Trojan project

Greetings /prog/, I'm from a neighboring board.

I'm on the last phases of a trojan project. I'm having a hard time finding the best way of hiding my malicious executable within another executable. I assume the best way of doing that would involve assembly? "Code caves" seem to be what I'm looking for, though there doesn't seem to be much tutorials for it. Any advice? Thanks in advance.

Just in case anybody is curious. The features for this trojan include (but are not limited to) the following:

-Reverse connection by way of an IRC channel.
-Offline/online keylogging.
-Desktop screencapure.
-Twofish/AES encryption for all command & control operations sent/read to the trojan--because safety for the users of this trojan comes first. (Fun fact: The CBC used with this implementation of AES uses a (pseudo) random-number generator for each instance of the executable.)
-A login/register system for guest who use this trojan.
-Limited RAT functionality, because I don't really have much time to mess around with the Windows API.

In other words, a bare bones, backdoor trojan with a little encryption added for paranoia. Also, when executed this trojan leaves virtually no trace to the user (or author) of the trojan. The random server and channels it connects to are determined at run-time with the help of a little entropy. And all inbound and outbound messages are encrypted. I'd feel secure running this type of trojan in the office of a senior FBI agent's office.

>>21
The reason I'm saying this is because while it's perfectly possible to write it all in C, your stub will look pretty bloated if you were to just relocate your executable at the end of the host and keep all the libc bloat that comes with the territory
With my current level of skill, I'd be lucky if I could accomplish that.