This problem is where the tediousness took off. I’ve spent about 5 hours on this. While I understood some of the hints, everything took a while to get right, especially when I started making assumptions. I used CyberChef to work out the solution manually, and so I don’t know if that’s considered cheating, and my filteration is a bit weak. I hope it won’t come back to bite me in the future challenges.

The Problem

A message has been XOR-ed against a single character, resulting in this hex encoded string:


Find the message. Importantly, don’t search for the string manually, let your code select the right one.

The Solution

After reading the problem statement, I didn’t know the format for the key. Questions I had were:

  • Was the key a “single character” as in the body of the problem statement?
  • Was the key a “Single-byte” as in the title of the question?

I decided to go with a single hex character for no other reason than the fact that it was what was given in the last question. However, I still got bad results. I ran the problem through CyberChef and saw the solution, but I still couldn’t reproduce it in Python.

Until now, I’d been passing the single character as a decimal value, but now I decided to convert it to a hexadecimal string $k$ and try checking if a key $K$ with following three formats works. Note that $n$ is the length of the input string:

  1. $K \leftarrow (\{0\}^{n-1} \mathbin\Vert k)$
  2. $K \leftarrow (k \mathbin\Vert \{0\}^{n-1})$
  3. $K \leftarrow \{k\}^{n}$

The third format worked, when I manually scanned the output and I was left with filtering for the right solution. I used two phases:

  1. I filtered the bytearrays that could be converted to a native string without raising an error
  2. I filtered the strings that contained common English phrases

I created the common English phrases by hand, and it isn’t well done. I realize most common letter in the English language is e, but that kind of frequency analysis wouldn’t work for this phrase because the most common letter here was an o.

The decoded string is Cooking MC's like a pound of bacon, and the key in hexadecimal is 58


import binascii

input = "1b37373331363f78151b7f2b783431333d78397828372d363c78373e783a393b3736"

def xor_hex_and_char(input_hex: str, guess_character_decimal: int) -> bytes:
    """XORs a hexadecimal numbers and a single character decimal number

        input_hex (str): an input string in hexadecimal format
        guess_character_decimal (int): a single decimal value betwen 0 and 255 inclusive

        bytes: a bytestring
        str: string
    input_decimal = int(input_hex, 16)  # to base10

    # convert the guess character into a repeating fixed length string
    guess_character_hex = f"{guess_character_decimal:x}"  # to base16
    guess_hex = guess_character_hex.rjust(
        2, "0"
    )  # ensure it's two characters, e.g. `a` -> `0a`
    guess_hex = guess_hex * (len(input) // len(guess_hex))  # repeat until
    guess_decimal = int(guess_hex, 16)  # to base10

    result_decimal = guess_decimal ^ input_decimal  # XOR
    result_hex = f"{result_decimal:x}".rjust(len(input), "0")

    return binascii.unhexlify(result_hex), guess_character_hex

def filter_strings(possible_answers):
    """filters bytestrings for those that can be decoded into native strings"""
    sentences = []
    for (answer, key) in possible_answers:
            answer = answer.decode()
            sentences.append([answer, key])
        except UnicodeDecodeError:
    return sentences

def check_phrases(sentence):
    """checks if a sentence contains common phrases in the English language

        Boolean: True if it does, false if not

    # A list of common phrases in the English language
    COMMON_PHRASES = [" a ", " of ", " an "]

    if any(phrase in sentence[0] for phrase in COMMON_PHRASES):
        return True
    return False

if __name__ == "__main__":
    # all the possible answers
    possible_answers = (xor_hex_and_char(input, number) for number in range(256))

    # filtered for answers that can be decoded into native strings
    filtered_answers = filter_strings(possible_answers)

    # filtered further for those that contain common English phrases
    final_answer = [answer for answer in filtered_answers if check_phrases(answer)]

    # print out the array containing the single final answer