Extend Mt19937 Predictor
Project description
Extend MT19937 Predictor
Predict and Backtrack MT19937 PRNG by putting 32 * 624 bits generated numbers.
Python "random" standard library uses mt19937, so we can easily crack it.
Usage
Install
$ pip install extend_mt19937_predictor
Predict
After putting 32 * 624 bits numbers, the internal state is uniquely determined. And the random number can be predicted at will.
import random
from extend_mt19937_predictor import ExtendMT19937Predictor
predictor = ExtendMT19937Predictor()
for _ in range(624):
predictor.setrandbits(random.getrandbits(32), 32)
for _ in range(1024):
assert predictor.predict_getrandbits(32) == random.getrandbits(32)
assert predictor.predict_getrandbits(64) == random.getrandbits(64)
assert predictor.predict_getrandbits(128) == random.getrandbits(128)
assert predictor.predict_getrandbits(256) == random.getrandbits(256)
Backtrack
Besides prediction, it can also backtrack the previous random numbers.
import random
from extend_mt19937_predictor import ExtendMT19937Predictor
numbers = [random.getrandbits(64) for _ in range(1024)]
predictor = ExtendMT19937Predictor()
for _ in range(78):
predictor.setrandbits(random.getrandbits(256), 256)
_ = [predictor.backtrack_getrandbits(256) for _ in range(78)]
for x in numbers[::-1]:
assert x == predictor.backtrack_getrandbits(64)
Advanced
check
param is True by default. It is ok to put more than 32 * 624 bits numbers when initializing. It will automatically check whether the excess number is the same as the predicted number, and also change the internal state.
When setting check
param to False, it will directly overwrite the state without checking.
import random
from extend_mt19937_predictor import ExtendMT19937Predictor
predictor = ExtendMT19937Predictor(check=True)
for _ in range(1024):
predictor.setrandbits(random.getrandbits(32), 32)
for _ in range(1024):
assert predictor.predict_getrandbits(32) == random.getrandbits(32)
import random
from extend_mt19937_predictor import ExtendMT19937Predictor
predictor = ExtendMT19937Predictor(check=True)
for _ in range(624):
predictor.setrandbits(random.getrandbits(32), 32)
_ = predictor.setrandbits(0, 32)
# ValueError: this rand number is not correct: 0. should be: 2370104960
Besides "random" standard library function getrandbits
, these functions can be predicted.
random
randrange
randint
uniform
But only these functions can be backtracked, because of cannot determine how many times the base functions are called by the others.
random
uniform
Reference
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for extend_mt19937_predictor-19937.0.3.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | fb5c4b9bb12c8f2b5cdebaf3547f4e1cf7296ae334bee378d5b57b1706a9d09a |
|
MD5 | af0ad36ff47b3ed03c20c99e9b104e5b |
|
BLAKE2b-256 | 3f650aab1a7e0ba75a6bff5ef28fcfc15d3f3002f42da849800cbf61086a50b5 |
Hashes for extend_mt19937_predictor-19937.0.3-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2ef80db1f5a82547abbb1f3375a5566fe0f1935c12856ff260f02541e971cbdd |
|
MD5 | 2f9730fad9ca6a1774d2c346bf806b26 |
|
BLAKE2b-256 | 1535e63036a95392af21b5c8801a2ff164c68133d6075305d4a2a58566e90dc9 |