In each lab, you are asked to solve a set of challenges (typically 10 challenges, except for the first two weeks). For each challenge, you have to submit three things to the submission site (-A section, -O01/-OCY section): the flag you captured from the challenge, the exploit that you wrote to get the flag, and a write-up that summarizes how you formulated the exploit (see below).
A flag is a 512-byte hex string (like shown below)
/proc/flag on the CTF server.
$ cat /proc/flag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
Your job is to read this flag by exploiting the provided challenges.
Taking actions #1: registration¶
Register your account and get an api-key:
Input your username (preferably your real name), GTID number, and e-mail address (we only accept @gatech.edu accounts), and click the Email api-key button.
You will receive an email with your api-key. You can then use this to log in to the course website.
Taking actions #2: course workflow¶
Log in to your account:
Go back to the submission site to use the Login menu.
Input your received api-key and click the Submit api-key button.
You can now connect to the CTF server and begin your lab challenges. You can find ssh details for each lab on Canvas after they unlock.
# log in to one of the CTF servers # ** Refer to Canvas for ssh information! ** [host] $ ssh lab01@ctf_server_address # let's start lab01! [CTF server] $ cat README [CTF server] $ ./bomb
While solving the challenges on the CTF server, you might need to write your own exploit scripts. Feel free to create a directory inside
/tmpto store your files. Note that the CTF server and the lab account are shared with all other users, so avoid using directory names that are easily guessable (e.g.,
/tmp/cs6265), and do not reveal your secret directory name anywhere. Listing files (
/tmpis not allowed. Also, we encourage you to keep backups of your files by copying them from the CTF server to your local machine, since
/tmpmight get emptied occasionally.
Submit your flags and write-ups.
Login → Problems → (lab) → Submit (Flag / Writeup)
Remember that your score is only valid after you submit both a flag and a write-up.
Taking actions #3: building local environment (OPTIONAL)¶
Although you can solve all challenges on the CTF server, it may be
inconvenient because you cannot install your own tools.
In this case, you can build an environment
to study the challenges on your own machine.
Please note that you still have to read
on the CTF server to get the real flag.
To build your own environment:
Download and install VirtualBox:
Note: Ubuntu users may want to use the following command to install VirtualBox:
[host] $ sudo apt-get install virtualbox
Download and install Ubuntu 18.04.2 in VirtualBox:
Install essential tools inside the VM:
To do labs (lab03 and after):
After launching VirtualBox and your Ubuntu VM:
# disable ASLR (before lab05) [vm] $ echo 0 | sudo tee /proc/sys/kernel/randomize_va_space # enable ASLR (starting lab05) [vm] $ echo 2 | sudo tee /proc/sys/kernel/randomize_va_space # download labXX challenges [vm] $ scp -r labXX@server:~/[challenge] [your_directory] # download libc used by the server for analysis (future labs) # 32-bit [vm] $ scp labXX@server:/lib/i386-linux-gnu/libc-2.27.so [your_directory] # 64-bit [vm] $ scp labXX@server:/lib/x86_64-linux-gnu/libc-2.27.so [your_directory] # tackle challenges [vm] $ cd [your_directory_for_challenge] [vm] $ cat README ...
Feel free to ask for help on Ed Discussion or at office hours if you have any trouble with setup.
Points are earned as follows:
Each lab has one or two tutorial(s) and 10 challenges.
Each challenge is worth 20 points.
You must submit a flag, a write-up, and your exploit to earn the points.
If you do not submit the write-up, you get 0 points for that challenge.
As an exception, you do not need to submit write-ups for the tutorial challenges.
The maximum points you can get in each lab is 220 if it has one tutorial (20 × 11 flags), or 240 if two (20 × 12).
Bonuses: starting with lab 3, first and second bloods (i.e., fastest solvers) for each challenge will get 2 and 1 bonus points, respectively, for that challenge.
Hints: we will provide up to two hints per challenge.
You can view them on the challenge page (Submission site → Lab N → Problem name → Hints: Show).
Hints cost one point to open.
Later labs generally have few or no hints available.
Late policy: challenges submitted within one week past the due date will receive 50% of the original points.
You must submit a write-up for each challenge to get the actual points.
Your write-up must be in plain text formatted with Markdown syntax:
Your write-up must contain
a simple description of how you solved the challenge, and
the actual exploit (i.e., code!).
Your write-up must be specific enough that we can run and validate your exploit ourselves.
You don’t need to submit write-ups for the tutorial challenges.
Here is an example write-up:
1) Write-up: This binary stores its return addresses and frame pointers in a separate stack; while we can overwrite those values on the regular stack, it accomplishes nothing. `main()` puts a pointer to `exit()` in its stack frame, which it jumps to after `start()` finishes. We can overwrite this during `start()`'s buffer overflow to instead point to `jump_to_here()`, which prints the flag. Conveniently, due to the return-address protection and the fact that `start()` doesn't have many meaningful local variables, we don't need to be careful about the other values we clobber on the stack on the way to `main()`'s local function pointer. The correct return addresses are restored at the end of each function. 2) Exploit: #!/usr/bin/env python3 from pwn import * p = process('./target') p.sendline(cyclic(52) + p32(0x8048948)) # (jump_to_here()) p.interactive()