A neat piece of coding indeed, it targets vulnerabilities in all 3 operating systems – the sad thing? The malware itself is vulnerable to a basic directory traversal exploit, which means rival gangs can actually commandeer the infected targets.

They went to lengths to keep it secure and unseen (encrypted communications etc) – but didn’t program the malware itself securely…

From the department of cosmic justice comes this gem, spotted by researchers from Symantec: a trojan that targets Windows, Mac, and Linux computers contains gaping security vulnerabilities that allow rival criminal gangs to commandeer the infected machines.

Known as Trojan.Jnanabot, or alternately as OSX/Koobface.A or trojan.osx.boonana.a, the bot made waves in October when researchers discovered its Java-based makeup allowed it to attack Mac and Linux machines, not just Windows PCs as is the case with most malware. Once installed, the trojan components are stored in an invisible folder and use strong encryption to keep communications private.

The bot can force its host to take instructions through internet relay chat, perform DDoS attacks, and post fraudulent messages to the victim’s Facebook account, among other things.

Now, Symantec researchers have uncovered weaknesses in the bot’s peer-to-peer functionality that allow rival criminals to remotely steal or plant files on the victim’s hard drive. That means the unknown gang that took the trouble to spread the infection in the first place risks having their botnet stolen from under their noses.

“Even though it’s encrypted and even though it was written in Java to make it cross-platform, it was still vulnerable to basically a directory transversal exploit,” Dean Turner, director of Symantec’s Global Intelligence Network, told The Reg. “From a technical perspective, it goes to show that even if you have all those things where you’re building in a secure platform, if you’re not building application security into your malware, other bad guys will probably take advantage of it.”

It’s somewhat of an odd decision though, in terms of numbers obviously Windows machines far outnumber Linux and OSX desktop installations. On the web-server front perhaps Linux is a valuable target – but on consumer desktops? Is it really worth the effort for malware creators to make cross-platform trojans? Personally I don’t think it is, maybe it was just an experiment.

The number of Apple machines is certainly growing, the next big market we are going to see is tablets and smartphones I believe. I’d be on the lookout for more iOS and Android worms/trojans in coming months.

A self-replicating stealthy Android trojan with a previously unpatched zero-day remote root exploit could be devastating.

Jnanabot’s P2P feature is designed to make botnets harder to take down by providing multiple channels of communication. After sending an infected machine a single GET request, a website can discover all the information needed to upload any file to any location on the host’s file system. Attackers can then install a simple backdoor on a user’s machine by, for instance, writing a malicious program to a computer’s startup directory.

Attackers can use the same vulnerability to steal files on infected machines.

Turner said the number of Jnanabot infections so far is “measured in the thousands,” rather than the hundreds of thousands for some of the better-known trojans. Still, infection statistics gathered by Symantec in December are surprising. They show that about 16 per cent of infections hit Macs. They didn’t show any infections on Linux machines. Turner said that Jnanabot attacks on the open source platform weren’t able to survive a reboot.

The bot was discovered spreading over Facebook posts that planted the following message on infected users’ Facebook pages: “As you are on my friends list I thought I would let you know I have decided to end my life.” An included link leads recipients to a cross-platform JAR, or Java Archive file that can run on Windows, Mac, or Linux. Once the recipient is infected, his Facebook page carries the same dire warning.

It seems like the trojan theoretically can attack Linux, but so far hasn’t been seen in the wild and it can’t survive a reboot. Not that it really matters as from my experience most Linux users never reboot anyway except for kernel upgrades (which isn’t that often).

Perhaps it just doesn’t work that well on Linux, or Linux users don’t believe in installing JVM – it doesn’t usually come standard with OS installs as it’s considered non-free software.

The chosen vector for replication seems to be Facebook and a rather dramatic faux-suicide note – which sadly I think will be very effective.

Source: The Register

There have been quite a few security concerns with Facebook, especially with the amount of personal information it collects on it’s users.

Of course there is Koobface and it’s many variants which have been propagating all kinds of spam through Facebook wall posts and messages.

I’m glad someone is offering a solution for free, yes they benefit from it too by being able to gather data on Facebook activity and the quantity of malicious posts occurring on Facebook.

Security vendor Websense if offering Facebook users and businesses a new free ‘firewall’ service that monitors their pages for malicious posts, links and spam.

Defensio 2.0 checks all posts to Facebook in real time against Websense’s ThreatSeeker Network, a database of problem URLs, before deciding whether to categorise a post as malicious or unwanted. This also draws from data gathered by US ISP Radialpoint and URL shortening service bit.ly before performing further heuristic analysis as a final check.

If a bad post is detected, the system logs and informs the user who makes the final decision. As with the original Defensio system – acquired a year ago when Websense bought the company of the same name – it can also monitor web pages for rogue posting, pre-emptively blocking those it deems unwanted.

“We are seeing real threats to Facebook such as Koobface,” said Websense senior research manager, Carl Leonard.

It seems to work on a ‘moderation’ model so if the software detects any suspicious automated messages/links or other dodgy activity it will block the post/message and allow the user to approve/deny the request.

But then it’s only going to be effective if take-up is good amongst the non-tech savvy users where the problems tend to be a lot more common.

Sadly this seems highly unlikely as only people who read sites like this will know about it, unless it get’s heavily promoted on Facebook..but then you have to contend with ad-blindness problems.

According to Leonard, an advantage of Web 2.0 monitoring was that it gave security companies a way of following criminals inside the otherwise closed world of social media, something that many security vendors can’t yet do. “We can have visibility into threats on these social networks, and have a fantastic feed of information that can benefit all our customers,” he said.

Leonard was not able to say when or if the monitoring might be available other social media sites or feeds such as twitter, where rogue behaviour can be difficult to spot.

The service is free for anyone with fewer than 50,000 posts per month, and for companies with 15 employees of less. For professional sites or sites with larger volumes of posts, the service starts at $5 (£3) per month, per site.

It’s free for most people, I’d imagine very few companies are making 1500 posts per day! Even if you need to pay it’s pretty cheap.

I hope to see more initiatives from companies like this, and ideally someone working with Facebook themselves to increase pro-active security measures on the site.

Obviously that’s not their first priority and with the recent brouhaha about their new privacy terms and default settings..you should be concerned about what information of yours they intend to utilise.

Source: Network World

A new variation of Koobface has popped up aggressively on Facebook and is attempting to steal login credentials for other social networking sites.

It’s heavy on the user of Social Engineering by using the old familiarity ploy and trying to lure users into execution the spurious malware.

Researchers at Trend Micro report that a new variant of the Koobface worm is squiggling through Facebook, infecting users and attempting to steal cookies with log-in information for sites such as MySpace.com, MyYearbook.com, Bebo and Hi5 Networks. The Koobface worm first appeared in 2008.

Researchers at Trend Micro are reporting that a new variant of the Koobface worm is spreading on Facebook.

Koobface first appeared in 2008, with separate variants striking members of Facebook and MySpace.com. Now the Koobface worm is back again, with an eye toward stealing cookies for other social networking sites.

According to Trend Micro, the new variant sends Facebook messages claiming to be from a friend. The messages link to a spoofed YouTube video. In an interesting social engineering ploy, the malicious landing page not only displays the friend’s name, but also a picture pulled from the person’s Facebook profile.

It’s a pretty standard modus operandi for malware distributors, a new version of Flash seems to be a very common theme amongst these kinds of worm.

It seems like quite a lot of thought has gone into the design of this worm and due to the pretty slick distribution method it might become quite a large scale infection. It’s endgame is to create some kind of botnet and to steal the real information of value (banking details, credit card numbers etc).

The page prompts the user to install a new version of Adobe Flash. Users who agree are redirected to a download site for the file setup.exe, which is the new Koobface variant. Trend Micro detects the worm as WORM_KOOBFACE.AZ, and reported March 1 that its researchers had seen more than 300 unique IP addresses hosting the .exe file.

Trend Micro is expecting to see more.

“We’re only flagging a few hits at the moment, but the complexity with which this threat has been created shows how much work has been done to social-engineer social networks with the end game of creating [botlike] accounts to send out third-party links to almost anything,” said Jamz Yaneza, a threat researcher at Trend Micro.

The latest iteration of the worm runs on Windows 98, ME, NT, 2000 and XP and Server 2003. It sends and receives information by connecting to several servers, allowing hackers to remotely execute commands on a compromised machine.

Once infected the malware will search the users machine for Cookies from other social networking sites such as MySpace.com, Hi5 Networks, MyYearbook.com and Bebo.

As always, warn whoever you know that might use Facebook and generally tell people not to install or run anything that they aren’t expecting or didn’t specifically look for.

If they want to update Flash tell them ONLY to ever do it at the Adobe site.

Source: eWeek