Spam-Bot Intrusion Caught — Now What?

An anonymous reader wonders: "I've recently detected and halted an intrusion on my home computer, taken some actions to prevent further intrusions, and located the software that was running a bot agent. Cursory examination showed that the bot software is intended for acting as an agent for spamming. Configuration files distinctly point at the user/host/domain of several bot-herders — damning evidence. Nothing would please me more than to see this botnet to be caught and disassembled, I'm sure much of the internet-using community would support this. Thanks in advance for your suggestions. So, to whom should I disclose this information for appropriate investigation, follow up, and countermeasures? "

Comments

one word

[Jbcarpen]

Spamhaus.

Re:one word

[XenoPhage]

If somebody's ISP is blindly rejecting mails due to nothing more than a positive Spamhaus hit then that's the fault of the ISP! This is like discussing religion or OS preference...

What would you have ISPs do to stop spam? Spamassassin, properly tuned, does a decent job, but it doesn't solve the underlying problem. If an ISP allows *every* incoming connection and relies on spamassassin to detect and mark mail, then they have to ensure that there is sufficient storage for the spam. In most cases, the amount of spam incoming to a system is over ten times more than normal mail.

Using something like spamhaus helps out considerably because it does block a lot. Unfortunately, like every single other system out there, it has flaws. As with other approaches, the goal is to find a happy medium of sorts. The result is, however, that you can't please them all.

I've spent quite a bit of time on spam prevention for my own server and it's definitely not easy. I have about 5 tiers of spam detection at this point and, while it's catching about 99% of the spam, some still gets through. As a technically savvy user, I can deal with this and the level of detail required. For the normal ISP user, however, it's a different story. They don't have the technical know-how to tune their mail filters, nor do they generally have any interest in doing so.

So, until someone comes up with the perfect filtering system (which the spammers will likely adjust to within a few days), there's not much else to do. Personally, I don't have the time or money to deal with every single incoming spam and blocking some based on a well-known RBL is fine for me.

Places to report to...

[caitriona81]

1) Don't contribute to the problem. Attacking botrunners directly, or vigilante action doesn't help, and may actually be harmful - by teaching them how to build better drones. See http://fm.vix.com/internet/security/superbugs.html

2) As for US gov't agencies, if you or the attacker seem to be in the US, http://www.ic3.gov/ is likely to be interested. http://www.cert.org/csirts/national/contact.html can also put you in touch with nationial computer security incident response teams, who will also be interested (you only need to contact the one local to you, please don't shotgun complaints to all of them.)

3) As for private companies and research organizations, if the bot isn't already clearly and specifically detected by antivirus, report it to them, following their reporting guidelines. Shadowserver (http://www.shadowserver.org) seems to be interested in researching and gathering intelligence on botnets also.

Re:Places to report to...

[Anonymous Coward]

Attacking botrunners directly, or vigilante action doesn't help

The spirited attack on and destruction of Blue Security and the spam flood that followed, does not support that assertion. Somebody wanted them gone badly, for a reason.

Re:Places to report to...

[caitriona81]

I should probably rephrase and clarify, attacking them directly without legal action to back that up is bad - ie, if you are going after a bot runner, it needs to be in a manner that not only takes away their toys, but also puts them in jail, for a long period of time. If you can't take away their freedom in the process, then you aren't doing us any favors by teaching them how not to get caught -- botnets, and their means of control get more and more sophisticated, with overall trends towards plausible deniability and robust survivable command and control networks, designed to either resist attack, or be reconfigured after the fact to retain control of compromised hosts.

This is a far cry from when botnets were controlled "in the open" on public IRC networks - the kiddies are clearly learning something with each iteration, and they are sharing that knowledge amongst themselves. Also of note is more use of packers, executable encryption and anti-debugger routines, which were completely absent from early botnet executables. Use of rootkits, as well as secondary backdoors (to regain access after the system owner detects the intrusion) are also on the rise.

Re:Places to report to...

[tacocat]

I disagree. If you could determine the physical location of such bot herders and disclose that to the internet at large, I'm sure that there would be a final solution applied that people would be willing to turn their backs on. Especially if you could post photographs, names, and physical addresses.

You could always try private sector...

[BinarySkies]

There is an organization, ShadowServer (www.shadowserver.org if I recall right) that specializes in mucking about with Botnets. They'd probably have the right contacts and such to deal with that.

What actions?

[dbIII]

Were the actions to install from scratch on a new disk / take a disk image to look at later + reformat + reinstall / poke around for a bit with the thing not on the network before reformat + reinstall / rely on external sources for info and just wipe the thing / or did you take the common and lazy approach now of just fixing the obvious damage and hoping the rest of the system is not compromised? The real pain is you can't even trust the backups in some cases especially if the people responsible for the machine ignore it most of the time - it may have been rooted for a while.

Preaching to the converted here but I'm amazed how many people do not realise that an owned computer is exactly that - there is nothing at all you can trust absolutely so you have to look at what is on the disk with something else and have to wipe it and start again. On *nix script kiddies love to put things in unexpected spots in the init scripts like in /etc/init.d/functions or the equivalent, or replace things like ntpd that you expect to talk to the outside world - so they would have control well before you get a shell. Some linux rootkits changed the generally useless ext2/ext3 file attributes in a cute effort to make cleaning up harder for those prone to try - it made it trivial to find their stuff becuase it would be the only thing on the volume with attributes set. Even then you can't trust that is all they did - it's just an obvious sign that you cannot trust anything on the machine.

Name and shame

[Anonymous Coward]

How did you get the infestation? What did you download?

contact the ISP/registrar

[sp1n]

You have the bot herder address. To do the most "damage", get it shut down. Contact the ISP abuse department who hosts it. If there's a DNS name, also contact the ISP hosting the authoritative DNS zone and possibly the registrar, who may elect to terminate the domain. If you don't get a response from the ISP, contact their upstream provider(s) (if a smaller Tier 3 ISP).

Whois is your friend.

Re:contact the ISP/registrar

[bernywork]

Fantastic. Get the persons account shut down, like most people these days, who have multiple domains, internet links and everything else, he will be offline for what? A couple of hours? Your just going to piss him / her off.

No, the best thing to do here is kill the whole problem. All the machines in the botnet need to be cleaned and updated so that they don't get re-infected, otherwise they will get taken over by someone else (Yes, I know most people when they infect a system DO update it so that someone else can't take over, but they leave back doors). The person running the botnet needs to see the beak (Judge). It might be that the beak decides that a slap on the wrist is the appropriate action, but I think just cutting off one point of access / control of a bot net which I am sure that they have other control over is just silly.

Re:contact the ISP/registrar

[mandelbr0t]

Usually you won't get anything from the ISP. I start with ARIN and move to RIPE, APIC as the search suggests. I run into one of two scenarios:

1) There is a properly listed contact for abuse reports to whom I send the complete relevant log entries in text format. I usually don't hear from them again, but I also don't see any further network abuse from that netblock owner.

2) The owner of the IP block is a complete and utter joke. Examples: they don't correctly configure their reverse DNS, so they will claim that you have the wrong IP address, they list an abuse contact that doesn't speak English, they send spam in reply to your abuse complaint (that actually made me laugh for a moment). In this case, you also won't hear anything, but you should probably go to the effort of banning such an irresponsible network at your firewall.

Generally you won't hear anything. You won't know if someone has seen or acted on your complaint. Just think of how many network abuse complaints a large, responsible network would have to deal with daily. There's also dozens of fly-by-nights that make it clear that they won't make their network behave no matter how much you complain.

Surprisingly, I've found that larger netblock owners are quite responsible. A threat to block their entire netblock at your firewall is an effective one, easy to carry out and perfectly justified. Just be sure to remove the block if they show that they have fixed your complaint.

Use your government

[tigersha]

Easy.

Hack into the US Navy weapons control website.

Search for a file called "city-coords.txt".

Find out what the lat and long is of the spammer.

Change the line "Al Queda Base 4:xxx" to reflect the new coordinates.

Dress as Osama and make a press release with a big "Base 4" sign behind you. Use a good make-up artist if you want.

Two days leater and BAM!!! the spammer is gone. Your tax dollars at work for you!

Re:Use your government

[WrongSizeGlass]

Um ... I think you forgot "Profit!".

Re:In Soviet Russia...

[Opportunist]

You can leave the "Soviet" out of this sentence to actually make it true...

Re:Law enforcement comes to mind first

[Opportunist]

Don't make me laugh. Law enforcement usually looks at you with a rather blank stare and says something along the lines of "And ... what should we do now about it?"

It's not that the nets would be unknown. Every security researcher worth his salt has a fairly good idea where those botnets are and how they work. The problem is, nobody with the legal muscle to do anything about it would care.

The sad answer: Nothing

[Opportunist]

Clean your computer and go on with your life. Everything else is a waste of precious time, energy and nerves.

What could you do? You could inform your local law enforcement. Which will invariably end up in a file cabinet within moments because they have no clue how to deal with it.

You could go a step higher and contact your country's equivalent of some sort of "internet police". Most countries have that today. They will look at the info, find out where the spammer sits and depending on where he sits it goes different roads. Either he is in a country within reach, i.e. your country or one where Interpol/Europol actually has some muscle. In this case, they will maybe even go through the hassle of dealing with the provider hosting the spam controller, and within 2-3 weeks they finally got all the papers necessary to shut the machine down. A day later, the spammer opens up a new one and the party continues.

If the machine is somewhere in Russia, far east or some country ending in -stan, nothing is being done and it just continues from the same machine.

The spammer himself (or rather, the individual registering the server) is invariably sitting in some of the countries mentioned in the previous paragraph and thus untouchable anyway.

In short, the best you can achive is to annoy a spammer. Just in case the server switch wasn't due anyway because you can only use a spamcontroller for a certain amount of time before the ISP gets interested and starts to "persuade" you to move.

Was in similar situation

[mattr]

I had my own server broken into for the first time, wasn't a botnet but a bank of america style phishing site. I discovered it when trying to make a subdomain with the control panel didn't work right.. the provider said they cleaned some out but couldn't be sure and then in fact I found the servers myself, in /root and /tmp disguised as other files. I mailed yahoo and google since both had email addresses being used, and told the isp. Guess what? I got no response from google, and none from the isp (they totally suck too, I've been down for a month after being told to erase the disk and they upgraded me - to Fedora Core 2! - and are so incompetent it is not even usable anymore. So I'm changing to a better managed hosting company rsn.)
I did get a thank you from Yahoo. But, the first one was clueless, ignoring the content of my letter. I got a second one from them saying thanks. But that they couldn't accept attachments. So couldn't send them the proof.

At any rate, what I did is erase the disk, restore from backup and some checked files, and lose a lot of time. There is probably little more you can do than simply report to one of the links below that you have a botnet address then as quickly as possible erase it.

I also found a number of commands changed in /bin however I couldn't tell if it was the crackers or the isp who did that. It was running out of date software, and though they failed lots of ftp login probes it looks like they got in through an out of use user's login somehow and promoted to root.

Moral of the story? If you use a managed hosting service, keep a FULL backup locally. Run tripwire or something similar, I will from now on. Use a hosting service that is not completely clueless. Do not try an upgrade or anything afterwards. Have a portable hard disk you can use - my ipod was very useful. The most annoying thing was having to spend lots of time on the phone with admins, and having my email and website hanging in the air. The answer is to immediately cut all your losses, get another system maybe on another provider. Possibly you could even do this with a local machine and dyndns temporarily but if you're busy the last thing you have time to do is mess with crooks. Best thing that came from it is I discovered several other hosting companies from friendly clients who helped me get my jobs done.

Re:Was in similar situation

[bleh-of-the-huns]

Yahoo and google etc are not clueless, just over worked. I have worked security for large ISPs, UUNET (prior to MCI getting involved), AOL Time Warner, and a couple of others. They get far too many complaints to be able to respond to each, so you are lucky if you get an autoresponce, but don't expect them to contact you, there is just no time for it. The attachment problem is due to the fact that in many cases, complaints are placed into a tracking system, so instead of an attachment, you end up with uuencoded text, its a pain to have to reassemble that manually for every complaint, and if you hit up the security pages of those websites, they clearly state not to use attachments.

Unless the botnet has caused more then $5k in proven damages, with tangible evidence, law enforcement will not get involved, this is at the federal level, not sure about state and local, as they rarely deal with cyber crimes of this type, they prefer to deal with cyber stalking and threats to individuals in their localities. If you must report a botnet, report it to USCERT (run by DHS), they may not be able to get to the root if its in one of those countries listed, but they can research it, and they are capable, and if something can be done, it will be done in the background.

Re:Virginia Tragedy

[Anonymous Coward]

[quote]We [b]will[/b] win in 2008, and you [b]will[/b] lose your guns. And there isn't a damn thing you can do about it[/quote] And criminals [b]will[/b] still, well, be criminals and obtain weapons illegally. Banning guns won't stop crime committed with guns.

So instead you're advocating a ban on angle brackets?

SANS

[gunnk]

The good folks at SANS do their best to act as early warning and protection for the net. They'd likely be interested in helping break this up AND they have the appropriate contacts in government and law enforcement to do so.

You can contact them here: http://isc.sans.org/contact.html and see if they are interested or can direct you to the appropriate person or agency contact.