Server-Side Request Forgery (SSRF) attacks allow an attacker to make requests to any domains through a vulnerable server.
Usually, attackers make the server connect back to itself, to some internal service or resource, or to its own cloud provider.
SSRF attacks abuse any type of URL or file upload functionality in application to send malformed URLs.
Vulnerable places are parameters that accept URLs or files.
An SSRF exploit that causes connections to external third-party systems might result in malicious attacks that appear to originate from the organization hosting the vulnerable application, causing potential legal liabilities and reputational damage.
Successful SSRF attacks usually result in accessing unauthorized data. That data can be server files, cloud provider metadata, open ports.
Besides accessing data, SSRF attacks can also scan internal networks. This lets a hacker see which services are running and let them connect to a service and start chaining exploits.
SSRF can also be chained into other attacks that can be much more damaging, These range from reflected XSS all the way to removing code executions.
There are two main types of SSRF attacks. The first is where the attack is done against the server itself by using a loopback network interface (
The second is where the trust relationship between the abused server being attacked and other devices on the same network that can’t be accessed through other means. Or even using that server to gather information about the cloud provider that is being used (
AWS, GCE, Azure...).
SSRF is injected into any parameter that accepts a URL or a file. When injecting SSRF payloads in a parameter that accepts a file, the attacker has to change
text/plain and then inject the payload instead of a file.
Accessing internal resources can mean a couple of different things. It can be achieved by accessing the
/admin panel that is only accessible from within the internal network. Reading files from the server. This can be done using the file schema (
Some common exploits for accessing internal pages include:
Internal files via URL scheme
Attacking the URL scheme allows an attacker to fetch files from a server and attack internal services.
Some common exploits for accessing internal files include:
Internal services via URL scheme
You can use a URL scheme to connect to certain services.
For file transfer protocols:
Makes request like:
Makes request like:
From: [Attacker] <email@example.com>
Date: Fri, 13 Mar 2020 03:33:00 -0600
You've been exploited :(
Cross-Site Port Attack (XSPA) is a type of SSRF where an attacker is able to scan the server for its open ports. This is usually done by using the loopback interface on the server (
localhost) with the addition of the port that is being scanned (
21, 22, 25...).
Some examples are:
Besides scanning for ports an attacker might also run a scan of running hosts by trying to ping private IP addresses:
Cloud provider metadata
With SSRF an attacker is able to read metadata of the cloud provider that you use, be it AWS, Google Cloud, Azure, DigitalOcean, etc. This is usually done by using the private addressing that the provider listed in their documentation.
For AWS instead of using
127.0.0.1 attackers use the
169.254.169.254 address for exploits.
Significant information can be extracted from AWS metadata, from public keys, security credentials, hostnames, IDs, etc.
Some common exploits include:
Additional links can be found in the official documentation of AWS.
Similar to AWS, DigitalOcean uses
169.254.169.254 for their services and checks the documentation for more information.
Azure is more limited than other cloud providers in this regard. Check the official documentation for more information.
Azure requires header
Oracle cloud uses the
One way to protect against SSRF is to blacklist certain domains and IP addresses. Yet, blacklisting isn’t really a good defense technique as hackers can use bypasses to avoid your security measures.
Bypass using HTTPS
Common blacklists blocking everything on port
80 or the
http scheme. but the server will handle requests to
https just fine.
Instead of using
The most common blacklist is blacklisting IP addresses like
127.0.0.1 or localhost. To bypass these blacklists you can use:
[::], abuses IPv6 to exploit
- With domain redirection, useful when all IP addresses are blacklisted
- With CIDR, useful when just
- With IPv6/IPv4 address embedding, useful when both IPv4 and IPv6 are blacklisted (but blacklisted badly)
- With decimal IP location, really useful if dots are blacklisted
http://0220.127.116.11/ --> (127.0.0.1)
http://2130706433/ --> (127.0.0.1)
http://3232235521/ --> (192.168.0.1)
http://3232235777/ --> (192.168.1.1)
- With malformed URLs, useful when port is blacklisted
- With shorthanding IP addresses by dropping zeros, useful when full IP address is whitelisted
- With enclosed alphanumerics, useful when just plain ASCII characters are blacklisted but servers interpret enclosed alphanumerics as normal.
- With bash variables (cURL only)
curl -v "http://attacker$google.com"; $google = ""
- Against weak parsers (these go to
Bypass 169.254.169.254 address
The most common bypass for AWS addresses is changing them to get past the blacklist of
http://425.510.425.510– dotted decimal with overflow
http://2852039166– dotless decimal
http://7147006462– dotless decimal with overflow
http://0xA9.0xFE.0xA9.0xFE– dotted hexadecimal
http://0xA9FEA9FE– dotless hexadecimal
http://0x41414141A9FEA9FE– dotless hexadecimal with overflow
http://0251.0376.0251.0376– dotted octal
http://0251.00376.000251.0000376– dotted octal with padding
The easiest way to remediate SSRF is to whitelist any domain or address that your application accesses.
Blacklisting and regex have the same issue, someone will eventually find a way to exploit them
Never send a raw response body from the server to the client. Responses that the client receives needs to be expected.
Allow only URL schemas that your application uses. There is no need to have
ftp://, file:/// or even
http:// enabled if you only use
And if you do use other schemas make sure that they’re only accessible from the part that needs to access them and not from anywhere else.
Make sure that authentication is enabled on any service that is running inside your network even if they don’t require it. Services like
memcached, redis, mongo and others don’t require authentication but can be exploited.
Input sanitization and validation
Never trust user input.
Always sanitize any input that the user sends to your application. Remove bad characters, standardize input (double quotes instead of single quotes for example).
After sanitization make sure to validate sanitized input to make sure nothing bad passed through.
SSRF is a very dangerous vulnerability that can leak a lot of critical information about internals of a network, and then be chained to other much more dangerous attacks, but with the correct steps it can be remedied.