A decryption program.

Is it possible to make a program that can decrypt any intercepted message?

There's probably a series of programs that can decrypt specific messages. There is no "skeleton key" program for it though. And I highly doubt someone can make one. Raindbow Cracking is popular for hash's, but those can take hours among hours to run. If you add in other "possible algorithms" for your program to check, it could take weeks.

Why? This sounds slightly malicious. ;)

I wouldn't worry about it being malicious, it is at the very least totally impossible. Is there a decryption program so good that it can crack any code? NO!

Suppose that all codes would scramble a message into a string of ASCII symbols (not too hard to suppose, since all bytes can be interpreted as ASCII, even if they are junk). If you had the universal cracker, I could send a random string of bytes in a message, and any meaning the universal cracker found in it would be incorrect simply because there was NO meaning in the original message.

Now suppose that the random bytes were not in a single message, but were interspersed at regular intervals in a crackable code. The universal cracker would not know which of the bytes in the message were meaning, and which were garbage unless it knew the pattern of the junk. If the pattern were one byte of meaning per packet, a message would be huge, but the number of crackable messages in the junk would probably be pretty big. This would make it impossible for a universal cracker to figure out which message was the true message unless it knew in advance which bytes held the message, which would not be information encoded in the message itself.

It is possible to unscramble any encrypted message but the time/processing power required is greater than the expected lifetime of the universe.

I don't quite agree. In theory, it is possible to unscramble any code, but the problem is that once you get to a certain size, it seems that the chances of unscrambling to a meaningful, but wrong, plaintext is increasingly likely.

After all, if you take the scrambled letters "sdfr s rergg as asdgsag" Simple replacement (a = b kind of stuff) might only produce one phrase, but no good encryption routine is that simple, so wouldn't there be several 'false interpretations'?

Cyrptography functions are written on purpose with high complexity. So any brute-force cracking attemp takes forever. Often cryptographic programmers want the complexitiy of their algo to run exponential (2^n).

Just my 2 cents.