I am planning to create an open source education web app where people can add and edit the content (a bit like Wikipedia).
However I wish to add another feature that allows the user to add their own interactive content using JavaScript. (similar how JSFiddle does it)
What are the security concerns in doing this?
Optional question: How can these issues be overcome?
Yes you could use HTML5 Sandbox to only load user scripts in an IFrame.
You should only host user content from a different domain than your main site. This will prevent any XSS attack if an attacker convinces a user to visit the page directly (outside of the sandbox). e.g. if your site is www.example.com you could use the following code to display the sandboxed IFrame (note .org rather than .com, which is an entirely different domain):
<iframe src="https://www.example.org/show_user_script.aspx?id=123" sandbox="allow-scripts"></iframe>
This will allow scripts, but forms and navigation outside of the IFrame will be prevented. Note that this approach could still risk a user hosting a phishing form to capture credentials. You should make sure that the boundaries between your site and the user content are clear within the user interface. Even though we haven't specified allow-forms, this only prevents a form from being submitted directly, it does not prevent form elements and JavaScript event handlers from sending any data to an external domain.
The HTML5 Security Cheat Sheet guidance on OWASP states this is the purpose of the sandbox:
Use the sandbox attribute of an iframe for untrusted content
You should test whether sandbox is supported first, before rendering the IFrame:
<iframe src="/blank.htm" sandbox="allow-scripts" id="foo"></iframe>
var sandboxSupported = "sandbox" in document.createElement("iframe");
if (sandboxSupported) {
document.getElementById('foo').setAttribute('src', 'https://www.example.org/show_user_script.aspx?id=123');
}
else
{
// Not safe to display IFrame
}
It is safer to do it this way by dynamically changing the src rather than redirecting away if sandboxSupported is false because then the iframe will not accidentally be rendered if the redirect doesn't happen in time.
As a simpler alternative, without the need to check whether the sandbox is supported, you can use the srcdoc IFrame attribute to generate the sandboxed content, making sure that all content is HTML encoded:
e.g.
<html><head></head><body>This could be unsafe</body></html>
would be rendered as
<iframe srcdoc="<html><head></head><body>This could be unsafe</body></html>" sandbox="allow-scripts"></iframe>
Or you could construct a data blob object, being careful to HTML encode again:
<body data-userdoc="<html><head></head><body>This could be unsafe</body></html>">
<script>
var unsafeDoc = new Blob([document.body.dataset.userdoc], {type: 'text/html'});
var iframe = document.createElement('iframe');
iframe.src = window.URL.createObjectURL(unsafeDoc);
iframe.sandbox = 'allow-scripts';
</script>
Of course you could also set the unsafeDoc variable from a JSON data source. It is not recommended to load an HTML file, as this has the same problem of it having to be from an external domain, as the attacker could just entice the user to load that directly.
Also, please don't be tempted to write user content into a script block directly. As shown above, data attributes is the safe way to do this, as long as correct HTML encoding is carried out on the user data as it is output server-side.
In these cases you can leave src as blank.html as older browsers that do not support srcdoc will simply load that URL.
As #Snowburnt touches upon, there is nothing stopping a user script from redirecting a user to a site where a drive-by download occurs, but this approach, assuming a user is up to date on patches, and there are no zero day vulnerabilities, this is a safe approach because it protects its end users and their data on your site via the same origin policy.
One big issue is cross-site scripting where users add code that tells the browser to open and run code from other sites. Say they add something that creates an iFrame or a hidden iFrame pointing to a site and starts downloading malicious code.
There's no simple way around it (thanks to Bergi in the comments) to make sure no elements are created and no ajax calls are made.
I've been a member of sites that provided this functionality, but for those sites I paid for my own space so any vulnerabilities I add are inconveniencing my own clients, in that case it's a little more okay to let that slip by since it's not a security leak for everyone.
One way around this is to create customizable controls for the users to use to add interactivity. The plus is that you control the javascript being added, the minus is that your user base will have to request and then wait for you to create them.
Related
Would it be possible to load an external page inside a container and replace text elements?
We work with ad campaigns and earn a percentage whenever a user signs up.
Can a script replace certain words? For instance “User” to “Usuario” or “Password” to “Contraseña” without affecting the original website or its functions.
Note: These links always pass through a redirection.
Example:
http://a2g-secure.com/?E=/0yTeQmWHoKOlN6zUciCXQwUzfnVGPGN&s1=
Note 2: Using an iframe is out of the question due to “Same-origin policy”.
I'm not sure if this answers your question, but you might find it useful.
(Perhaps you might give a step-by-step example of what you're trying to accomplish?)
If we assume that a browser attempts to retrieve page P from a proxy which first retrieves the content of page P from its actual home and then performs some transformation on its content before returning that page content to the browser, what you're describing is a Reverse HTTP Proxy and is a very well-known page serving technique.
Rather than performing complex transformations at the server (which require specialized knowledge of the page layout), this technique is usually used to inject a single line into the retrieved source that calls a JavaScript file to actually perform the required transformation at the browser.
So in essence:
Browser requests Page P from Proxy 1.
Proxy 1 retrieves the actual Page P from its real home, Server 2.
Proxy 1 adds the line <script src="//proxy1.com/transform.js"></script> to the source of Page P.
Proxy 1 then returns the modified source of Page P to Browser.
Once the Browser has received the page content, the JavaScript file is also retrieved, which can then modify the page contents in any way required.
This technique can be used to solve your "Same origin policy" issue by loading an iframe from a URL that points to the same server as that which provided the parent or owning page of the iframe which acts as proxy, like:
http://example.com/?proxy_target=//server2.com/pageP.html
Thus, the browser only "sees" content from a single server.
You would need to load the external page server-side, and then you can do whatever you want with it. You can do serverside string replacement, or you can do it later in javascript.
But, remember that as soon as you add a whole webpage into for example a div in your own page, the css from your page will affect it.
Plus, you would need to manipulate all the links in the documents, to have absolute urls. If the page depends on ajax, there is pretty much no way to accomplish what you want to do.
If on the other hand the pages you will be loading are static html, it is possible, though there are a lot of things you need to take care of before you can actually present the page to the user, like adjusting links, urls to stylesheets and so on.
It seems you are trying to localize a website on the fly, using your server as a proxy for that content. Does it make sense? If that's the case, depending on the size of your operation, there are several proxy translation services out there (I'll name them if needed).
Basically, they scrape a website, providing a way for you to translate and host the translated content. Of course, this depends on your relationship with the content providers. You should also take this into consideration, since modifying content, even for translation, can be a copyright problem.
All things considered, if you trust the provider's javascript, the solution involves scraping the content, as mentioned in other answers, and serving that modified content. You really need to trust the origin...
update per request
http://www.easyling.com
http://www.smartling.com
http://www.motionpoint.com
http://www.lionbridge.com/solutions/translation-proxy/
http://www.sajan.com/translation-proxy-technology-and-traditional-website-translation-understanding-your-options/
They are all aimed at enterprise-grade projects, but I would say Easyling is the most accessible.
Hope this helps.
Using the .load() callback function, this will replace the text
$(function(){
$("#Content").load("http://example.com?user=Usuario",function() {
$(this).html($(this).html().replace("user", +get param value+));
});
redirection u can use
// similar behavior as an HTTP redirect
window.location.replace("url");
// similar behavior as clicking on a link
window.location.href = "url";
The answer is NO, not without using a server-side proxy. For a really good overview of how to use a proxy, see this YUI page: https://developer.yahoo.com/javascript/howto-proxy.html (Be patient, as it will take time to load, but the illustrations are worth it!)
When I try to do this in jsfiddle to see what data that the 3 parameters contain, then the error below appears:
$(function() {
$(this).load('https://stackoverflow.com/questions/36003367/load-external-page-and-replace-text', function(responseText, textStatus, jqXHR){
debugger;
});
});
ERROR:
XMLHttpRequest cannot load Load external page and Replace text.
No 'Access-Control-Allow-Origin' header is present on the requested resource. Origin 'https://fiddle.jshell.net' is therefore not allowed access.
You might know HN, but you maybe also do not like the fact of clicking around in many tabs/going back and forward. I thought about making some page on which both webpage and links from HN are. So I made this: http://goodfrontpage.com/direct
But there are 2 Problems:
First: How to determine if a page doesn't allow in it's http-headers to open it in something like this:
<iframe class="webpage" src="{{post.url}}" ></iframe>
or this:
<object class="webpage" data="http://asfasfasfa.com" >
<embed class="webpage" src="http://asfasffvasf.com" > </embed>
Error: Webpage not accessible!
</object>
This is true for pages like github.com, eff.org or youtube.com
Second: Is there any possibility to fetch the sites in a different way allowing me to display all pages?
If you want to embed a webpage within another one then you should use the iframe element:
<iframe src="http://asfasffvasf.com"/>
You can style this like any other block element, and set an explicit width and height.
Some pages ask browsers not to include them within iframes (using the X-Frame-Options header). I don't think there's an easy way to solve this on the client side, but you could create a simple backend or proxy to request the page you want and return the content. This gets round the iframe restriction because you're now including content from your own domain.
This does have a couple of security issues to be aware of:
You've now made a backend which can be used to download any page on the Internet. There's a denial of service vulnerability if someone makes lots of requests to download huge pages.
The pages you're including will no longer be restricted by the same origin policy. Scripts on those pages will be able to interact with everything on the parent page. This may be a problem if you plan on creating login functionality in the future.
Seems like you are looking for iframe property.It specifies an inline frame.
An inline frame is used to embed another document within the current HTML document.
<iframe src="http://asfasfasfa.com"></iframe>
I'm building a Chrome extension and using the db.js wrapper to utilize the indexeddb. The problem is, I've got several subdomains and I'd like to be able to share the information across them.
When I use the Chrome Dev tools to view Resources, all of the individual subdomains have their own copy of the schema I'm creating, and each has it's own data.
The only thing I knew to try was to set the document.domain but that didn't help. I wasn't surprised.
Documentation on indexeddb is very slim it seems. I keep finding the same 2 or 3 blog posts copied word for word in several different blogs and nothing specifies that this is possible or impossible.
You can't access the same database from multiple subdomains, the access scope is limited to html origin.
html_Origin = protocol + "://" + hostname + ":" + port + "/";
As #Xan mentioned, if you can use a common origin owned by the extension itself, rather than by the content pages, that sounds like it would be by far the easiest solution. If for whatever reason you can't do that (or for readers who got here wanting to know about regular page javascript or Greasemonkey-style userscripts, rather than extensions), the answer is:
Yes, though it's a slightly awkward and takes some work:
Since you're using a number of related subdomains, (rather than completely unrelated domains), there's a technique you can use in that situation. It can be applied to IndexedDB, localStorage, SharedWorker, BroadcastChannel, etc, all of which offer shared functionality between same-origin pages, but for some reason don't respect modifications to document.domain.
(1) Pick one "main" subdomain to for the data to belong to. i.e. if your subdomains are https://a.example.com, https://b.example.com, and https://c.example.com, you might choose to have your IndexedDB database stored under the https://a.example.com subdomain.
(2) Use it normally from all the the https://a.example.com pages.
(3) On https://b.example.com and https://c.example.com, use javascript to set document.domain = "example.com";. Then also create a hidden <iframe>, and navigate it to some page on the https://a.example.com domain (It doesn't matter what page, as long as you can insert a very little snippet of javascript on there. If you're creating the site, just make an empty page specifically for this purpose. If you're writing an extension or a userscript and so don't have any control over pages on the example.com server, just pick the most lightweight page you can find and insert your script into it. Some kind of "not found" page would probably be fine).
(4) The script on the hidden iframe page need only (a) set document.domain = "example.com";, and (b) notify the parent window when this is done. After that, the parent window can access the iframe window and all its objects without restriction! So the minimal iframe page is something like:
<!doctype html>
<html>
<head>
<script>
document.domain = "example.com";
window.parent.iframeReady(); // function defined & called on parent window
</script>
</head>
<body></body>
</html>
If writing a userscript, you might not want to add externally-accessible functions such as iframeReady() to your unsafeWindow, so instead a better way to notify the main window userscript might be to use a custom event:
window.parent.dispatchEvent(new CustomEvent("iframeReady"));
Which you'd detect by adding a listener for the custom "iframeReady" event to your main page's window.
(5) Once the hidden iframe has informed its parent window that it's ready, script in the parent window can just use iframe.contentWindow.indexedDB, iframe.contentWindow.localStorage, iframe.contentWindow.BroadcastChannel, iframe.contentWindow.SharedWorker instead of window.indexedDB, window.localStorage etc. ...and all these objects will be scoped to the https://a.example.com origin - so they'll have the this same shared origin for all of your pages!
The "awkward" part of this technique is mostly that you have to wait for the iframe to load before proceeding. So you can't just blithely initialize IndexedDB in your DOMContentLoaded handler, for example. Also you might want to add some error handling to detect if the hidden iframe fails to load correctly.
Obviously, you should also make sure the hidden iframe is not removed or navigated during the lifetime of your page... OTOH I don't know what the result of that would be, but very likely bad things would happen.
And, a caveat: setting/changing document.domain can be blocked using the Feature-Policy header, in which case this technique will not be usable as described.
However, there is a significantly more-complicated generalization of this technique, that can't be blocked by Feature-Policy, and that also allows entirely unrelated domains to share data, communications, and shared workers (i.e. not just subdomains off a common superdomain). #Xan alludes to it in point (2) of his answer:
The general idea is that, just as above, you create a hidden iframe to provide the correct origin for access; but instead of then just grabbing the iframe window's properties directly, you use script inside the iframe to do all of the work, and you communicate between the iframe and your main window only using postMessage() and addEventListener("message",...).
This works because postMessage() can be used even between different-origin windows. But it's also significantly more complicated because you have to pass everything through some kind of messaging infrastructure that you create between the iframe and the main window, rather than (for example) just using the IndexedDB API directly in your main window's code.
HTML-based storage (indexedDB, localStorage) in Chrome extensions behaves in a way that might not be expected, but it's perfectly natural.
In the background page, the domain is chrome-extension://yourextensionid/, and this is shared by all extension pages and is persistent.
In the content scripts though, you're sharing the HTML storage with the domain you're operating on. This makes life difficult if you want it to share/persist things. Note that sometimes this behavior is actually helpful.
The universal solution is to keep the DB in a background script, and communicate data/requests by means of Messaging API.
This was the usual solution for localStorage use until chrome.storage came along. But since you're using a database, you don't have a ready extension-friendly replacement.
I am attempting to share data across subdomains using Safari. I would like to use an HTML5 database (specifically localStorage as my data is nothing but key-value pairs).
However, it seems as though data stored to example.com can not be accessed from sub.example.com (or vice versa). Is there any way to share a single database in this situation?
Update 2016
This library from Zendesk worked for me.
Sample:
Hub
// Config s.t. subdomains can get, but only the root domain can set and del
CrossStorageHub.init([
{origin: /\.example.com$/, allow: ['get']},
{origin: /:\/\/(www\.)?example.com$/, allow: ['get', 'set', 'del']}
]);
Note the $ for matching the end of the string. The regular expression in the above example will match origins such as valid.example.com, but not invalid.example.com.malicious.com.
Client
var storage = new CrossStorageClient('https://store.example.com/hub.html');
storage.onConnect().then(function() {
return storage.set('newKey', 'foobar');
}).then(function() {
return storage.get('existingKey', 'newKey');
}).then(function(res) {
console.log(res.length); // 2
}).catch(function(err) {
// Handle error
});
Check https://stackoverflow.com/a/39788742/5064633
There is simple way to use cross-domain anything, just create simple page that will be included as proxy iframe hosted on domain you try to access, send PostMessage to that iframe and inside iframe you do your LocalStorage database manipulation. Here is a link to article that do this with lcoalStorage. And here is demo that send message to different page in subdomain check the source code, it use iframe and PostMessage.
EDIT: New version of sysend.js library (used by above demo) use BroadcastChannel if browser support it, but still it require Iframe. Recent version also simplify using of Cross-Origin messages, you have html of the iframe in repo, that you can use (or you can use simple html file with single script tag with the lib) and in parent you just need to call one function sysend.proxy('https://example.com'); where example.com need to have proxy.html file (you can also use your own filename and different path).
Google Chrome blocks localStoage access from an iFrame in another domain by default,unless 3rd party cookie is enabled and so does Safari on iPhone...the only solution seems to be opening the parent domain on a different domain and then sending to to the Child via window.postMessage but looks ugly and shifty on phones...
Yes. This is how:
For sharing between subdomains of a given superdomain (e.g. foo.example.com vs bar.example.com vs example.com), there's a technique you can use in that situation. It can be applied to localStorage, IndexedDB, SharedWorker, BroadcastChannel, etc, all of which offer shared functionality between same-origin pages, but for some reason don't respect any modification to document.domain that would let them use the superdomain as their origin directly.
NOTE: This technique depends on setting document.domain to allow direct communication between iframes on different subdomains. That functionality has now been deprecated. (As of April 2021 it continues to work in all major browsers however. From Chrome v109 the feature will be disabled unless an Origin-Agent-Cluster: ?0 header is also sent.)
NOTE: Be aware that this technique removes the same-origin defences that block malicious script on a subdomain from affecting the main-domain window, or visa versa, potentially broadening the attack surface for XSS attacks. There are other security implications for shared hosting as well - see the MDN document.domain page for details.
(1) Pick one "main" domain to for the data to belong to: i.e. either https://foo.example.com or https://bar.example.com or https://example.com will hold your localStorage data. Let's say you pick https://example.com.
(2) Use localStorage normally for that chosen domain's pages.
(3) On all other https://*.example.com pages (the other domains), use JavaScript to set document.domain = "example.com"; (always the superdomain). Then also create a hidden <iframe>, and navigate it to some page on the chosen https://example.com domain (It doesn't matter what page, as long as you can insert a very little snippet of JavaScript on there. If you're creating the site, just make an empty page specifically for this purpose. If you're writing an extension or a Greasemonkey-style userscript and so don't have any control over pages on the example.com server, just pick the most lightweight page you can find and insert your script into it. Some kind of "not found" page would probably be fine).
(4) The script on the hidden iframe page need only (a) set document.domain = "example.com";, and (b) notify the parent window when this is done. After that, the parent window can access the iframe window and all its objects without restriction! So the minimal iframe page is something like:
<!doctype html>
<html>
<head>
<script>
document.domain = "example.com";
window.parent.iframeReady(); // function defined & called on parent window
</script>
</head>
<body></body>
</html>
If writing a userscript, you might not want to add externally-accessible functions such as iframeReady() to your unsafeWindow, so instead a better way to notify the main window userscript might be to use a custom event:
window.parent.dispatchEvent(new CustomEvent("iframeReady"));
Which you'd detect by adding a listener for the custom "iframeReady" event to your main page's window.
(NOTE: You need to set document.domain = example.com even if the iframe's domain is already example.com: Assigning a value to document.domain implicitly sets the origin's port to null, and both ports must match for the iframe and its parent to be considered same-origin. See the note here: https://developer.mozilla.org/en-US/docs/Web/Security/Same-origin_policy#Changing_origin)
(5) Once the hidden iframe has informed its parent window that it's ready, script in the parent window can just use iframe.contentWindow.localStorage, iframe.contentWindow.indexedDB, iframe.contentWindow.BroadcastChannel, iframe.contentWindow.SharedWorker instead of window.localStorage, window.indexedDB, etc. ...and all these objects will be scoped to the chosen https://example.com origin - so they'll have the this same shared origin for all of your pages!
The most awkward part of this technique is that you have to wait for the iframe to load before proceeding. So you can't just blithely start using localStorage in your DOMContentLoaded handler, for example. Also you might want to add some error handling to detect if the hidden iframe fails to load correctly.
Obviously, you should also make sure the hidden iframe is not removed or navigated during the lifetime of your page... OTOH I don't know what the result of that would be, but very likely bad things would happen.
And, a caveat: setting/changing document.domain can be blocked using the Feature-Policy header, in which case this technique will not be usable as described.
However, there is a significantly more-complicated generalization of this technique, that can't be blocked by Feature-Policy, and that also allows entirely unrelated domains to share data, communications, and shared workers (i.e. not just subdomains off a common superdomain). #jcubic already described it in their answer, namely:
The general idea is that, just as above, you create a hidden iframe to provide the correct origin for access; but instead of then just grabbing the iframe window's properties directly, you use script inside the iframe to do all of the work, and you communicate between the iframe and your main window only using postMessage() and addEventListener("message",...).
This works because postMessage() can be used even between different-origin Windows. But it's also significantly more complicated because you have to pass everything through some kind of messaging infrastructure that you create between the iframe and the main window, rather than just using the localStorage, IndexedDB, etc. APIs directly in your main window's code.
I was unable to find a solution to the following problem: I want to display any website in an iframe and add elements (divs) to this iframe.
Due to the cross-scripting prevention in the browser this seems to be not possible.
Is there any way to do this?
If you do not actually host the domain in the iframe, no. Best you can do is have a server-side proxy page which copies the html and inserts divs.
What is your goal?
If you want to reformat the content or run scripts on their page then you probably need to proxy and copy the html (as meder suggested)
If you simply want to overlay or "insert" something in their page you can try placing a div on your own page and settings its position over top of the iframe where you want it.
I've had to do this to stick iframes over swf files before.
Cross-domain scripting rules explicitly prevent this. Even if you find a solution now, i'd be concerned about it in future, it's likely a hack. If you don't own/host the domain in your iFrame, it is forbidden.
A proxy or facade application that manipulates the HTML and serves that up might be possible, but I can't imagine it's easy.
It's likely that this is of no use to you, but the only sanctioned method I know of is to use HTAs, so if your application is running in a trusted environment (like a company intranet, or a kiosk, etc), you can take the HTA approach: http://msdn.microsoft.com/en-us/library/ms536496%28VS.85%29.aspx
Personally HTAs give me the creeps.
If you could do that, you could easily load any webpage into the iframe (possibly a website where the user has logged in) and extract content from there.
Putting the topic of security aside, there is another problem with your code: you need to create the element within the iframe's document, not the current document, and once again append it to the iframe's document, not the iframe object itself:
function addDiv() {
var doc = myFrame.contentDocument,
newdiv = doc.createElement("div");
newdiv.innerHTML = "foo";
doc.appendChild(newdiv);
}