I've written two programs that scan wordlists. Feel free to modify the programs in any way; I provide no guarantee about your data or anything, though neither one does anything out of the ordinary.

• lb.c. Usage: lb n < filename. Finds all words or phrases in the wordlist "filename" which contain exactly n distinct letters. Useful for solving letter banks, or making anagram searches with grep easier to type.
• crypt.c. Usage: crypt ciphertext < filename. Finds all words in the wordlist "filename" which could be plaintext for the ciphertext; this includes the requirement that no letter can encode itself. An "@" before the first letter indicates it should be considered a capital. Capital letters within the ciphertext are treated as plaintext; the program matches them to plaintext instead of trying to substitute.

The code was written to compile under gcc. If there are any problems, or you have programs you'd like to share, just e-mail me at kutin@math.uchicago.edu.