Security watchers are counting down to a change in how the infamous Conficker (Downadup) worm updates malicious code, due to kick in on Wednesday 1 April.

Starting on 1 April, Windows PCs infected by the latest variant of the Conficker worm (Conficker-C) will start attempting to contact a sample of 50,000 pre-programmed potential call-home web servers from which they might receive updates, a massive increase on the 250 potential web server locales used by earlier variants of the code.

"Conficker-C isn't going to contact all 50,000 domains per day," explained Niall Fitzgibbon, a malware analyst at Sophos. "It's only going to contact a randomly-chosen 500 of them which gives each infected machine a very small chance of success if the authors register only one domain. However, the P2P system of Conficker can be used to push digitally signed updates out to other infected machines that don't manage to contact the domain."

Whether anything will actually be offered for download, much less what the payload might be, is unclear. No particular function or payload currently within the malicious code is due to activate on 1 April. It's also possible that a payload will only be offered up for download days or week after the new call-home routine comes into effect.

If updates are successfully made, infected machines are programmed to suspend call-home activity for 72 hours, as an analysis by Sophos explains.

Lessons from the call back routines of previous variants of the worm provide few clues as to what might happen. Sophos said it never observed the previous Conficker-B variant ever downloading malicious payloads, other than updates to Conficker-B++ and Conficker-C. As a result, there isn't much history to draw upon for any speculation as to the eventual goal of the Conficker botnet.

Anti-virus firms are keeping a close eye on what Conficker might do early next month while downplaying concerns that Downadup will either "erupt" or "explode" on 1 April, deluging us with spam or swamping websites with junk traffic in the process.

"Let's not forget that history has shown us that focusing on a specific date for an impending malware attack has sometimes lead to nothing more than a damp squib," notes Graham Cluley, senior technology consultant, at Sophos.


Although nothing might happen it's never a bad time for sys admins to check for infection by Conficker on their network. Such infections have already caused widespread problems.

Symantec said that the worm, which had initially focused solely on spreading "has since developed into a robust botnet, complete with sophisticated code signing to protect update mechanisms, as well as a resilient peer-to-peer protocol".

Windows PCs infected with Conficker (Downadup) are programmed to dial home for updates through a list of pseudo-random domains. Microsoft is heading a group, dubbed the anti-cabal alliance, to block unregistered domains on this list. The more complex call-home routine deployed by Conficker-C comes in apparent response to this move.

Rik Ferguson, a security researcher at Trend Micro, added that blocking call-back domains associated with the latest variant of the worm will be "almost impossible" not only because of the daily volume, but also because there is a possibility that legitimate domains might be hit as a result. Even earlier versions of the worm, calling far fewer domains every day, used algorithms that threw up addresses that coincided with legitimate domains.

Birth of a superworm

Variants of the Conficker worm, which first appeared back in November, spread using a variety of tricks. All strains of the superworm exploit a vulnerability in the Microsoft Windows server service (MS08-067) patched by Redmond in October.

Once it infects one machine on a network, the worm spreads across network shares. Infection can also spread via contaminated USB sticks. This combined approach, in particular the worm's attempts to hammer across corporate LANs, have made Conficker the biggest malware problem for years, since the default activation of the Windows firewall put the brakes on previous network worms such as Nimda and Sasser.

Compromised Windows PCs, however the infection happens, become drones in a botnet, which is yet to be activated. It's unclear who created or now controls this huge resource.

Estimates of the number of machines infected by Conficker vary, from barely over a million to 12 or even 15 million. More reliable estimated suggest that between 3-4 million compromised systems at any one time might be closer to the mark.

SRI reckons that Conficker-A has infected 4.7m Windows PC over its lifetime, while Conficker-B has hit 6.7m IP addresses. These figures, as with other estimates, come from an analysis of call-backs made to pre-programmed update sites. Infected hosts get identified and cleaned up all the time, as new machines are created. Factoring this factor into account the botnet controlled by Conficker-A and Conficker-B respectively is reckoned to be around 1m and 3m hosts, respectively, about a third of the raw estimate.

Estimates of how many machines are infected by the Conficker-C variant are even harder to come by.

But however you slice and dice the figure its clear that the zombie network created by Conficker dwarfs the undead army created by the infamous Storm worm, which reached a comparatively lowly 1 million at its peak in September 2007. Activation of this resource may not come next week or even next month but the zombie army established by the malware nonetheless hangs over internet security like a latter-day Sword of Damocles.

Some security watchers are sure it will get used eventually, if not on 1 April. Sam Masiello, a security analyst at MX Logic, said: "Why go through all of this effort to create such a huge botnet then not utilize it for something?"

Security Update Arrives Weeks After Notice Issued

Adobe has released a security update to address a flaw the company first warned users of in February.

The company said that the update should patch a flaw in Reader 9 and Acrobat 9 which could allow an attacker to use a specially-crafted PDF file for cause a crash and take control of a targeted system.

The security fix will update both the Mac and Windows versions of Adobe Reader and Acrobat to version 9.1. The company is planning to release fixes for the Unix version of the software as well as earlier versions of both applications later in the month.

Along with Adobe, security experts from the US Computer Emergency Response Team and Sans are recommending that users update to the 9.1 versions of the software if at all possible.

The update comes more than two weeks after Adobe first warned of the threat, which has been actively exploited in the wild. At the time, the company estimated that the first patches for the flaw would not be out until March and users were advised to disable Javascript code within PDF files.

However, just days after Adobe released its advisory on the attacks and a timeline for a fix, an independent researcher constructed a home-made patch for Windows systems.

A company worth billions of dollars which is supposed to have the best programmers, the kind of company that won’t leave any security wholes in the system. Yahoo! system that is!

XSS bugs are already yesterday’s news when we talk about Yahoo! They are all over the place on the *.yahoo.com subdomains.But we are not talking here about minor XSS bugs. We mean serious business. We are talking about the kind of security which exposes the privacy of millions of users of the “trustable” Yahoo! services.

We are talking about SQL Injection. One of the worst kinds of security breach.

Here you have one of the pages vulnerable to SQL Injection:

http://in.jagran.yahoo.com/article/index.php?choice=homepage_getnews&state=1&city=87%20union%20all%20select%201,concat_ws(0x203a20,version(),user()),3,4,
5,6,7,8,9,10,11,12,13,14,15,16,17,18

What do we find here? Information about the SQL server, its version and the current user SQL user:



A list with SQL users and passwords:


And of course, much more information available at the hand of an attacker.

Moreover, this SQL Injection can be used as an XSS, especially for session hijacking:



The sad part is that Yahoo! didn’t adopt any policy whatsoever regarding this kind of problems. They dont admit they have a problem, nor do they give any credits to those who find them.

Following in the footsteps of other sites, Yahoo! could learn to gain from this. Vast majority of those who find bugs don’t disclose them anymore precisely for the fact that Yahoo! is in total denial. By coming out clean, Yahoo! would also reduce the amount of hacked/stolen accounts and other shameful security breaches like the one we present here.

Google confessed to a serious bug in its Docs sharing system over the weekend, but downplayed the security cockup by claiming only a tiny number of users had been affected.

The internet search kingpin said that less than 0.05 per cent of Google Docs accounts were hit by a privacy breach after documents were shared “inadvertently” with other users.

Mountain View said in a blog post, penned by Docs product manager Jennifer Mazzon, that the security lapse was “limited to people with whom the document owner, or a collaborator with sharing rights, had previously shared a document.”

She claimed that very “few users” would have been affected by the bug “because it only could have occurred for a very small percentage of documents, and for those documents only when a specific sequence of user actions took place.”

Google said the error was limited to its Docs system within Google Apps and did not affect its spreadsheet system, though some presentations were also hit by the error.

The company fixed the bug by using what it described as an “automated process to remove collaborators and viewers from the documents” that had been exposed to the security glitch.

In other words it stripped all sharing privileges from the documents affected by the bug and then informed affected users that they would have to manually re-share their documents.



“We're sorry for the trouble this has caused. We understand our users' concerns (in fact, we were affected by this bug ourselves) and we're treating this very seriously,” said Mazzon.

Google has recently been attempting to woo businesses away from desktop-based Office suites in favour of adopting the company's cloud-based Apps system.

In January Google confirmed it had inked deals with IT resellers to sell its online applications to biz customers. From the end of this month authorised resellers will be able to flog, customise and support premium versions of Google Apps.

However, this latest bug could lead some businesses to conclude that pushing their personal information up into the clouds simply poses too big a security risk.

Source: The Register