The DMZ

Ulkoma

16:00

Nice answers

AJ Henderson

@Xander ah good, someone else hit on the bodyguard thing too

I've never found a good way to catch up on chat, doesn't matter which side you start on

Jeff Ferland

Security expert usually overlies with some other profession. Security/systems admin, security/programmer, security/policy management... and security/cryptographer is an overlay as well. Practical application of algorithms is distinct from from the algorithm itself.

AJ Henderson

but yes, at it's most fundamental, "security expert" is a knowledge and awareness thing, not a job

but even there, there are different break downs

Thomas Pornin

@AJHenderson It is a job when you manage to get paid just for being awesome.

AJ Henderson

in terms of what area of security someone is focused in

Rоry McCune

16:03

that's amusing... drupal vuln notification

Drupal core vulnerability sa core 2014 005 is now public: http://cgit.drupalcode.org/drupal/commit/?id=26a7752 Platform.sh customers receiving notifications shortly.

Tweeted by robertDouglass on October 15, 2014 at 4:02 PM

and the site's down

Terry Chia

@RоryMcCune Probably written in PHP.

Rоry McCune

@TerryChia you knows it

Jeff Ferland

Pro tip of the day: isolate your processes. LXC that shit.

Xander

@RоryMcCune Ha, figures.

Ulkoma

Can I become a security bartender?

Adnan

16:04

nvm

Jeff Ferland

@Ulkoma DEFCON

Adnan

@JeffFerland I like that game

AJ Henderson

@Ulkoma that might be hard, but it wouldn't surprise me if some high end bouncers have security expertise

Jeff Ferland

Executive protection, physical security design.

AJ Henderson

identity verification

Rоry McCune

16:07

@Adnan ta :)

Adnan

@RоryMcCune drupal.org/drupal-7.32-release-notes

Here it is :D

Jeff Ferland

Security expert:

Mr. T

Mr. T (born Laurence Tureaud; May 21, 1952) is an American actor known for his roles as B. A. Baracus in the 1980s television series The A-Team, as boxer Clubber Lang in the 1982 film Rocky III, and for his appearances as a professional wrestler. Mr. T is known for his trademark African Mandinka warrior hairstyle, his gold jewelry, and his tough-guy image. In 2006 he starred in the reality show I Pity the Fool, shown on TV Land, the title of which comes from the catchphrase of his Lang character. == Early life == Mr. T was born in Chicago, Illinois, the youngest son in a family with twelve children...

Rоry McCune

@Adnan that's nasty, I forsee pwnage in peoples future

Terry Chia

@Adnan Wow.

Adnan

Unauthenticated non-blind SQL injection

Sweeeeet

Terry Chia

16:09

Highly critical indeed.

AJ Henderson

@Adnan wow, that and SSL3 breaking in the same day. Good times

need some other major announcement for a trifecta

better start feeding the rumors

Terry Chia

@AJHenderson Did you miss the MS code execution vuln yesterday?

Adnan

@TerryChia Meeeeh

AJ Henderson

@TerryChia I guess so

Adnan

> Please download this file and run it on your computer

AJ Henderson

16:10

oh wait, I might have heard a little bit

Adnan

Who knew that can lead to code execution?

AJ Henderson

it was an old one they patched yesterday right?

used by some targeted spy malware

Terry Chia

@Adnan It's not interesting but it's definitely high impact.

AJ Henderson

I read briefly, but was distracted by the prospects of SSL3.0 breaking entirely

I kind of passed over it as not all that interesting though

Digital Fire

Damn I love you guys!

Terry Chia

16:12

Technical details on Drupal Vuln: sektioneins.de/en/advisories/…

Digital Fire

That drupal sql would of hit a few of my sites =P

Adnan

@DigitalFire We love you too, as friends.

> To handle IN
statements there is an expandArguments function to expand arrays.

:D :D :D

Shit.. I... I have some all-night coding to do

I handle IN statements pretty much the same way Drupal is handling them

Rоry McCune

@Adnan oopsie :op

Adnan

Oh, I don't seem to be vulnerable. There's no way to create a keyed array

Phew!

Anybody want in on a Poc for this Drupal thing?

Only 2600 results "WHERE x IN"

Undo

16:31

I've got a Ubuntu 12.04 server running Apache, what do I need to do to make my poodle stop bleeding?

Xander

I do believe that Twitter has some issues handling bi-di text.

raz

@Undo just hit undo

Thomas Pornin

@Undo Cauterize the wounds with hot iron.

Adnan

Anybody has a Drupal site they don't really care about?

I need to test this

Xander

@Undo Wouldn't you know it, we have just the thing for you.

1

A: SSL3 "Poodle" Vulnerability

To disable SSL v3.0 support: In Clients: Mozilla Firefox Either Install the Mozilla add-on called "SSL Version Control" Or Type about:config into the navigation bar and press [Enter] Accept the warning and proceed Search for tls Change the value of security.tls.version.min from 0 to 1 (0...

Undo

16:40

yay

Rоry McCune

@Adnan @Adnan here you go shodanhq.com/search?q=drupal these people obv. don't care as they've not bothered removing the drupal version info from their headers :)

Undo

wait, I don't even have an SSL certificate. So I guess this doesn't really matter for me :P

Adnan

@RоryMcCune Thanks!

I think I'll be able to publish a PoC before SektionEins do

fingers crossed

Kisunminttu

The donut has landed

Crazy

Inb4 Adnan gets v&'d and CFAA'd.

Xander

16:44

@RоryMcCune Oh good, the LinkedIn Developers Network. That's fun.

Adnan

Guys.. anybody?

I don't really know much about Drupal, so navigating the files is quite difficult

I have 41 hits on WHERE x IN, 4 of them are in the user handling stuff (might be login or registration)

At least 10 keyed arrays

COME ON!

Digital Fire

@Adnan so did you get your poc to work?

Adnan

@DigitalFire Nope. Can't get the array to initialize with key values I want

This definitely needs two more eyes

Lucas Kauffman

AJ Henderson

anyone know if there is a way to tell which cipher suite is being used by IIS?

I'm trying to redirect people using SSL 3.0 instead of shutting it down outright if possible

but haven't had much luck figuring out ~~a good~~ any way to do it

raz

16:57

Are you trying to find which ones are enabled?

Adnan

@AJHenderson You're trying to cause redirects with ASP.NET code or with IIS configs?

AJ Henderson

@raz no, that I know, I'm just trying to give a more friendly failure

Lucas Kauffman

@AJHenderson why not just put it as the least prefered cipher?

Adnan

AFAIK, you cannot see SSL info on either of those levels

AJ Henderson

@LucasKauffman because of forced failures

Adnan

16:58

@LucasKauffman The attack causes the connection to downgrade

AJ Henderson

an attacker could force to SSL3.0

so I don't want to allow someone to actually use the site as that

but I want to have a page that explains to them why they need to update from IE6.0

Lucas Kauffman

@Adnan still it's not THAT bad that we need to disable SSLv3.0 all of a sudden

@AJHenderson seriously

Adnan

@AJHenderson You can't

Lucas Kauffman

@AJHenderson You have people running IE6 and your biggest concern is the SSLv3 vulnerability?

AJ Henderson

@Adnan that's unfortunate, but corresponds with what I was starting to think :/

@LucasKauffman hopefully no one accessing my site is using IE6

but I don't know who is going to try to access it

and I'd prefer not to have a negotiation failure, particularly since I force SSL on my sites

and I'd also rather not disable forcing SSL

Adnan

17:00

@AJH I can think of some ways to do it, though

The problem is that they all involve modifying your SSL lib

At the level where SSL stuff are being negotiated, you cannot do any redirects, so you need your SSL lib to keep track of that and "report" it to your IIS (or something running on it, like ASP.NET) which can cause the redirect

AJ Henderson

the programmer in me is rather bothered by the fact that IIS doesn't bother to share the connection details

Thomas Pornin

@AJHenderson If they are still using IE 6.0, then it is doubtful that a mere Web page will be sufficient to make them amend their ways.

AJ Henderson

@Adnan yeah, that I know

Adnan

@AJHenderson Why would it? These are very low low level stuff

AJ Henderson

@Adnan because of security concerns like this

it isn't exactly unforseen that old protocols would break over time

Adnan

17:03

So.. nobody knows Drupal here?

AJ Henderson

@Adnan not particularly well

Thomas Pornin

@Adnan I tried it once; I didn't like it.

AJ Henderson

I've fiddled a bit

Adnan

I'm trying to produce a PoC for the latest vuln.

It's just sooo slow since I'm learning Drupal's structure

A lot easier than Wordpress, but not that easy

AJ Henderson

@Adnan it reports cert details

Adnan

17:06

@AJHenderson Hmmm.. then it doesn't really make sense not to "report" the ciphers

AJ Henderson

@Adnan yeah, exactly

Digital Fire

@Adnan I know it.. somewhat

AJ Henderson

but it just reports if SSL/TLS is used or not, not which suite is actually used

Adnan

Okay, possible hit

The injectable area seems to be where you change the language of the Drupal site

YESS! This is, in theory, looking very good.

Jeff Ferland

The DMZ: releasing exploits early since 2014

AJ Henderson

17:14

yeah, I somewhat wonder if we should setup a room similar to the CTF room for that kind of thing

though I suppose if we can do it, chances are other people that are less friendly could too

Adnan

Yes! Managed to get it to mess the query!

Jeff Ferland

Until we break something, let's keep it the way we are.

Adnan

Now let's see if I can inject something

Anybody knows PDO stuff?

@Simon

Jeff Ferland

@Adnan SQL is less noticeable if you do it between the toes.

Adnan

PDO people, are multiple queries as straightforward as query1;query2?

Undo

17:16

I don't think so

IIRC PDO has some "protection" around that

Adnan

Confirmed: stackoverflow.com/questions/6346674/…

> PDO_MySQL is a more dangerous application than any other traditional MySQL applications. Traditional MySQL allows only a single SQL query. In PDO_MySQL there is no such limitation

Well, multiple queries seem to be working in an isolated env.

Let's try in my injection

YES! Found another possibly injectable area

modules/user/user.module:3633

Digital Fire

That looks to be the page that displays all the content per user.

Simon

@Adnan Oh for fucks sake, I could have sweared that this wasn't possible with PDO.

Adnan

@DigitalFire Hmm.. accessible without login?

Digital Fire

It should be given that the user or the site doesn't have any ACL put in place to restrict anon users

Adnan

17:30

@DigitalFire While this is still good, I think I'll keep looking until I find something that allows anon by absolute defult

@Digital Any info on modules\profile\profile.module?

Digital Fire

That has permissions attached to it

not anon by default

but i did find something interesting thou on one of my production sites.

Adnan

@DigitalFire Do tell

Digital Fire

Anyone else notice that you can enumerate the usernames by doing: user# ?

user / #

Adnan

@DigitalFire Do you have URL?

a URL demonstrating this, I mean

Ah, it's been a long day. I'll call this one a tie.

If anybody is interested, the juicy stuff is in includes\path.inc starting with line 83 until line 117

another important hint is HTTP_REFERER

Ooooh.. manged it in the URL itself with no referrer play! :D

Digital Fire

18:14

drupal.org/node/1004778 WHY?! OH WHY! YOU DO THIS DRUPAL SECURITY TEAM!

so it is an on going issue...

bah

Xander

@DigitalFire ??

Digital Fire

@Xander They don't consider user enum to be a vulnerability because they have third party modules that can mitigate it... dunno. doesn't sit right with me..

Xander

@DigitalFire I don't think that was the primary point. What they really said was that it isn't a vulnerability because it isn't in their threat model, which is reasonable.

AJ Henderson

@DigitalFire it's usernames

Xander

@DigitalFire The mitigations (third-party modules) are a nice option for specific installations where that is in the threat model...That doesn't mean that it should automatically be added into the the core threat model.

Digital Fire

18:30

Ehhh... I imagine a distributed bruteforce similar to the attack on Wordpress earlier this year.

While Drupal out of the box has a lockout. It just seems like bad practice to allow user enumeration.

Xander

@DigitalFire If usernames are public, which it sounds like Drupal intends them to be, user enumeration is a feature, not a bug. This is idential to Google's decision to declare that open redirects on Google properties are not a vulnerability, despite the fact that they're in the OWASP Top 10.

raz

@Xander Everyone says, "It's a feature not a bug." And every software package has SO many features.

Xander

@raz The point is, it's the Dupal developers who get to decide if something is desirable behavior, undesirable behavior, or irrelevant behavior. If they have use cases for user enumeration they want to support, then it's desirable behavior, whether you particularly like it or not. So, you're certainly free to disagree with their decision, but that doesn't mean that it's the wrong decision and they should change it.

Gilles

18:47

0

Q: How can I migrate a bully (reaver alternative) session from Kali to Ubuntu?

I'm running bully on Kali Linux and this is the output : [!] Received disassociation/deauthentication from the AP [+] Rx( ID ) = 'NoAssoc' Next pin '99728298' [+] Rx( M5 ) = 'Pin1Bad' Next pin '19268293' [+] Rx( ID ) = 'EAPFail' Next pin '19268293' [!] Received disassociation/deauthe...

raz

That's a cool story

raz

19:11

@Xander have you ever had the Unibrou 17 Grand Reserve? I see you like belgian strong dark ales, I highly recommend it.

Xander

@raz Yes, it's fantastic. Terrible is my all time favorite Unibroue brew, but the Grand Reserve is right behind, and well, everything else I've ever had of theirs has been excellent.

Polynomial

evenin' all

Xander

@Polynomial Evenin'!

Polynomial

could I perhaps bother y'all to check over something for me?

I wrote a thing on the POODLE bug for work: labs.portcullis.co.uk/blog/…

raz

@Xander Unibrou has never failed me either. It's been a while since I've had Terrible

Polynomial

19:13

if you spot any bugs or anything that is really confusing / unclear about the way I've written it, let me know

AJ Henderson

@Polynomial I've got one thing that's unclear... what's SSL?

;)

Polynomial

@AJHenderson quiet you ;]

Undo

@Polynomial minor nitpick: This gives rise to a padding oracle bug, which is how BEAST worked too. sounds a little clunky, maybe This gives rise to a padding oracle bug, which BEAST also made us of. or something

AJ Henderson

@Undo or possibly "This gives rise to a padding oracle bug, similar to how BEAST works."

Undo

that'd work too

raz

19:19

@Polynomial It's more of a Man on the Side type situation. MitM would be full on decryption of your bits.

Polynomial

@raz erm... ok? it's not really that at all. you have to be in the middle to modify it. MitS would be sniffing mainly.

AJ Henderson

@raz yeah, what poly said, you need to be able to modify to get the failure

and it's active

raz

@Polynomial But you're also just forwarding packets on to the server right?

AJ Henderson

man on the side would be passive

as I understand it you are fiddling with the padding

raz

@AJHenderson Right I see that

I guess I misunderstood MitM

It's a bit more broad than I thought it was.

AJ Henderson

19:24

mitm just means you have to be able to modify stuff actively

you don't actually have to play host

it's just one of the most common ways as that is easiest if authenticity is not established

Rоry McCune

@Polynomial was it not a bit too long... surely "ask the bear" would have been more apposite :oP

Polynomial

@RоryMcCune shrug fair enough

Rоry McCune

@Polynomial more seriously looked good to me but then I am a crypto n00b

Xander

@Polynomial Looks good to me. Tiny things: "Connection uses SSLv3" Might should include "and a CBC ciphersuite" (or block, since all block ciphersuites in SSLv3 use CBC) I know the rest of the article makes this obvious, but since SSLv3/RC4 isn't vulnerable, could be worth mentioning in the prerequisites list.

Polynomial

@Xander ah yeah I should change that over.

Xander

19:32

@Polynomial Only other thing which you may or may not want to add would be a brief mention of the server-side component of TLS_FALLBACK_SCSV

AJ Henderson

@Xander that's not really a mitigation though

and it isn't supported by most browsers apparently

rather than support it, Firefox is just disabling SSL3

Xander

@AJHenderson Yeah, that's why I was ambivalent on whether it should be added or not.

AJ Henderson

after reading more on that feature, it is really more of a Google brag

that they are the only browser that doesn't ignore it

Jeff Ferland

Buying high and selling low....

AJ Henderson

and it looks like anyone who happens upon my website with IE6 is just going to be doomed to negotiation failures

looks pretty confirmed that the data about the cipher suite doesn't make it out of the SSL library, even if cert details do

Thomas Pornin

19:42

@Polynomial Looks good. You may want to insist a bit more on the specific SSL 3.0 feature, i.e. that the padding bytes are ignored; and say that this is fixed in TLS 1.0 (published January 1999, so not exactly new).

Polynomial

@ThomasPornin sweet. danke :)

Thomas Pornin

@Polynomial To be fair, what Oracle did (set all bytes to 0 except the last one) is not exactly a mistake; the SSL 3.0 specification allows it.

If you unfairly attribute mistakes to Oracle, this may summon Larry Ellison.

Polynomial

@ThomasPornin true, though doesn't it specify that the bytes should be set? (i.e. not MUST but SHOULD)

@ThomasPornin Is that like Bloody Mary? Say "Oracle sucks" 3 times in the mirror and BAM! Larry Ellison eats your face.

Thomas Pornin

@Polynomial As far as I can tell, RFC 6101 says absolutely nothing about the value of the padding bytes.

Gilles

@Polynomial worse: if you unfairly attribute mistakes to Oracle, they charge you for a license

3

Polynomial

19:46

@ThomasPornin yeah I noticed that, but then I google'd and found another RFC that spec'ed it for block ciphers in SSL

@Gilles Holy shit that's brutal.

@ThomasPornin it noted that a "PKCS#7 compatible padding" was preferred

Thomas Pornin

@Polynomial If you have a reference I'd be happy to have a look at it.

Polynomial

@ThomasPornin I'll see if I can find it again after dinner. Mmmm, ravioli!

Polynomial

20:11

@ThomasPornin ah, my bad, I was looking at an RFC for IPsec ESP

must've missed the title when I opened it

Simon

@Polynomial You donut.

Polynomial

@Simon Appropriate!

Simon

@Polynomial Don't tell me you're actually eating donuts?

Polynomial

@Simon Not currently.

Rоry McCune

@Simon he's in the UK ... we don't eat those things here... doughnuts on the other hand :op

Simon

20:21

@RоryMcCune Doughnuts sound disgusting, ewwww.

Thomas Pornin

@Polynomial I had not realized it, but until today I had no idea what you real name was (insofar as you used your real name for your twitter account).

Polynomial

@ThomasPornin heh, well now you know!

and knowing is half the battle!

Simon

@Polynomial What's the other half, doing?

Polynomial

@Simon Depends on the context, but it could also be getting attacked by bees.

2

Xander

@Simon No you donut, it's not knowing. Half the battle is knowing, and half is not knowing.

Simon

20:27

@Polynomial That makes sense!

@Xander That makes no sense.

Xander

@Simon That's because you're a donut.

Simon

@Xander I am what I eat.

Polynomial

@Xander Partial ignorance is partial bliss!

Xander

@Simon And clearly, given that you're confused about what constitutes the components of the battle, you're losing the battle as well.

Gilles

65

Q: How do I patch/workaround SSLv3 POODLE vulnerability (CVE-2014-3566)?

After BEAST, Heartbleed now I've heard about a new vulnerability in SSL/TLS called POODLE. How do I protect myself against being exploited? Are only servers or also clients affected? Is this OpenSSL/GnuTLS specific? What kind of services are affected? Only HTTPS or also IMAPS, SMTPS, OpenVPN, e...

security ssl tls

@Xander and knowing that you don't know is half the peace?

Polynomial

20:46

Anyone wanna join me in some headbanging to Nordic viking music?

►

VALHALLAAAAAAAAAAAAAA

And stick with it 'till the chorus.

Gilles

21:08

1. Define a one way trapdoor function, 2. fall through it :) — owlstead 4 hours ago

(Question: How would I write my own asymmetric encryption algorithm (not RSA, DES etc)?)

Simon

21:33

@Xander Troy Hunt's article you just rt'd is pretty cool.

Xander

@Gilles Transmogrify yourself into @ThomasPornin, define Sosemanuk. Oh wait, that wasn't from scratch, it was build on Serpent and Snow. Nevermind. You're screwed.

@Simon The Whitehat article? Yeah, it's fun. :-)

Simon

@Xander I guess that's vulnerability scanners are testing too.

Xander

@Simon Yeah, definitely.

Adnan

21:58

YES!!

We have a PoC!

Should I submit it somewhere?

(It's for the Drupal thing)

Oh.. fuck. Looks like I'm not the only one :(

Oh well. Others haven't went public with theirs, so what the hell, why not

POST /?q=node&destination=node
name[0]=nothing&pass=nothing&name[1 ;DELETE FROM users;%23]=garbage&form_id=user_login_block&op=Log+in

Simon

That's a fun command to inject.

Adnan

22:16

Oh, looks like somebody posted another PoC with a query to reset the username and password for the user 1 (god mode) reddit.com/r/netsec/comments/2jbu8g/…

@Simon This one is more fun :D

Simon

@Adnan Haha, nice.

Adnan

I used Tamper Data addon in Firefox to inject, by the way.

Headers captured by a honeypot pastebin.com/vejBF9xN

I've now made an automated tool to test for the vuln.

This is how it looks when you succeed, by the way.

Simon

22:37

Cute.

Adnan

I've managed to trigger this in other parts of Drupal, but I'm not sure if I should go public with it yet

Jeff Ferland

23:24

@Adnan CVE! CVE!

Transcript for