Bolt is in beta phase of development which means there can be bugs. Any production use of this tool discouraged. Pull requests and issues are welcome. I also suggest you to put this repo on watch if you are interested in it.
Bolt crawls the target website to the specified depth and stores all the HTML forms found in a database for further processing.
In this phase, Bolt finds out the tokens which aren't strong enough and the forms which aren't protected.
This phase focuses on detection on replay attack scenarios and hence checks if a token has been issued more than one time.
It also calculates the average levenshtein distance between all the tokens to see if they are similar.
Tokens are also compared against a database of 250+ hash patterns.
In this phase, 100 simultaneous requests are made to a single webpage to see if same tokens are generated for the requests.
This phase is dedicated to active testing of the CSRF protection mechanism. It includes but not limited to checking if protection exsists for moblie browsers, submitting requests with self-generated token and testing if token is being checked to a certain length.
Various statistical checks are performed in this phase to see if the token is really random. Following tests are performed during this phase
- Monobit frequency test
- Block frequency test
- Runs test
- Spectral test
- Non-overlapping template matching test
- Overlapping template matching test
- Serial test
- Cumultative sums test
- Aproximate entropy test
- Random excursions variant test
- Linear complexity test
- Longest runs test
- Maurers universal statistic test
- Random excursions test
Scanning a website for CSRF using Bolt is as easy as doing
python3 bolt.py -u https://github.com -l 2
Where -u
is used to supply the URL and -l
is used to specify the depth of crawling.
Other options and switches:
-t
number of threads--delay
delay between requests--timeout
http request timeout--headers
supply http headers
Regular Expressions for detecting hashes are taken from hashID.
Bit level entropy tests are taken from highfestiva's python implementation of statistical tests.