Base words (feature request)
Lexus89 opened this issue · comments
Hi @digininja First of all thank you for all the hard work you put in this tool. Pipal does a great job analyzing passwords. What I noticed is that the top 10 base words are not really the "base" words. For example, it sometimes happens that "p@ssw0rd" and "p@ssword' are (yes also separately) mentioned as a base word, while you would think in the basis it should be "password" (the others are variations).
Perhaps an idea to merge the functionality of deleet with pipal (https://github.com/digininja/deleet)? It could really improve the value of the output.
Very very late reply - perhaps usable for someone having this issue as well - a comparison to dictionary words may help determining what substitutions were used (for deleet). Convert back to the potential original words, search in a custom dictionary file, and the one found is likely to have the correct base word.
I've just checked and I've not done anything significant to this code for 10 years! It has probably been about that long since I last touched Ruby for anything more than a few line script.
I'll see if I can remember how any of it works and make some changes.
I've just checked and I've not done anything significant to this code for 10 years! It has probably been about that long since I last touched Ruby for anything more than a few line script.
I'll see if I can remember how any of it works and make some changes.
Haha no worries, I didn't even expect a reply as it was so long ago (I do hope it has not been 10 years already haha! Time flies). Just wanted to share the idea, even if for archiving purposes ;)
Try the new deleet_checker in here:
I've tided up that code, done a few more little fixes, and pushed it all into the main branch.
Hopefully it is all working correctly, any problems, let me know.
Thanks a lot! Looking at the output of deleet, it might bring more concrete results when deleet is focusing on the base word output instead of the raw passwords. With some luck this will return 'potato' as the main base word, which can be used for creating new (client tailored) word lists. When deleet-ing those they can also be merged again (p0t4t, p0t@t, etc. => single 'potat' statistic).
Password list
potato123!
123!potato
potato!
p0t4t0
pot@t0
p0t@to2023!
Top 10 base words
potato = 4 (57.14%)
p0t4t = 1 (14.29%)
pot@t = 1 (14.29%)
p0t@to = 1 (14.29%)
Top 10 passwords (from deleet)
potato = 2 (28.57%)
potatoi2e! = 1 (14.29%)
i2e!potato = 1 (14.29%)
potato! = 1 (14.29%)
potato2o2e! = 1 (14.29%)
There will be some weird cases like 'p0tat0' => 'potat' (due to ltrim/rtrim), which would make things a bit more complicated, but I am hoping the overall result will be more in line with what to feed to the password cracker. With larger results a print all instead of top X could also be useful.
My apologies for these ideas, I tend to overcomplicate things.. :)
I've just pushed a change that will deleet the base word rather than the main word.
This:
p@ssword
P@ssw0rd1
123Pas5word22
123P@ssw0rd1
fish
pa55word
Used to go to:
password = 2 (33.33%)
passwordi = 1 (16.67%)
i2epassword22 = 1 (16.67%)
i2epasswordi = 1 (16.67%)
fish = 1 (16.67%)
Now goes to:
password = 5 (83.33%)
fish = 1 (16.67%)
The base words are detected a lot better now, really useful thanks a lot! Small thing I noticed; sometimes it also prints an empty base word ("" = x%), probably because the string is empty after deleet-ing, but that doesn't really matter it's good enough for me ;)
I've just pushed a fix for this.