context aware wordlist generation
Icecold is a tool designed to generate wordlists based on websites, it does this by scraping words and combining and modifying them into forms that are commonly used as passwords.
This is a very CTF-esque tool, as a common CTF category is creating passwords based off a specific theme or topic. It still has applications in real world password recovery, it could have been used that time you found a Raspberry Pi in your closet and you were pretty sure the password was Love Live themed.
Icecold does not take duplicates into account in order to keep ram usage and disk read/write low, if this is of great concern (enough to take the cost of weaker wordlist generation), then CeWL is another option.
The command icecold.py https://google.com -d 0 -C "?" --no-smush will give output that looks similar to the following
...
savings
are
now
live
the
google
store
cyber?monday
monday?savings
savings?are
are?now
now?live
live?the
the?google
cyber?monday?savings
monday?savings?are
savings?are?now
are?now?live
now?live?the
live?the?google
...
Icecold has many command line flags to customize the output wordlist. It should be noted that icecold works best when use in tandem with other wordlist generation tools, such as hashcat and John the Ripper.
git clone https://github.com/Shensd/icecold
pip3 install -r requirements.txt
This program is inteded to be used with python3, and compatability with python2 was not taken into account during development.
usage: icecold.py [url] [options...]
-u, --url Base url to scrape
-U, --url-file Use a file with a list of urls, with one on each line
-o, --output Output file location to dump wordlist (default stdout)
-m, --min-word-len Minimum word length to accept (default 3)
-M, --max-word-len Maximum word length to accept (default 15)
-c, --chain-len Maximum length to make word chains (default 3)
-h, --help Display this help page
--ignore-unresponsive Do not halt program when a website doesn't respond, skip website instead
-d, --depth How many links deep to spider pages (default 1)
--no-leave-domain Disallow leaving domain while spidering
--ua Set the user agent used in requests
-C, --charset Charset to use when making word chains, (default '_-')
--words-only Do not make word chains, just print single words (macro for -c 1)
--no-smush Do not create word chains that only push words together, only use connecting characters