Christmas Hackers 2010

 The 2010/2011 cybercrime season has been one of the most prolific I remember. There have been the usual script-kiddie attacks, wasting bandwidth. These largely consist of morons trying to guess passwords using an automated script, and they’re doomed to failure because no serious UNIX administrator would have left guessable passwords on proper accounts. And besides which they’re guessing system account names you only find on Windows or Linux.

What seems to be a bigger feature this year is compromised “web developer” software written in PHP. This is set up by designers, not systems people, and they really don’t understand security – hence they’re a soft target.

This year it appears that phpMyAdmin has been hit hard. This seems to be a vulnerability caused by poor installation (leaving the configuration pages up after use) and using a weak version of the code that was actually fixed a year ago. When I looked I found several copies of the old version, still active, and dating from the time when the web designer had initially commissioned the site.

The criminals appear to be using a mechanism that’s slightly different from the original exploit documentation, but is fairly obvious to any programmer looking a the setup.php script. It allows arbitary uploads to any directory that Apache has write access too.

The nature of the attacks has also been interesting. I’ve seen scripts dropping .htaccess files into all likely directories, redirecting accesses elsewhere using the mod_rewirte mechanism. This appears to intended as a simple DoS attack by overloading target servers (homelandsecurity.gov and fbi.gov being favourite targets).

That this is the work of script kiddies there is no doubt. They’ve left botnet scripts written in perl and python all over the place on honeypot machines. Needless to say this makes them really easy to decode and trace, and you can probably guess which part of the world they seem to be controlled from.

My advice to users of phpMyAdmin (a web based front end for administering mySQL) is to learn how to use SQL properly from the command line. If you can’t do that (or your hosting company won’t let you, which is a problem with low-cost web hosts), at least secure it properly. Upgrade to the latest version, keep it upgraded and remove it from the server when not in use. If you don’t want to remove it, at least drop a .htaccess file in the directory to disable it, or make it password protected.

chkrootkit finds bindshell infected on port 465

The current version of chkrootkit will throw up a warning that bindshell is INFECTED on port 465 in some circumstances when this is nothing to worry about. What it’s actually doing (in case you can’t read shell scripts, and why should you when there’s a perfectly good ‘C’ compiler available) is running netstat and filtering the output looking for ports that shouldn’t be being used. Port 465 is SMTP over SLL, and in my opinion should very definitely be used, but it is normally disabled by default.

As to whether this should worry you depends on whether you’re using secure SMTP, probably with sendmail. If you set up the server you should know this. If someone else set it up and you’re not too familiar with sendmail, the tell-tail line in the .mc file is DAEMON_OPTIONS(`Port=smtps, Name=TLSMTA, M=s')dnl. Note the ‘s’ on the end of smtp.

Assuming you are using SMTPS, you can easily stop chkrootkit from printing an error (or returning an error code) simply by modifying the bindshell() subroutine to remove 465 from the list of ports to check. It’s on line 269 on the current, 0.49, version of the script.

I’m not so convinced that chkrootkit is any substitute for an experienced operator, but it’s out there, people use it and its better than nothing.

FBI hacks every VPN on the planet

Can VPN’s be trusted?

I got wind of an interesting rumour yesterday, passed to me by a fairly trustworthy source. I don’t normally comment on rumours until I’ve had a chance to check the facts for myself, but this looks like it’s going to spread.

Basically, the FBI paid certain developers working on the OpenBSD IPsec stack to and asked for back-doors or key leaking mechanisms to be added. This occurred in 2000/2001. Allegedly.

The code in question is open source and is likely to have been incorporated in various forms in a lot of systems, including VPN and secure networking infrastructure.

Whilst I have names of the developers in question and the development company concerned, it wouldn’t be fair to mention them publicly, at least until such code is found. If you’re using the IPsec stack in anything might want to take a good look at the code, just in case.

However, if the code has been there for nearly ten years in open source software, how come no one has noticed it before?

Error 0x8002007 installing Security Essentials

Good one this! If you’re trying to install Microsoft Security Essentials and it crashes out with Error 0x8002007, clicking on the Help link doesn’t really help.

If you read the technet blurb it relates to the Windows Update service not working, and if you believe this you’re going to waste a lot of time trying to repair it. I did. But the solution was really simple.

If you’re using Windows XP the Microsoft site will give you the Vista/Windows 7 version by default! Hunt around for the Windows XP 32-bit version, download that and it’ll probably work. Just don’t click the “Download Now” button because it doesn’t check which one you need – or give you the choice.

Some genuis programmers at Microsoft didn’t bother to check the version number as soon as start to run the installer. I wonder why not.

The one you get by default is:

mssefullinstall-x86fre-en-us-vista-win7

The one you probably want is:

mssefullinstall-x86fre-en-us-xp

Google is innocent (ish)

So Google’s streetview cars have been driving around harvesting people’s email passwords have they? Well this is probably true. Let’s sue/fine/regulate them!

Actually, let’s not. They haven’t done anything wrong. What Google’s surveying vehicles did was record the wireless Ethernet radio activity as they went along, to get an idea of where the WIFI hotspots are. This is a really useful thing for someone to have done – there’s no other way to find out what’s really where than by doing a ground-level survey.

In order to determine what kind of service they’re receiving you need to record a bit of the traffic for analysis. If it’s a private service, this traffic will be encrypted so it really doesn’t matter a jot – they’d be mostly recording gibberish. If it’s an open, public service they’d get the clear text of whatever happened to be transmitted at the time if the luser’s weren’t using application-layer encryption. If some technological dunderhead decides to do a radio broadcast of his unencrypted passwords, Google (and anyone else in the vicinity) will end up receiving that too.

Look at it another way – if someone wrote their password on a big sign and stuck it in the front of their house, anyone walking down the road couldn’t help but capture it. Are the pedestrians doing something wrong, or is the owner of the house an idiot?

It’s no good the idiots bleating on about Google. That won’t give them brains. It might, however, give them some of Google’s money and this could be the real motive.

The Information Commissioner, Christopher Graham, has come up with some surprising statements about Google. But on review, they’re only surprising to someone understanding the technical issues here. Does this mean Graham is a technological klutz? It’s one theory – at times it seems like everyone the government appoints to deal with technology requires this as a qualification. However I think it’s far more likely a case of bowing to media/political pressure on the subject and wishing to be seen to be doing something about it.

Then, last Friday, Google signed an undertaking with the Information Commissioner’s Office to train their staff that they mustn’t do naughty things (just in case they were ever tempted). In return for this the ICO promises to leave them alone. Read it for yourself – it’s only three pages long.

http://www.ico.gov.uk/~/media/documents/library/Data_Protection/Notices/google_inc_undertaking.ashx

What’s sad about the whole affair is that the ICO is, first and foremost, a political/media driven entity even if there are some level heads at work behind the scenes. But what a waste of time and money…

Oliver Drage makes mockery out of RIPA

Oliver Drage, suspected trader in child pornography, has just been sent down for refusing to disclose the password he’d used to encrypt his PC. This is an offence under RIPA (the Regulation of Investigatory Powers Act 2000). So if you’ve got something dodgy on your computer, you’ll get locked up whether or not the cops can decrypt it (or you’ve lost the password).

A spokesman for Lancashire police was pleased: “Drage was previously of good character so the immediate custodial sentence handed down by the judge in this case shows just how seriously the courts take this kind of offence.”

Really. Drage is going to gaol for sixteen weeks  (read “two months”) . How long would he have been locked up for if he’d given them the password so they could decrypt whatever it’s alleged he was hiding? Five years? Ten years? Lock up and throw away the key?

This is not what I call “taking it seriously”.

The penalties under RIPA for not disclosing passwords are far lower than the likely sentence assuming someone’s been up to anything of interest of the authorities in this way. They don’t take it seriously at all.

Comment spam from Volumedrive

Comment spammers aren’t the sharpest knives in the draw. If they did their research properly they’d realise that spamming here was a stupid as trying to burgle the police station (while it’s open). You’ll notice there’s no comment spam around here, but that isn’t to say they don’t try.

Anyway, there’s been a lot of activity lately from a spambot running at an “interesting” hosting company called Volumedrive. They rent out rack space, so it’s not going to be easy for them to know what their customers are doing, but they don’t seem inclined to shut any of them down for “unacceptable” use. For all I know they’ve got a lot of legitimate customers, but people do seem to like running comment spammers through their servers.

If you need to get rid of them, there is an easy way to block them completely if you’re running WordPress, even if you don’t have full access to the server and its firewall. The trick is to over-ride the clients Apache is prepared to talk to (default: the whole world) by putting a “Deny from” directive in the .htaccess file. WordPress normally creates a .htaccess file in its root directory; all you do is add:

Deny from bad.people.com

Here, “bad.people.com” is the server sending you the spam, but in reality they probably haven’t called themselves anything so convenient. The Apache documentation isn’t that explicit unless you read the whole lot, so it’s worth knowing you can actually list IP addresses (more than one per line) and even ranges of IP addresses (subnets).

For example:

Deny from 12.34.56.78
Deny from 12.34.56.89 22.33.44.55
Deny from 123.45.67.0/24

The last line blocks everything from 123.45.67.0 to 123.45.67.255. If you don’t know why, please read up on IP addresses and subnet masks (or ask below in a comment).

So when you get a a load of spammers from similar IP addresses, look up to see who the block belongs to using “whois”. Once you know you can block the whole lot. For example, if you’re being hit by the bot using Volumedrive on 173.208.67.154, run “whois 173.208.67.154”. This will return:

NetRange: 173.242.112.0 - 173.242.127.255
CIDR: 173.242.112.0/20
OriginAS: AS46664
NetName: VOLUMEDRIVE
NetHandle: NET-173-242-112-0-1
Parent: NET-173-0-0-0-0
NetType: Direct Allocation

<snip>

If you don’t have whois on your comptuer (i.e. you’re using Windoze) there’s a web version at http://www.whois.net/.

In the above, the CIDR is the most interesting – it specifies the block of IP addresses routed to one organisation. I’m not going in to IP routing here and now, suffice to say that in this example it specifies the complete block of addresses belonging to volumedrive that we don’t want – at least until they clean up their act.

To avoid volumedrive’s spambots you need to add the following line to the end your .htaccess file:

Deny from 173.242.112.0/20

If this doesn’t work for you the the web server you’re using may have been configured in a strange way – talk to your ISP if they’re the approachable type.

I have contacted Volumedrive, but they declined to comment, or even reply; never mind curtail the activities of their users.

This isn’t a WordPress-only solution – .htaccess belongs to Apache and you can use it to block access to any web site.

Perhaps there’s some scope in sharing a list these comment spambots in an easy-to-use list. If anyone’s interested, email me. This is a Turing test :-)

Why and how to hack a mobile phone

Anyone outraged that News of the Screws journalists have been “hacking” in to mobile ‘phones needs to get a grip on reality. They’re investigative tabloid journalists; what do you expect them to be doing?

To call it “hacking” is grossly overstating the case anyway – what they did required no technical knowledge other that that available in any playground in the country. All you need to do to retrieve people’s voice mail messages is dial their number, and when you get through to voice mail, enter the PIN. Most people leave the PIN as the system default.

You might argue that this is a gross breach of privacy and so forth. But it’s no more so than camping out on someone’s doorstep to see who goes in and out, following them, or tricking them into telling you something they wouldn’t if they knew your were a journalist.

New Labour was very keen to suppress the traditional liberties of the population in general and passed various dodgy laws to protect the lives of the guilty from prying journalists. In 2000, listening to other people’s voice mail was made a specific offence. “And quite right too!”. Wrong! It’s just another example of those in power making it difficult for us to check up on what they’re doing. We have (or had) a free press with a tradition of snooping on politicians, criminals and anyone else they wanted to using whatever means, as long as it was “In the public interest”.

Journalists are also out to sell papers, so the “public interest” defence is often strained to its limit, or broken. However, it should remain as a defence in a court of law and people should be able to argue their case there. It should be all about intent. But New Labour had other ideas.

People are uneasy about voice mail because it’s technological, so lets look at another example.

Suppose a journalist was camped outside someone’s house, noting down who came in and out. Another invasion of privacy, but right or wrong?

Well that depends – if it’s some innocent person then the journalist will probably end up throwing the notes away, so no harm done. If someone uses information collected in this way in the pursuance of a crime (e.g. Blackmail), that’s another matter, but journalists don’t do that.

Now supposing the journalist is investigating a suspected terrorist, and checking up to see who they’re associating with – or even a politician associating with a known crook. Clearly this information in the public interest.

It’s all about intent.

You could argue that investigations of this nature shouldn’t be carried out by private individuals but should be left to the security forces. That argument doesn’t bear scrutiny for more than a couple of seconds. The public needs the right to snoop as well as the government agents – anything else is known as a ‘police state’

As to the current difficulties – anyone who knows anything about the press will tell you that these and many other tricks are employed as a matter of course, although journalists won’t make a big noise about using them. It’s conceivable that an editor like Andy Coulson would neither know nor care exactly what his investigation teams were doing to come up with the information; you don’t ask. It’s also inconceivable that only the hacks on the News of the World had thought of it. Sources need protection.

It’s clearly a political stunt by old new Labour. Could they be upset that the press, including Mr Coulson’s old rag, turned against them? They used to be friends with the News of the World. At the time of the original scandal, it appears that the first politician to call Andy Coulson to commiserate with him about having to resign was none other than Gordon Brown. Apparently he went on to suggest that someone with his talent would soon find another job where he could make himself useful. (Source: Nick Clegg at today’s PMQs).

Intel has just bought McAfee

Intel has just bought its neighbour in Santa Clara.

Well there’s a surprise. According to today’s Wall Street Journal it’s a done deal at $48/share (about £5bn). Paul Otellini (Intel’s CEO) has been saying that “security was becoming important” in addition to energy efficiency and connectivity. This lack of insight does not bode well.

I’ve been expecting something like this since Microsoft really got its act together with “Security Essentials”, its own PC virus scanner by another name. Unlike other PC virus scanners, Microsoft’s just sits in the background and gets on with the job without slugging the PC’s performance. Why would anyone stick with McAfee and Symantec products in these circumstances?

Whether PC virus scanners have much benefit in today’s security landscape is questionable, but at least the Microsoft one does no harm.

Intel has (apparently) paid about £5bn in cash for McAfee. I wonder if they’ve paid too much. It’ll generate revenue while lusers and luser IT managers are too scared to stop paying the subscription, but as anti-virus becomes built in to Windows this is going to dry up. I suspect McAfee was aware of this situation ad was moving on to mobile device security – not by developing anything itself, but by buying out companies that are.

When McAfee bought Dr Solomons in 1998, it was basically to pinch their technology for detecting polymorphic viruses and close down their European rival, which they did – everyone lost their jobs and the office closed. (Declaration of interest: Dr Solomons was a client of mine). Whether McAfee has any technology worth plundering isn’t so obvious, so presumably Intel is buying them as a ready-made security division.

McAfee does, of course, have some good researchers in the background – we all know the score.

BlueWatchDog Review

This is an almost brilliant idea. The BlueWatchDog is a thick credit-card sized device that picks up the signal from your paired Bluetooth ‘phone, and if it gets separated from it, sounds an alarm. Great if you’re the type to leave your Blackberry behind or you iPhone is pinched from you handbag. At just £40 it could save you a lot of hassle.
BlueWatchDog
I said it was “almost” brilliant. The snag is that it requires an application running on the mobile device. It’d have been better if it could pair with anything Bluetooth, at least as an option. The application can be used to set the range before the alarm is set off but this is functionally you could live without. As it stands it works with Android, RIM and Apple mobiles. Apple, incidentally, didn’t like the idea of them giving the App away but the company have struck a deal to make this possible.

I managed to speak with the inventor and suggested a version that would work with any Bluetooth unit – possibly by treating it an audio device. Watch this space (and I hope he sends me a sample!)

www.mindyourit.co.uk
0800 999 2177