CH.A.S.M.

CHaracter Aware Split Method

The main idea behind Chasm is to reuse password patterns in a way that is likely to produce probable password candidates.

So how does this work?

Requirements

• libjudy
  On debian based systems use:
  $ sudo apt install libjudy-dev

Building

$ make
Or get the precompiled binaries. They are statically linked against libjudy.

Chasm usage

$ chasm -l 30 -k 2 out/chasm < dictionary.txt

Some user definable options that Chasm supports:

• -a (Analyze frequency, will output the frequency of the split part)
• -c Create rules for dictionary attack eg left-side becomes prefix rules ^ conversely right-side becomes $ suffix rules
• -s Sort by frequency, highest first
• -o [charcode] Only splits if the split character matches the character code, eg -o 101 will only split on the letter 'e' at the defined midpoint/range
• -l [string len] Will not split if the length of the input string if greater than this number -l 20 will skip all strings greater than length 20.
• -k [min occurance] Will not write splits where occurrence is less than defined occurrence -k 2 will not output splits which only occur once
• -m [mid point] Instead of splitting the string in half (len/2) you can specify a position to perform the split on
• -r [number] Will split X chars around the middle including the middle, X denotes how far you want to branch out