this still makes it possible to send spam, mine the data of the user, and modify the user profile, at least it prevents further deeper compromising of the running system).
That doesn't magically solve all malware problems in the universe. But at least it makes the life of malware writer a little bit more complicated. And the 5-step virus relies on a work-around of the first rule. Which has been since then corrected.
Back then, this no-brainer principles were NOT followed by Windows XP, making it even easier to write worms spreading over e-mail. Thankfully, since then Vista has arrived and has brought UAC dialogs in these situations (now how much dialogs can help security problems when the users are used to "okay" click on everything, that remains to be seen).
Or did you think android runs on Windows?
Android is a completely different beast and instead of unix-like userland it uses it's very own userland (a Java-like system).
Though it too doesn't allow execution of arbitrary e-mail attachment too. It's not impossible to write android malware, even malware that finds a way to look legitimate to android's capability system.
But at least the scenario "Here are some pics of hot lesbian teens! Click on the attachment to view them!" doesn't work on modern OSes. Except windows (and that's until WinXP, starting from Vista, you get an UAC dialog telling you that you run an executable from an untrusted source - now how many idiots will click on "okay" anyway is a different story).
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Many of them. All alone.
You looked at them all to 'make sure,' huh?
"First they came for the slanderers and i said nothing."
dud example
There are no examples, and the "5 easy steps" from the linked page haven't worked for years.
One of the reasons Linux is more secure is that the community responds far more quickly to potential threats.
Hairyfeet always gets to +5 with votes from the Apple/Windows crowd here, but he's never been able to show a single current instance of actual Linux malware in the wild. Much like the 235 patents, it's always threats from the future or the past.