by mr96
These are two very similar challenges: the first one has a bug so that you can solve it in an unintended way, so the organizers realized a second “fixed” challenge. The following method works for both of them.
We’re given a web page that gets a random integer and asks us to guess it, multiple times. After guessing correctly 20 times, it will return a flag.
From the title we can see that the prng is of lcg type (linear congruential generator), that is, given 3 integers \(a,b,n\) and a seed \(x_0\), the sequence is constructed as \(x_i=ax_{i-1}+b \pmod{n}\). It turns out that the website tells you what value it randomized even if you fail, so observing about 10 values we can break the prng as follows:
We did this using the Python script from msm from p4team and then did it on the website by hand (input 20 times is faster than scripting). Here’s the code:
from functools import reduce
from gmpy2 import gcd
def egcd(a, b):
if a == 0:
return (b, 0, 1)
else:
g, x, y = egcd(b % a, a)
return (g, y - (b // a) * x, x)
def modinv(b, n):
g, x, _ = egcd(b, n)
if g == 1:
return x % n
def crack_unknown_increment(states, modulus, multiplier):
increment = (states[1] - states[0]*multiplier) % modulus
return modulus, multiplier, increment
def crack_unknown_multiplier(states, modulus):
multiplier = (states[2] - states[1]) * modinv(states[1] - states[0], modulus) % modulus
return crack_unknown_increment(states, modulus, multiplier)
def crack_unknown_modulus(states):
diffs = [s1 - s0 for s0, s1 in zip(states, states[1:])]
zeroes = [t2*t0 - t1*t1 for t0, t1, t2 in zip(diffs, diffs[1:], diffs[2:])]
modulus = abs(reduce(gcd, zeroes))
return crack_unknown_multiplier(states, modulus)
obs = [9933, 16949, 696, 30323, 33563, 25061, 14546, 13243, 12116]
n,k,d = crack_unknown_modulus(obs)
state = obs[-1]
while True:
state = (state*k+d) % n
print(state)
input()
crypto
prng
lcg