Raymond James CTF

I went to Tampa, Florida last weekend to participate Raymond James CTF. My team got 3rd place with $2500 award.
The weather in Florida is so0O gO0od: 24 degree C, meanwhile it’s like 3 degree C in Baltimore.

The team photo:
My eyes were closed lol.

The trophy:

The coin from gam3z:

 

The onsite-CTF was 70% forensics, 20% binary reverse and 10% Misc. I’m not good at forensics so I didn’t contribute much on that. However I earn 185 out of 215 points from the pre-challenge by solving a binary reverse problem. Those points from pre-challenge give our team huge advantage during the on-site CTF.


Write up for the binary reverse problem from pre-challenge(without IDA):

Step 1: Understand the program

Use the password in the email from t0k3nz@gam3z-inc.com to extract the binary from the Zip file.

file Ghost_Protocol_Access

Ghost_Protocol_Access: ELF 64-bit LSB executable, x86-64, version 1 (GNU/Linux), statically linked, for GNU/Linux 3.2.0, BuildID[sha1]=cf8c2aebbde52c274e2f71dc3a83bb63e4a629e3, stripped

I copy the binary to my Ubuntu VM and run the binary:

The program print out two messages:
"This version of bash is terribly out of date....your environment is not secure and this program will not run."
"You're not even authorized to use this program!"

I load the binary in Hopper Disassembler(https://www.hopperapp.com/) and search for the string: “This version of bash is terribly out of date....

The string was used in function: sub_401c99. I then load the binary in radare2, I prefer reading assembly code in r2 because it looks nicer:
r2 Ghost_Protocol_Access
aaa
s 0x401c99
pdf

The program check for things like “gdb”, “r2”, and “bash”. If “gdb” or “r2” is found, the program will print out string:
"Seriously....you're trying to debug and disassemble me?!?\n ....ridiculous....I don't even know what to say.....so disappointed...."
If “bash”  is found, the program will print:
"This version of bash is terribly out of date....your environment is not secure and this program will not run."

Now I understand the function “sub_401c99” does some environment check and exit the program. My next step is to find out where this function gets called in the program.
I reload the binary in radare2:
r2 Ghost_Protocol_Access
aaa
pdf

Noticed the function address “0x401c99” was one of the function argument for the next function call(sub.libc_start_main_40). To display the instruction in function sub.libc_start_main_40 in r2, I used following command:
s sub.libc_start_main_40
pdf

I can’t really understand what the program is doing inside the function “sub.libc_start_main_40” by reading the assembly code only, I need to run the program in GDB to understand the program flow.
gdb ./Ghost_Protocol_Access
The program starts at address 0x004015f0:

 

I put break point at the beginning of the program:
tbreak *0x004015f0
run

Run the program line by line(in assembly) in GDB with “next” and “step”. Pressing “n” in gdb for a while and the program stop at address: 0x402347:

whats in eax? It’s the function(sub_0x401c99) I mentioned before, the function that checks environment and quit the program.

At this point I understand after entering the function “sub.libc_start_main_40“, the program call “sub_0x401c99” and the program die afterword.

I am not sure what to do at this point, I just randomly do a “strings Ghost_Protocol_Access | grep t0k3n” and got some interesting result:

Since I know one of the token named “acc3ss”, I start searching the string “acc3sst0k3n” in the program. I search the string in r2 with command "/ $string":

In hopper, it shows the strings was used in function sub_401987:

I notice the function sub_401987 doesn’t get call in the program:

Hopper has the feature to show where a function get called, if a function get called, it will show something like this in hopper:

I want to see what function sub_401987 does, my plan is to patch the program and let the function sub_401987 run.


Step 2: Patch the binary to get the Token(Flag):
First Patch:
Remember at the beginning of the program, the environment checking function: sub_0x401c99 was moved into the register and get called in “sub.libc_start_main_40” function. I first patch the binary to change the value from sub_0x401c99 to sub_401987, so sub_401987 will be called afterwords.

The opcode 48c7c7991c40. means 48c7c7991c4000, the small dot in the end represent two zero “00”. I use this website to convert between opcode and assembly: https://defuse.ca/online-x86-assembler.htm#disassembly2
I need the opcode for mov rdi, 0x401987:

Patch the binary in r2:
oo+(enable write)
s 0x0040160d
wx 48C7C787194000
q
Run the program again:

The program ask for a “Confirmation Code”, I am not sure what the code is so I enter some random string:


Second Patch:
Check the sub_0x00401987 function in r2:
s 0x00401987
pdf

There are two place call the function “call sub.You_re_not_even_authorized_to_use_this_program_70d”. The jump come from CODE XREF from main (0x401b42) and CODE XREF from main (0x401b0d)
Both section does the same thing: call a function, compare the output to a variable, jump if not equal:

I patch the jump instruction to let the program can keep running without doing the jump:
oo+
s 0x401b42
wx 909090909090
s 0x401b0d
wx 909090909090
q
Change jne instruction to nop:

 

Run the program again to get the flag:


Step 3: Fix the image file:
After the program print our both token, it also output a file called: “acc3sst0k3n”:

My teammate told me this file can be base64 decode to a image, but the file is somewhere corrupted and they can’t the flag in the image.

I did remember there is a super long string in function sub_0x0401987:

I am not sure what was this at the beginning, I went to address “0x00496498” in r2 and type:
s 0x00496498
ps 100000

In r2 it output huge amount of base64 encoded string, the image below shows the end of the string. 

I search for "XXH+Kk" in “acc3sst0k3n” file and found a “NUL”.

I delete the "NUL" and run the following command to convert base64 encoded file to a image:

base64 -D acc3sst0k3n > a.png

View the image:

 

One thought on “Raymond James CTF

Leave a Reply

Your email address will not be published. Required fields are marked *