TorBot vulnerable to Inefficient Regular Expression Complexity in validate_link

Summary

The torbot.modules.validators.validate_link function uses the python-validators URL validation regex. This particular regular expression has an exponential complexity which allows an attacker to cause an application crash using a well-crafted argument…

Details

https://github.com/DedSecInside/TorBot/blob/d2b89192379ad033ffc7044efff26e16ccc02d5a/torbot/modules/validators.py#L13

An attacker can use a well-crafted URL argument to exploit the vulnerability in the regular expression and cause a Denial of Service on the system.

PoC

I have uploaded a secret gist containing a PoC (https://gist.github.com/ikkebr/6041055314f1cfb8e65b2a1acbaae12c). By adding one special character at the end of the user argument of the URL, the regular expression will take exponentially longer to compute.

For a string of size 10k, the regex will take 0.01s without the well-crafted URL and 1.3s with the well-crafted URL exploit.
For a string of size 50k, the regex will take 0.03s without the well-crafted URL and 35s with the well-crafted URL exploit.
For a string of size 100k, the regex will take 0.05s without the well-crafted URL and over 200s with the well-crafted URL exploit.

The regular expression used in the validators library versions [0.20, 0.11] is vulnerable to this attack. Version 0.21 appears to be unaffected, but it no longer contains a single regular expression.

Impact

An attacker could exploit this vulnerability to cause a denial of service or increased resource usage.