AutoHotkey Community

[Titan]

As the title suggests, this is a basic yet highly effective encryption/decryption algorithm. It provides almost instantaneous encryption and keeps the cipher text the same size as the original text.
Credits: Laszlo (for various of things) and PhiLho (for performance tips)

Example:


Here's a speed test, you will need to include the original function in the script for it to work: BasicEncryptionSpeed.ahk

Usage guide:
Enc(type, str, pass [, l, h])
type: boolean - true for encryption or false for decryption
str: the string to be encrypted/decrypted
pass: the password for encryption/decryption
l & h are advanced params that you should omit (working ASCII bounds)

Download (< 1kb)

[kiu]

mmm, maybe useful for many purposes
_________________
____________________
______________________
kiu

[Laszlo]

Well, it is that: a (really) Basic Encryption script. There are a few basic problems with it:
- The encryption is dependent only on the first character of the password and its length
- The ciphertext can contain non-printable characters, what makes printing or saving it complicated
- With long passwords, the encrypted character codes can grow beyond the byte boundary, and disappear.
- The floor function does not have a function there.
- Only a constant is added to each character code, so trying all the possible offsets (less than 200) trivially breaks it.

Still, it can be useful to hide text from your kids. For a little more security and robustness, I'd suggest a few changes:
- restrict the input and output to printable characters (like in here).
- step through cyclically and use all characters in the password.

Even with these changes, the resulting cipher is weak: Subtracting two ciphertexts removes the effects of the password, providing the same result as subtracting the original texts. If an attacker knows a long enough plaintext-ciphertext pair, with this subtraction he can decrypt your messages.

[Titan]

I was going to AHKisize the TEA algorithm, I never knew you done it already ... fantastic.

In that case I'll abandon this script. But like you said, it can still prove useful in some cases.

[Laszlo]

[Titan]

Allright, I'll take your suggestions and write the next version, cheers.

[Titan]

Laszlo (or anyone else who can help),
I was in the process of writing the next version when I hit a dead end. I was able to encrypt the text but decrypting wasn't possible.

The encryption works by rotating the ASCII value of each character in the string against a list of alphanumeric characters. I'm not a cryptographic expert so I don't know if my encryption is seen as mathmatically legal or whatnot, but here it is (try it out first):

[Laszlo]

- You don't need % after -= or +=, but it is not an error.
- For speed, I would check if encryption or decryption is requested, outside of the text-processing loop, but it is still OK the way it is.
- Your input contains other characters, than listed in "c", like Space, "-", "," and ".". You never get them back with decryption. It looks safer to allow all ANSI (ASCII) characters between 32 and 126.
- I don't understand the function of the innermost loop, which runs once, twice or three times. Does it provide an extra level of complexity for the cipher? The length of the password is subtracted, which tends to be 5..8, making the effects of the cipher very easy to guess. I would subtract the characters of the password instead, which exhibit larger variations.
- If I understand right, the characters of the password are checked if their ANSI code is 0..49, 50..99 or 100..150, and this determines the number of times the innermost loop runs. For example, all the codes of the letters in google are larger than 100, so they all make the innermost loop run 3 times. Most of the small letters are like this, making the encryption often the same with different passwords of the same length.
- You restart processing the password at each letter of the input, so they are (cyclically) translated by the same amount. This is very easy to guess. As I said, it is easier and more secure to use the next password character directly. Something like this

[PhiLho]

Very nice!
I like it, I may use it if I need to put a password in a script that must run unattended.
The protection is very brittle, but at least, it protects against a casual glance at a script. To break the "protection", you just have to add a msgbox after the decryption... But it is the case for any script that must send the password, whatever the protection scheme.

I played with the code to make some small performance enhancements (probably unnoticeable anyway...) and to avoid duplicate code (but resulting in less elegant code...):

[Titan]

Thanks a lot Laszlo
I have modified your original script and improved it a little (usings PhiLho's suggestions), the script is now version 2.0.
_________________

[Laszlo]

[Laszlo]

The easiest to handle `r, `n, `t (tab) characters is to leave them alone. Although it leaks some information (the formatting), it also has some advantages: the ciphertext will not contain very long lines (for nicely formatted plaintexts), or have many line breaks not in the plaintext. The modification is trivial:

[Titan]

[PhiLho]

[Laszlo]