Assignment 6: Passwords Solution

Introduction




This assignment explores a few di erent aspects of the password and authentication ecosystems. In particular, you will observe how passwords are stored on Linux systems, gain experience crack-ing passwords using existing password-cracking software (albeit in con gurations you will have to choose), leverage the information you gain in a rate-limited online attack, and take advantage of a side channel to learn more about a system.




In this assignment, the ags you will be aiming to capture are passwords for di erent users. Before getting to the problems, we discuss the rules for this assignment and how you can access




the assignment infrastructure for your attacks.




Rules




Collaboration policy. Please respect the following collaboration policy: You may discuss the problems with up to 3 other students in the class, but you must write up your own responses and your own code. You should never see your collaborators’ writing, code, ags (passwords in this assignment), or the output of their code. At the beginning of your submission write-up, you must indicate the names of your (1, 2, or 3) collaborators, if any. You may switch groups between assignments but not within the same assignment.




This should go without saying, but you should not capture anyone else’s ags.




Sources. Cite any sources you use. You may Google liberally to learn basic Python, learn how to navigate Linux, and learn how to use Hashcat. Unlike some past assignments, you are welcome to perform additional Google searches on other topics related to this assignment, but be warned that these are unlikely to help you much. You must, however, note at the end of your solution for each task any topics you ended up Googling to complete that task.




Campuswire. We encourage you to post questions on Campuswire, but do not include any of your code in the Campuswire posts. If you have a question that you believe will reveal secrets you have discovered while working on the assignment, post privately to just the instructors. If you have a question that you believe will be of general interest or clari es the assignment, please post publicly. If you are uncertain, post privately; we will make public posts that we believe are of general interest.




Outside attacks. Your attacks for this assignment should be those discussed below. Do not attempt to compromise our server, sni your classmates’ network tra c, or do other nefarious




1



things. You will not receive credit for breaking into the server. In fact, you will lose credit for doing so.




Grading. Responses will be graded for correctness and clarity.




Assignment Tech Set-Up and Overview




This assignment begins (Problem 1) on your VM from the previous assignment and ends with sending HTTP queries to our server (Problems 2-3).




For Problems 2 and 3, which require that you complete Problem 1 rst, you will be making queries to a domain that we surprisingly now own: https://uchicago.computer/. We have provided sample code for using both Python 2 and Python 3 to make queries to the server with a make query function. Di erent from past assignments, our queries to the server are not in Base64, but rather re ect basic URL-safe quoting (and unquoting on the server side).







As in past assignments, you may implement your code in whatever programming language you choose, although we recommend Python, for which we can provide the most support. You will be submitting your code (when applicable) alongside descriptions of what you did, as detailed in each task below.




What and How to Submit




You will submit a set of les:




A le <YOUR CNETID-writeup.pdf/txt that brie y describes how you solved each problem. This le is shared across all four problems.



The individual les (code and/or results) requested by each of the problems, detailed below.



You will upload these les to Canvas. You may zip/tar/gzip/7zip/etc. them or submit them individually.




Final Notes and Hints




Please don’t DDOS the server. It will support thousands of queries per student (which you will need), but not millions.




This assignment was built by Blase last year for this class, and it was modi ed this year based on feedback from last year. Please let us know if you nd bugs or need instructions eshed out by posting on Campuswire.




Problem 1: Password Cracking




200 of your closest friends have somehow created accounts on your VM. Well, now that they’re here, you might as well gure out their password.




On your VM, there is a le /etc/passwd that anyone can read, but this is not the le you want; you want /etc/shadow. Why? See https://en.wikipedia.org/wiki/Passwd. The problem is that if you try to access this le as a normal user, access will be denied. However, since we gave you sudo access for the last assignment, you can become a superuser and access this le. You now




2



have the /etc/shadow le. You will notice inside the le that there are 200 users with usernames that start with your CNetID followed by an underscore. We will refer to these 200 usernames as the users assigned to you. Your rst step for this problem is to extract only the lines corresponding to the 200 users assigned to you from the shadow le; there are a handful of accounts made by the system for various purposes.




Unfortunately, this shadow contains hashes, yet you want (plaintext) passwords. Thus, you’ll need to crack the passwords! You will probably want to download Hashcat and install the appro-priate GPU drivers (if you have an appropriate GPU); it also works on CPUs. It supports Linux, OS X, and Windows.




When you get a set of hashes, the rst step is to gure out what hash function was used. Typically, you would use an online version of the Hashtag script (http://www. onlinehashcrack.com/hash-identification.php) to identify possibilities. For instance, try 5f4dcc3b5aa765d61d8327deb882cf99, which is "password" hashed with MD5. Once you know which (or which set of) hashes are possible, you will need to know the numerical hash function parameter (the -m parameter) for Hashcat from their list (https://hashcat.net/ wiki/doku.php?id=example_hashes). In this case, though, you could also learn how the hashes are stored by simply examining the shadow le (see https://www.cyberciti.biz/faq/ understanding-etcshadow-file/). Before you try to crack any hashes, you will want to extract just the hash portion of each line and feed just these hashes to your cracking software. The examples below assume you saved this le as hashes.txt.




You can run Hashcat on your machine or your VM. To run Hashcat, go to a command prompt, navigate to the directory where the les are, and call the appropriate binary. For example, on Ubuntu, I call ./hashcat64.bin.




This won’t do anything, though. You need to point Hashcat towards the le of hashes you want to crack (giving it the full path if it’s not in the same directory as your Hashcat executable). You must also specify an attack mode. Your successful cracks will appear in the hashcat.pot le in the same directory as Hashcat, though you can change the output le using the -o option.




Here are some sample attack modes:




./hashcat64.bin hashes.txt -m 100 -a 3 ?l?l?l?l?l?l?l



try to crack hashes in hashes.txt that are hashed with SHA1 (-m 100) using the brute-force/mask mode (-a 3) of Hashcat, trying all 7-character strings of only lowercase letters




./hashcat64.bin hashes.txt -m 100 -a 0 pw.txt



try to crack hashes in hashes.txt that are hashed with SHA1 (-m 100) using the wordlist mode (-a 0) of Hashcat, drawing its guesses from the le pw.txt (which you would have to provide)




./hashcat64.bin hashes.txt -m 100 -a 0 -r ./rules/best64.rule pw.txt



try to crack hashes in hashes.txt that are hashed with SHA1 (-m 100) using the wordlist mode




(-a 0) of Hashcat, drawing its initial words from pw.txt...and also mangling those entries with the Best64 mangling rules




Note that you’ll need to edit the sample commands above to re ect the proper binary for your operating system, specify the mode (-m) that actually applies for the hashes you have, provide valid paths to the rulelist (if applicable) and wordlist (if applicable) you want to run, and so on.



















3



Here are some initial resources that may help, but you will almost certainly want additional wordlists and/or rulelists.




Lots of password breaches/other word lists: https://wiki.skullsecurity.org/Passwords



Additional wordlists: http://contest-2010.korelogic.com/wordlists.html



And this is how you get the password.




That said, you do not have to guess all passwords to receive full credit! If you successfully guess the passwords for 80 of the 200 users assigned to you who do have accounts on the server, and your code and write-up are su ciently descriptive, you can receive full credit for this section. If you nd as much joy in cracking passwords as we do, try to crack as many as you can. We’ll gure out some reward (snacks, etc.) for the people who crack the most.







What to submit. You do not need to submit any code for this problem.




You must submit the passwords you crack in one of two formats. If you used Hash-cat, you can submit the .pot le that Hashcat outputs to store successful cracks as <YOUR CNETID-problem1.pot. If you don’t use Hashcat (or if you really want to reformat the out-put for whatever strange reason), you may instead submit a le <YOUR CNETID-problem1.txt that includes the usernames and corresponding passwords you successfully guessed. Each line of this le should contain a username, followed by a tab character (\t), followed by the plaintext password that was a successful guess. Do not include the usernames of users whose passwords you did not successfully guess.




In your shared write-up le, <YOUR CNETID-writeup.pdf, describe your approach to solving this problem. In particular, be sure to note what machine you used, what con gurations of Hashcat you tried (word lists, rule lists, etc.), and comment brie y on the success or failure of the main cracking strategies you employed.




Problem 2: Using Side Channels to Identify Who Has An Account




In Problem 1, you discovered 200 accounts on the VM that were assigned to you (began with your CNetID). It turns out that some of them, but not all of them, also have accounts on uchicago. computer . You can try to log into the server as a given user by making a query to https:// uchicago.computer/<USERNAME/<PASSWORD/, replacing <USERNAME and <PASSWORD with the values you want to test. We have once again provided sample code for using both Python 2 and Python 3 to make queries to the server with a make query function. If you choose not to use this function, note that the username and password should both be percent encoded (see https:// en.wikipedia.org/wiki/Percent-encoding) as performed by the urllib.parse.quote function in Python3.







Use information that leaks from the server to determine who (among the 200 usernames assigned to you) does, and who does not, have an account. Note that time is of the essence.







What to submit.




Include a code le <YOUR CNETID-problem2.py (or substitute an appropriate extension for the language you used) that includes any code you wrote to solve this problem.




Include a le <YOUR CNETID-problem2.txt that includes just the usernames, one per line, of the users assigned to you who do have accounts on the server. If a user from Problem 1







4



appears not to have an account on the server, do not include them in this le.




In your shared write-up le, <YOUR CNETID-writeup.pdf, brie y describe your approach to solving this problem and hypothesize why this information about who has accounts is leaking. Subsequently, discuss how to x this leak.




Problem 3: Online Attacks (20 points)




In Problem 2, you identi ed that some of the users who had accounts on the VM also have accounts on uchicago.computer, whereas others do not. Problem 3 concerns only the users assigned to you who do also have accounts on uchicago.computer. You will be attempting to log into the server as them following the same query format as in Problem 3. Note that the server returns either \Success" or \Failure" for each query you make that is not timed out. You can observe this in action for the user \student" whose password for the server is \1234".




Note that this server employs conservative rate-limiting, so your strategy of making lots of guesses in Problem 1 will not work here. In particular, you may only make one guess every ve minutes against each account. If under ve minutes have elapsed since your last guess, the server will return \Timeout" instead. Thus, you’ll need to make every guess count. Hint: what you learned in Problem 1 about each user will help you greatly with this task, though sometimes a little twist is necessary. Note that you may not attempt to make guesses against any of your classmates’ accounts in this section. We will check the server logs!




You do not have to guess all passwords to receive full credit! If you successfully guess the passwords for 30% of the users assigned to you who have accounts on the server (i.e., 30% of the people from Problem 2), and your code and write-up are su ciently descriptive, you can receive full credit for this section.







What to submit. Include a code le <YOUR CNETID-problem3.py (or substitute an appropriate extension for the language you used) that includes any code you wrote to solve this problem.




Include a le <YOUR CNETID-problem3.txt that includes the usernames and corresponding passwords you successfully guessed. Each line of this le should contain a username, followed by a tab character (\t), followed by the plaintext password that was a successful guess. Do not include the usernames of users who do not have accounts on the system. If you did not guess a particular user’s password, don’t include them in this le.




In your shared write-up le, <YOUR CNETID-writeup.pdf, brie y describe your approach to solving this problem.

















































5