bogocrypt

The stupid encryption+obfuscation algorithm.

What the fuck is this shit?

This is a thing that came up to my mind one day. It was very simple and easy to implement, and I decided to write it, just for fun.

bogocrypt is a very silly algrithm, and it works in a very silly way. No complex maths involved. But lots of garbage.

And how does this nonsense work?

bogocrypt needs two things: an input blob and a key.

Assume that:

• blob is the input blob. It is an array of bytes.
• key is the key. It is also an array of bytes.
• n is the length of the key.
• cypher is the output blob.

Now, for every b bytes from blob, n bytes will be appended to cypher. n-1 of those bytes can be anything, really. I call them garbage bytes. Only one of those bytes actually has usefull content. I call it needle byte, which is XORed to the respective byte in key when written to cypher (wrapping around if n is greater than the length of blob).

Now here comes the tricky part. In order to make this whole mess useful (to some extent), the position of the needle byte needs to be different for every n garbage bytes. This creates a challenge for anyone that tries to retrieve information from cypher, as they will have a whole lot more garbage than usefull data. And since they don't (well, shouldn't) know the key, they can't technically guess which bytes are garbage and which ones are not.

"Please don't write your own encryption algorithms!!!111ONE!!"

I know. I don't expect this to be anything but a joke, really. That's why I called it bogocrypt, which was inspired by bogosort. I have not put this into any kind of theoretical analysis, and I don't intend to. If you take this thing seriously, you are probably not spending your time wisely (as.. well... me, in fact).

Usage

So you are really interested in seeing how this thing works, eeh? Well, good for you I have prepared a small demo.

Compile simply by running make. It has no make install rule. Stop being silly. Also, this was written in pure ANSI C and doesn't have any dependencies, so it should work anywhere.

Then use it:

$ bogocrypt -e "test" -k 1234
> 4546542307665F516B7D407B0C28475B
$ bogocrypt -d 4546542307665F516B7D407B0C28475B -k 1234
> test

It also has an interactive mode :D (now you're thinking "damn, this guy has put a lot of effort in this crappy shit" — yes I did).

Encrypt:

$ bogocrypt -i
> bogocrypt - The stupid encryption+obfuscation algorithm.
> Interactive mode: [e:encrypt|d:decrypt]: e
> Insert input (new line marks end - prefered format is base64): YmFuYW5hcyE=
> Insert key (new line marks end - prefered format is byte string): K()"t83d    
> 2110127B714C42794C7C0454191F4C246B183E646107614347535C574B5E1610680E074A2D7C304D1C57531B3B3D7D4F5B722B170C0D7B0450405C2057745F062810612B5F1B73663233032E6E5F682D7D6D45093F6E79484E14561349135F1D

Then decrypt:

$ bogocrypt -i
> bogocrypt - The stupid encryption+obfuscation algorithm.
> Interactive mode: [e:encrypt|d:decrypt]: d
> Insert input (new line marks end - format is HEX): 2110127B714C42794C7C0454191F4C246B183E646107614347535C574B5E1610680E074A2D7C304D1C57531B3B3D7D4F5B722B170C0D7B0450405C2057745F062810612B5F1B73663233032E6E5F682D7D6D45093F6E79484E14561349135F1D
> Insert key (new line marks end - prefered format is byte string): K()"t83d
> YmFuYW5hcyE=

Future work

Now you just said "reeeally??"

Yes. There are still things that bother me about this algorithm. One of them is that once the key wraps around the input blob, the positions calculated for each needle will also repeat. This could theoretically be used for pattern matching. All my attempts to solve this did not work.