DEF CON "Capture the Capture The Flag" Data

pablos writes "Each year DEF CON hosts the famed Capture The Flag contest. Hackers from all over the world duke it out on the network for 72 hours, hacking for the title. The Shmoo Group diligently logs every packet for posterity, we "Capture the Capture The Flag." Now is your chance to download by far the most interesting, 'sploit ridden, 5.8GB of intrusion collusion ever published. Free for the bandwidth endowed, this is the ultimate IDS testbed."

Making waves...

Hmm...my favorite was the sinusoidal IP address spoofing. Anyone else?

They cheated us.

At defcon 8 DPS was at defcon and Burrows straight up social engineered his way into the server room and rooted the main box. So technically we had just won but they disqualified him cuase they wouldn't acknowledge social engineering as valid. Before defcon 8 DPS (dead [protocol] society) had pretty much dominated the social engineering contests but defcon 8 was the first year they decided to stop doing the social engineering contests so we were forced to improvise.

Re:They cheated us.

Ghettohackers quite brutally owned my laptop. One of 'em started chatting with me, asked if he could check his email...though I watched the screen, it's always polite to look away when someone types in their password.

Except when that password is

notepad c:\flag.txt
ghi

Now, at the time I damn near killed someone over that...but I realized pretty quickly it was a damn slick hack. Ask, and ye shall receive. Even from me.

--Dan

Where do I download...

...The .pak files?

*ducks*

Site is slashdotted (almost), so here are mirrors.

Well, since the site is getting hit pretty hard, here is a direct link to all the mirrors:

Capture the Capture The Flag Mirrors

If you have a mirror up, please let me [mailto] know.

If you're using wget to pull the data, please use the following command:
&nbspwget -r -nd --no-parent -R "=A","=D" http://site/path/

US - Wisconsin (100Mbit):
http://www.wi2600.org/mediawhore/mirrors/shmoo/cct f-defcon9 [wi2600.org]

US - Colorado (100Mbit):
http://www.ucar.edu/temp/shmoo-defcon9-ctf/ [ucar.edu]

US - Pennsylvania (T1):
http://www.bitsend.com/defcon9-cctf [bitsend.com]

US - Alaska (DSL):
http://cctf1.shmoo.com [shmoo.com]

Please be sure to read the license [slashdot.org].

Bandwith Endowed

'sploit ridden, 5.8GB of intrusion collusion ever published. Free for the bandwidth endowed'

You would need alot of and bandwidth and evan more time on your hand to evan start on.

Now , let my see ... on a 56k modem ,(if my math is correct) , then thats about 10 days...

Well I hope they Capture the Slashdot Effect.

Putting a couple of Gigs data on the net and then having the bad luck to be posted on Slashdot is going to mean that their link will be unreachable for most of the day. :-) But hey it will probably make for neat graphs.

Bandwidth Cost

How are they going to pay for the bandwidth cos on this...if evan just 1000 people download it (and it has been slashdotted) then it will 5.8 Terabytes of information to be downloaded.

This won't exaclty be payed for by a banner ad.

Re:Bandwidth Cost

Did you read who these people are? I don't expect acquiring bandwidth is much of a problem, if you know what I mean.

Social engineering is the way forward

In the dim and distant past, before I became 'respectable' I used to be a hacker wannabe. I used to use my 1200 baud modem to dial into various systems, and lets just say, that had the law been the same then as it is now, I could have been arrested.

But after I became involved in tech support for major financial institutions, I realised that although security there was reasonably good, you could almost always circumvent it via social engineering.

My favorite trick to get into the server room was to put on an old hard-hat and a fluorescent jacket. I would stand outside the door until someone came along, then I would simply ask them to let me in. Which about 70% of the time, they did. At which point, I would point out to them that I could have been anyone, usually got an embarassed apology.

I was using social engineering to raise the security awareness of staff, but it was a real eye-opener to me just how easy it was to control people.

Re:Social engineering is the way forward

My favorite trick to get into the server room was to put on an old hard-hat and a fluorescent jacket.

Yes, my favorite way to get into the server room is to dress up as a member of the Village People, and then wait for some random person to agree to take me into the closet.

Mirror in the making

Mirror in the making at http://deimos.siliconinc.net/cctf

Its currently chugging away at about 250 kbps, so it should be done within a few hours. There is already 1+ gig of data up there for your browsing pleasure, and its chugging away at around 250kbps. Enjoy. If it breaks email me or something.

Traces collected using ??

Tcpdump? or what else?

n-ctf sucked this year, I hear...

from what I hear, n-ctf SUCKED this year...

From a friend whom was on one of the teams:

We set up some 'reflectors', using the MIRROR target of the Linux netfilter and almost got booted of the net by the judges for this unique solution.

Bleh.

a bit of hyperbole

The Shmoo Group diligently logs

every[*] packet for posterity

I don't know about defcon 9 (2001), but I seem to recall them only being able to get part of the traffic at defcon 8 (2000).

[*] my emphasis, not theirs

even better

the shmoo group's data gives an idea of the type of attack tools that are most commonly used in intrusion attempts, but if you want to know the tools and techniques that are the most likely to succeed, it would be good to talk to Caezar [caezarschallenge.org] or some other member of the ghettohackers [ghettohackers.net]. After all, they are the ones who win at capture the flag year after year....

Is there any commentary on the 5.8GB

Call me lazy (I am) is there a summary or commentary on all that raw information that can show us hacks attempted, both successful and unsuccessful. It even gives some hacker some reflected flame at deciphering and commenting on the information. If I was considerably less lazy I might do it myself.

Greed is Good - 1980's
Lazy is Good - 2001

Article's title

<P>I suppose that would then be "Metacapturing the Flag".
<P><H3>This useless comment was generated by a Cockpitful of Suicidal Fanatics for you</H3>

CTF Rules

The rules for CTF at DC9 were, unfortunately, not well tested prior to the actual event. The intent of the rules were to provide more targets to attack, by shifting the burden of providing targets to the competitors. However, with the rules as written at the beginning of the contest, it turned out to be (pointwise) not worth attempting to hack. The net effect of the rules were that most groups were simply putting up a server, getting the points and pulling it down. While this is a valid strategy for that ruleset, it doesn't make for much of a hacking competition. This constant churning of servers also made hacking difficult, with targets disappearing by the time you could identify them through the standard CTF network instability.

We (the GhettoHackers, with the much appreciated help of Jennifer Grannick) managed to slowly, over the course of the competition, convince Miles to change the rules to a set more conductive to an actual hacking competition. When teams began merging due to the rule changes, we merged with Digital Revelation, to both group's benefit. We gained their server points, and they gained our capture points.

Besides winning CTF, the GhettoHackers / Digital Revelation team also had the highest average Blood Alcohol Level of any group (check out http://cow.pasture.com/~tcroc for more details). As announced at the awards ceremony, we, the GhettoHackers, have retired from CTF after DC9. To help foster more competiton, and for a different application of our expertise, the GhettoHackers will be helping to run CTF at DC10.

Input from a member of the winning team.

Defcon 9 was my first time with CTF and I must say, it's not exactly what I expected. My buddy Thalakan got recruited to Digital Revelation and he recruited me over there. 90% of the time, everyone hacked systems that were difficult to hack. All the servers on the server segment (x.x.x.250-254) had either chrooted systems, patched servers and for a day and a half, nothing happened. During that time, the most exciting thing was when Dan got social engineered (see above link). However, 2 hacks did happen. I think it was prophet on digital revelation who rooted a win2k box with the unicode exploit. Then, the most exciting hack was the obsd 2.9 local exploit. Someone from the grey team finally setup a server with local access (he gave out login/password) and the race was on to apply the exploit. By this time, we were already merged with ghetto and everyone watched in anticipation. Eugene, from the ghetto hackers worked ferverntly and a bunch of us watched in anticipation. Because of the race condtion, two teams simultaneously rooted the server at the same time and split the points.

Since there was physical access to the box (they were located right next to the operator), I heard that people yanked network cables when they were about to be rooted.

There were many interesting systems and different programs that ran on the network but without source, 2 days is simply not enough time to do anything substantial. I hope next year, Caesar and the Ghetto Hackers will run a better job of providing more interesting hacks. I'm hoping the judges will put up servers that arent locked down. Those roots will be for maybe 10 points. Roots in servers with no known vulnerability (with source provided) will give 100 points. Something like that would provide with more hacks than the 3-5 roots we had. Having each team provide servers that are locked down is plain stupid.

-Nouveaux