Category Archives: Technical Insight

Oil Droplets and Your Banking Credentials

Warning: IANAQC (I am not a quantum cryptographer)

What does a droplet of oil have in common with the security of your banking credentials? Very little, you might think. However, there is research that came out a few months back, that confirms a theory made (and then dismissed) over 80 years ago about quantum effects. Bear with me.

For the last 80 years or so people have believed that particles are in two or more places at once, and only once measured do they “choose” a position and lock in. By virtue of being in two or more places at once, it is believed that a quantum computer can test all theories in a binary question simultaneously (on or off). Each binary question is effectively one bit of entropy, and if you get enough bits, you can build a very powerful computer.

From a computer security perspective it means that factoring large primes is a relatively easy thing to do – you get enough bits into your machine and you can factor the largest publicly available crypto-systems. While it is believed by the likes of Dr. Martin Helman (of Diffe-Hellman-Merkel key exchange infamy) that we have 10 years before such a machine is feasible and an additional 10 years before such a machine is usable, that still brings the time horizon of quantum cryptology into our lifetime.

That’s a scary thought if you have secrets that need to live beyond your lifetime – you only have 20 years before they are breakable by the military — or by anyone who could afford such a device in their evil lair; the only caveat would be the data collection and storage space for all that encrypted data.

However, recent findings suggest that particles might actually not be in multiple positions all at once; they might instead act like a droplet of oil dancing along the surface of a pool of water. Unlike a droplet of water, an oil droplet won’t go beneath the surface of the water (because of differing densities for oil and water). Instead, the oil droplet will bounce along on the surface. But when the oil droplet is first dropped, it causes a ripple, and that ripple will bounce around and could actually interact with the oil droplet again, causing it to move seemingly erratically.

So perhaps quantum particles, like oil, do not behave in “spooky” ways, but rather in very deterministic (as opposed to probabilistic) ways. That is to say that there may be no “magic” behind how particles behave – it may just be a very challenging fluid-dynamics problem. If we knew enough about the waves and the oil we might be able to predict exactly where the oil (or particle) would end up. Okay, but what does this have to do with your banking password?

If this theory is indeed true, cryptographers might be unable to build a quantum computer capable of being in a super-position (two positions at once), and therefore capable of factoring all possible variations at once in the way once envisioned. If that is true, we could be much safer in the near-term as our secrets stay safe from such a machine. That means that crypto-reliant technologies like SSL/TLS might actually have some greater longevity than previously thought (barring things like Poodle, BEAST, CRIME, etc.).

Although this is an unconventional theory, one that is not at all agreed upon by the scientific community (yet), and a difficult one to prove at that, it might make your banking passwords safe for a little longer than we previously thought. Who knew? Oil. Hmm! Read more about it in Wired.

#HackerKast 5: POODLE Attack, HackerKombat and Drupal SQLi Flaw

This week Jeremiah Grossman, Robert Hansen and Gabe Gumbs host HackerKast at Levi’s Stadium – the home of the SF 49ers – to discuss the recently announced POODLE Attack on SSL 3.0 and a critical SQLi flaw affecting Drupal making headlines. WhiteHat’s 6th HackerKombat capture the flag competition will also stream LIVE on Twitch.tv.

Watch HackerKombat LIVE starting at 3 pm PT on 10/17:
http://www.twitch.tv/hackerkombat

Other Resources:
POODLE Attack Information:
https://blog.whitehatsec.com/what-you-need-to-know-about-poodlessl-3-0-vulnerability/
http://googleonlinesecurity.blogspot.com/2014/10/this-poodle-bites-exploiting-ssl-30.html
https://www.openssl.org/~bodo/ssl-poodle.pdf

Drupal SQLi Flaw Advisory:
https://www.drupal.org/SA-CORE-2014-005
http://news.techworld.com/security/3581251/drupal-releases-patch-for-severe-sql-injection-flaw/?olo=rss

What you need to know about POODLE/SSL 3.0 vulnerability

UPDATE – 10/16 12:45 p.m. PT: For users with Akamai sites, Akamai has made the following updates:

  • Akamai is going to be disabling SSLv3 and SSLv2 support on an aggressive timeline
  • If SSLv3 support is necessary for legacy support clients can be exempted upon request

  • Users that utilize Akamai should contact Akamai for further details. Here are some Akamai blogs which clients may find helpful:
    SSL is dead, long live TLS
    Excerpt: How POODLE happened

    UPDATE – 10/15 7:15 p.m. PT: WhiteHat Security has added testing for the new POODLE attack. These vulnerabilities will be shown as ‘Insufficient Transport Layer Protection’ in the Sentinel interface. They will have the description ‘CVE-2014-3566 – POODLE Attack’. These tests will be run at the start of a new scan.

    Google researchers released a new SSL vulnerability yesterday nicknamed “POODLE Attack.” POODLE, which stands for Padding Oracle On Downgraded Legacy Encryption, is an attack that targets SSL version 3.0 and allows interception and compromise of supposedly secured data.

    Only SSL version 3.0 is known to be effected by this exploit. Although SSL 3.0 is extremely outdated, connection failures will result in older versions of SSL being used in an attempt to establish connection. Attackers can leverage this and force connection reattempts with SSL 3.0.

    Disabling SSL 3.0 will fix the issue however unforeseen compatibility problems may exist on sites. The Google researchers recommended supporting TLS_FALLBACK_SCSV. It’s also important to note that RC4 encryption has no padding, and as such is not vulnerable to this specific attack – although RC4 is not exempt from known issues as well.

    WhiteHat Security is currently researching a check for the POODLE Attack and will implement it as soon as it is possible.

    If you want to protect yourself in your browser, as Robert Graham with Errata Security has suggested, disabling SSLv3 in browsers is easy. On Chrome, Chromium and Aviator, use the command-line flag –ssl-version-min=tls1, and on Firefox set security.tls.version.min to 1. Mozilla also has an add-on available for disabling SSL 3.0 in Firefox. If you choose not to do this, please make sure you avoid unknown wireless connections until an official update is available for your browser.

    We will continue to update this blog as more information about POODLE is known and as more information for our customers becomes available. If you have any questions please contact WhiteHat Customer Support at support@whitehatsec.com.

    How I stole source code with Directory Indexing and Git

    The keys to the kingdom pretty much always come down to acquiring source code for the web application you’re attacking from a blackbox perspective. This is a quick review of how I was able to get access to a particular client’s application source code using an extremely simple vulnerability: Directory Indexing. Interestingly enough, they also had a .git repository accessible at https://www.[redacted].com/.git/ (although the ‘why’ still baffles me). If you have access to this you also have access to any commits and all logs that may exist in the repo.

    The following screenshots are from a recreation of the environment being run locally that I /etc/hosts mapped to http://demo.jkuskos.com. All client information has been redacted.

    image1 copy_Kuskos_10.14.14

    First, I confirmed that Directory Indexing was enabled. You’ll see why this is great in a moment.

    image2 copy_Kuskos_10.14.14

    The easiest way to download anything would be with a recursive wget(you simply need to set the flag -r).

    wget -r http://demo.jkuskos.com/.git/

    image3 copy_Kuskos_10.14.14

    Now let’s investigate. With the repository downloaded we can perform git commands on it.

    image4 copy_Kuskos_10.14.14

    Now that we can see which files exist in the repository, access to them is as simple as checking them out.

    git checkout *.php; ls;

    image5 copy_Kuskos_10.14.14

    This example is clearly simplified; however, the real site allowed me to find several SQL Injections and authorization bypasses that would have been cumbersome to find through dynamic blackbox testing alone. It also allowed me to find several files that would otherwise have been available only if you had the appropriate credential access. These types of flaws are easily found through static code analysis and much harder to find through a dynamic assessment only. As a hacker, turning a blackbox penetration test into a whitebox penetration test is always a victory.

    Shellshock Vulnerability – What It Is & Recommendations

    Shellshock VulnerabilityUPDATE – 9/26, 1:35 p.m. PT: Customers with WAFs (Web Application Firewalls), IPS’, and other security devices may have noticed that we have some checks already in place, with results / vulnerabilities coming out of the system. The nature of the Shellshock vulnerability requiring only a single http(s) request means that the number of attack vectors are numerous and as such we will be continuing to improve our testing methodology in the days and weeks to come. It is of the utmost importance that we reiterate the importance of checking your systems directly and patching as other services may be available such as SSH, CUPS and DHCP.

    UPDATE – 9/25, 5:00 p.m. PT: The WhiteHat Research & Development team has been working hard to dissect the Shellshock issue and deploy additional checks as necessary to Sentinel.

    Prior to the announcement of Shellshock, WhiteHat Sentinel Source had already been testing for applications making use of untrusted data in conjunction with the operating system’s shell interface to execute native commands and applications writing untrusted data to a system environment variable. In the Bash shell, injection into an environment variable can also lead to remote code execution. Failure to properly validate and or encode data utilized by the shell allows an attacker to execute arbitrary operating system commands. This is dangerous because environment variables can be used in other parts of the application, external process on the host, or even other applications. Many applications implicitly trust environment variables to be safe, so this data is often not checked for suspicious activity. Both of the checks in Sentinel Source are able to accurately identify the type of behavior that Shellshock is vulnerable to.

    The ‘Shellshock’ exploit (CVE-2014-6271) announced yesterday is a vulnerability found in the Bash command interpreter. Bash is the shell, or command language interpreter, whose name is an acronym for the ‘Bourne-Again Shell.’ Injection vulnerabilities in web apps are a death blow: they are the one class of vulnerability that accounts for more data loss than all other vulnerabilities. The Shellshock bug is a code-injection vulnerability that allows an attacker to pass commands to Bash to execute arbitrary code. This is a critical issue for any application that evaluates user input and calls other applications via a shell. The CVE severity score for Shell Shock is 10 on a scale of 1 to 10. Given that this vulnerability is known to be ‘wormable’ 10 almost seems like it is not high enough. This issue is likely to be of greater concern than Heartbleed (which we posted about here and here) was earlier this year.

    The extent to which this vulnerability affects the web is still unfolding. WhiteHat has confirmed that cgi-script based web applications may be vulnerable, especially those that call other applications via the shell. Apache servers using mod_cgi or mod_cgid are affected if CGI scripts are either written in bash, or spawn subshells. We have also observed several working pieces of exploit code in the wild that requires a minimal amount of technical expertise to execute. WhiteHat is implementing a detection for this vulnerability to identify the existence of this critical vulnerability in their web applications. At this time is highly advisable that you patch all systems running Bash. Additionally, there are several working mitigations currently available for this vulnerability:

    1. Upgrading to a new version of bash
    2. Replacing bash with an alternate shell such as zsh
    3. Limiting access to vulnerable services, or filtering inputs to vulnerable services

    Editor’s note: Want to learn more about Shellshock? Register for our town hall discussion.

    We will continue to provide regular updates as they become available.

    Other Resources for more information on this bug as it unfolds:
    GNU bash Environment Variable Processing Flaws Let Users Execute Arbitrary Code
    Shellshock DHCP RCE Proof of Concept
    [SECURITY] [DSA 3032-1] bash security update
    Bash specially-crafted environment variables code injection attack
    Bash ‘shellshock’ bug is wormable
    Everything you need to know about the Shellshock Bash bug
    Bash ‘shellshock’ scan of the Internet
    Quick notes about the bash bug, its impact, and the fixes so far
    Bash specially-crafted environment variables code injection attack
    Environment Bashing

    The Ghost of Information Disclosure

    Information disclosure is a funny thing. Information disclosure can be almost completely innocuous or — as in the case of Heartbleed — it can be devastating. There is a new website called un1c0rn.net that aims to make hacking a lot easier by letting attackers utilize Heartbleed data that has been amassed into one place.

    The business model is simple – 0.01 Bitcoins (Around $5) for data. It leaves
    no traces on the remote server because the data isn’t stored there anymore,
    it’s on un1c0rn’s server. So let’s play a sample attack out.

    1) Heartbleed comes out;

    2) Some time in the future un1c0rn scans a site that is vulnerable and logs it;

    3) A would-be attacker searches through un1c0rn and finds a site of interest;

    4) Attacker leverages the information to successfully mount an attack against the target server leveraging the data.

    In this model, the attacker’s first packets to the server in question could be
    the one that compromises them. But it’s actually more interesting than that. As I was looking through the data I found this query.

    un1c0rn

    For those of you who don’t live and breathe HTTP this is an authorization request with a base64 encoded string (which is trivial to reverse) that contains the usernames and passwords to the sites in question. This simple request found 400 sites with this simple flaw in it. So let’s play out another attack scenario.

    1) Heartbleed comes out;

    2) Some time in the future un1c0rn scans a site that is vulnerable and logs it;

    3) Site is diligent and finds that they are vulnerable, patching up immediately and switching out their SSL certificate with a new one.

    4) A would-be attacker searches through un1c0rn and finds a site of interest;

    5) Using the information they found they still compromise the site with the username/password, even though the site is no longer vulnerable to the attack in question.

    This is the problem with Information Disclosure – it still can be useful even long after the hole that was used to gather the data has been closed. That’s why in the case of Heartbleed and similar attacks not only do you have to fix the hole but you also have to expire all of the passwords, and remove all of the cookies or any other way that a user could gain access to the system.

    The moral of the story is that you may find yourself being compromised seemingly almost magically in a scenario like this. How can someone guess a cookie correctly on the first attempt? Or guess a username/password on the first try? Or exploit a hole without ever having looked at your proprietary source code or even having visited your site before? Or find a hidden path to a directory that isn’t linked to from anywhere? Well, it may not be magic – it may be the ghost of Information Disclosure coming back to haunt you.

    Spooky!

    SSL/TLS MITM vulnerability CVE-2014-0224

    We are aware of the OpenSSL advisory posted at https://www.openssl.org/news/secadv_20140605.txt. OpenSSL is vulnerable to a ChangeCipherSpec (CCS) Injection Vulnerability.

    An attacker using a carefully crafted handshake can force the use of weak keying material in OpenSSL SSL/TLS clients and servers.

    The attack can only be performed between a vulnerable client *and* a vulnerable server. Desktop Web Browser clients (i.e. Firefox, Chrome Internet Explorer) and most mobile browsers( i.e. Safari mobile, Firefox mobile) are not vulnerable, because they do not use OpenSSL. Chrome on Android does use OpenSSL, and may be vulnerable. All other OpenSSL clients are vulnerable in all versions of OpenSSL.

    Servers are only known to be vulnerable in OpenSSL 1.0.1 and 1.0.2-beta1. Users of OpenSSL servers earlier than 1.0.1 are advised to upgrade as a precaution.

    OpenSSL 0.9.8 SSL/TLS users (client and/or server) should upgrade to 0.9.8za.
    OpenSSL 1.0.0 SSL/TLS users (client and/or server) should upgrade to 1.0.0m.
    OpenSSL 1.0.1 SSL/TLS users (client and/or server) should upgrade to 1.0.1h.

    WhiteHat is actively working on implementing a check for sites under service. We will update this blog with additional information as it is available.

    Editor’s note:
    June 6, 2014
    WhiteHat has added testing to identify websites currently running affected versions of OpenSSL across all of our DAST service lines. These vulnerabilities will open as “Insufficient Transport Layer Protection” in the Sentinel interface. WhiteHat recommends that all assets including non-web application servers and sites that are currently not under service with WhiteHat be tested and patched.

    If you have any questions regarding the the new CCS Injection SSL Vulnerability please email support@whitehatsec.com and a representative will be happy to assist.