Word for Windows Password Cracker (R8)
Copyright by Fauzan Mirza, 1995
This program attempts to recover the password for Word for Windows encrypted
documents.
Usage: WFWCD <DocumentPath>
If the path to the protected document is not specified on the command line,
the user will be prompted to enter it (also facilitating running from within
Windows). Wildcards cannot be used in the document path.
There is a short delay during loading and analysis of the document, after
which a portion of the decrypted document and the password is displayed.
Technical Explanation
Word for Windows derives a 16 byte binary key from the password and cyclicly
XORs this key with the plaintext. This is called a repeated key cipher and
its weaknesses are well known.
In versions 1 and 2, the key was stored along with the encrypted document
(probably by accident). With a little intuition, it was trivial to retrieve
the decrypted text or to remove the encryption from the document:
Locate 16 bytes of pure key in the document header.
Cyclicly XOR the key with the ciphertext.
Marc Thibault used a similar routine in his public domain WORDUNP program.
In version 6, the document file format was changed and the key was no longer
stored in the file. This did not make it that much more difficult to solve
though; someone with a little knowledge of cryptanalysis would be able to
extract the plaintext from most encrypted documents:
Load ciphertext into a 16 column buffer.
For each key byte from 1 to 16,
Let key byte be the most frequent character in column XOR’d with 32.
Cyclicly XOR the key with the ciphertext.
Although this may be able to solve most files, most of the time the user
would need to correct guesses made by the routine. And if the document is
fairly short, or doesn’t contain enough spaces then it’s guaranteed to fail.
The new file format also made constructing a decrypted document quite
difficult.
WFWCD extends the above routine. It uses a much better algorithm to recover
the plaintext, and then to recover the password. The decryption engine is
based on the WP1 program, by Lyal Collins, but has been improved.
It assumes that the document consists predominantly of lower case letters,
and the letter frequencies are close to that of English. It may fail to solve
documents in which these conditions are not satisfied. For example, if there
is very little textual content in the file (less than 300 bytes) than it will
fail because of the lack of data for analysis.
Other Information
WFWCD (Word For Windows Cracker Demo) was developed and tested using gcc
under Linux, and was only recently ported to MSDOS. R8 is an indication of
the revision number. I had plans for R9, but it would take too much time to
complete and so I decided to release this preliminary version instead.
If you’d like to contact me, my current email address is:
Fauzan Mirza <fauzan@dcs.rhbnc.ac.uk>
Thanks to the following people to whom this program is dedicated:
Hamid Moosavi, Christopher Wilkinson, John Godley, and IRON MAIDEN.
WFWCD IS FREEWARE AND YOU ARE ALLOWED TO USE AND DISTRIBUTE THE PROGRAM
PROVIDED THAT IT IS NEITHER USED FOR AIDING DISHONEST OR CRIMINAL ACTIVITY,
NOR USED FOR PROFIT.