My current content security policy section of code allows inline. The app blocks rendering of the charts during the production build. How should I add permissions for google charts I can't find anything regarding this on their page:
https://www.npmjs.com/package/react-google-charts#installation
My current code (manifest.json):
"content_security_policy": "script-src 'self' 'sha256-/8hmd4DqwVs419dLsNQ5IDmaKuXZyjZnNQcgWS+mCk4='; object-src 'self'",
Needed to add:
<script src="https://www.gstatic.com/charts/loader.js"></script>
to the header to make it work
Related
I want to enforce CSP as a security measure in my web application.
From the server end , I have set the policy to "allow" self for all of its resources.
However there is one particular front end node_module file , which is throwing error as attached below.
The CSP header set is :
script-src 'self' 'node_modules/aurelia-webpack-plugin/runtime/empty-entry.js'; script-src-elem 'self'; style-src 'self'; img-src 'self'
Please help , I have been trying to find a solution for more than a week now!
Solutions tried:
Try the front end to make it ignore this file since it is an empty file. But I am not able to get it to ignore.
Trying from server end to bypass this particular file by changing the rules
TIA.
Update:
Error message
UnCaught EvalError: Refused to evaluate a string as Javascript because 'unsafe-eval' is not an allowed source of script in the following Content Security Policy directive: "script-src 'self'"
at ./node_modules/aurelia-webpack-plugin/runtime/empty-entry.js
The file seems to be allowed to load as it is loaded from the same source and you have allowed script-src 'self'.
The problem seems to be that the code in the file does eval(), new Function(), setInterval() or setTimeout(), which requires 'unsafe-eval' to be allowed. This is strange given that the file should be empty. Your console errer may provide you with a direct link to the offending code.
You could add 'unsafe-eval' to script-src. This would make your CSP less strict, but it is of course a lot better to set "script-src 'self' 'unsafe-eval';" than to not restrict scripts with a CSP at all.
I have ASP.NET Core application and i am using Telerik's UI for ASP.NET Core framework for certain widgets like Date, DropDownList, Charts etc.
Application is loading all javascripts, images, css from its own server. So i have enabled CSP policy as below
script-src 'self' 'unsafe-eval';
style-src 'self' 'unsafe-inline';
img-src 'self' data:;
font-src 'self';
media-src 'none';
object-src 'none';
child-src https://xxxx.yyyy.com;
report-uri http://myapplication/csp/report;
However when the page load i see error in chrome's console
Refused to execute inline script because it violates the following
Content Security Policy directive: "script-src 'self' 'unsafe-eval'".
Either the 'unsafe-inline' keyword, a hash
('sha256-oii70XYoqukWS9204nbwatxgYOYcr06+rftc4egdfUk='), or a nonce
('nonce-...') is required to enable inline execution
this error is repeated for several Kendo widgets that i'm using on that page. For example date widget. In cshtml i have configured date as below
#(Html.Kendo().DatePicker()
.Name("sbDate")
.HtmlAttributes(new { #class = "mydate" }))
which gets render in html as
<input class="mydate" id="sbDate" name="sbDate" type="date" value="" /><script>jQuery(function(){jQuery("#sbDate").kendoDatePicker({"format":"M/d/yyyy","footer":false});});</script>
I have already gone through Telerik's suggestion here and here for CSP. And as per the suggestion i only i have to add unsafe-eval to get widgets working. However looks like that is not true. I also have to add unsafe-inline to get widgets working.
But the whole point of enabling CSP is to not allow inline scripts.
Questions
Is there any way to solve this issue?
Update 1
Widgets only works in IE 11 without adding unsafe-inline. In chrome and IE edge widget requires unsafe-inline to work.
The solution below might work (i have not tried yet)
1> Add Deferred method for each kendo widget. Like
#(Html.Kendo().DatePicker().Name("BeginDate").Deferred())
2> In each view that is using Kendo control, at the bottom add the script tag
<script asp-add-nonce="true">
#Html.Kendo().DeferredScripts(false)
</script>
Note that asp-add-nonce is not out of the box from asp.core. You have to install Joonasw.AspNetCore.SecurityHeaders nuget library. I found this article from the author that shows how to configure CSP for .net core application and also how to generate new nonce for each request.
3>Dont configure CSP policy in web.config because we need different nonce value for each request, instead configure the CSP policy using Joonasw.AspNetCore.SecurityHeaders middleware as mentioned in the article.
4>I thinks we don't need to add unsafe-inline if we use random nonce (unless you are targeting to older browsers)
I can't really tell if this would solve your issue, but in our Application, we are using the Deferring feature, but for other reasons. Our reason is that we load all scripts (esp. jquery.js) at the end of the page. But the side effect is that no jQuery script calls are rendered into the page anymore. You can control the place to generate all scripts by executing #Html.Kendo().DeferredScripts() (but you will find more help in the docs above). Anyway, probably in the context of CSP, even that <script> block at the very end of the page is still considered "inline". But give it a try.
The Zemanta Chrome Extension fails to load with the following error message
loader.js:13 Refused to load the script 'https://static.zemanta.com/widgets/blogger.com/merged-blogger.js?v=1451290656'
because it violates the following Content Security Policy directive:
"script-src 'self' *.google.com *.google-analytics.com 'unsafe-inline'
'unsafe-eval' *.gstatic.com *.googlesyndication.com *.blogger.com
*.googleapis.com uds.googleusercontent.com https://s.ytimg.com
www-onepick-opensocial.googleusercontent.com www-bloggervideo-opensocial.googleusercontent.com
www-blogger-opensocial.googleusercontent.com *.blogspot.com https://www.blogblog.com".
I have changed the content_securiy_policy line to the following, and reload the extension
"content_security_policy": "script-src 'self' https://ssl.google-analytics.com https://*.zemanta.com; object-src 'self'"
Why is it that "https://*.zemanta.com" is not listed in the CSP directive in the error message above, and how can I make sure it is in the CSP?
It looks like an error thrown by a webpage with its own CSP, not inside your extension. So the CSP you set there has no effect.
It probably happens as a result of a content script trying to insert <script src="..."> into the page. That's subject to the page's CSP and can fail.
You can try to bypass the page's CSP by loading the script with XHR and inserting a <script> tag with the code included instead of src link. Note that this may fail at a later stage, since while the script will be executed this way, it will be subject to CSP in its own actions (so if it, say, tries to add a <script> tag as well, it will fail).
Alternatively, you could use webRequest API to intercept and modify the CSP header. That's risky since you're loosening the page's security in general.
I am writting a chrome extension that needs to have two domains in its whitelist for the content security policy. I've looked at the official docs, but I still can't seem to figure out the proper syntax.
The following does not seem to work:
"content_security_policy": "script-src 'self' https://foo.com https://example.com; object-src 'self'"
EDIT:
Both my content script and my popup are able to reach foo.com, however, neither can reach example.com.
Are chrome extensions capable of having multiple sources whitelisted in the CSP?
From what I know about CSPs, this looks syntactically correct. The HTML5 Rocks article on CSP agrees with your syntax, saying:
script-src https://host1.com https://host2.com would correctly specify both origins as valid.
However, your problem may be that either:
This CSP disallows all subdomains, including www.foo.com and www.example.com. You can add those subdomain hostnames explicitly, or you can use https://*.foo.com to allow all subdomains.
If any of your script requests redirect to a non-permitted domain, the request will fail. For example, if https://example.com/foo.js responds with a 301 or 302 redirect to https://notpermitted.com/foo.js (not-permitted origin) or https://www.example.com/foo.js (non-permitted subdomain), the request will fail according to the spec:
Whenever the user agent fetches a URI (including when following redirects)... if the URI does not match the allowed script sources, the user agent must act as if it had received an empty HTTP 400 response...
EDIT:
Just to confirm, yes, Chrome extensions can whitelist multiple HTTPS origins. You can build a simple extension to test this:
manifest.json
{
"name":"CSP Test",
"version":"1.0",
"manifest_version":2,
"browser_action":{
"default_popup":"csp_test.html"
},
"content_security_policy": "script-src 'self' https://www.iana.org https://ajax.googleapis.com; object-src 'self'"
}
csp_test.html
<script src="https://www.iana.org/_js/2013.1/jquery.js"></script>
<script src="https://ajax.googleapis.com/ajax/libs/jqueryui/1.10.3/jquery-ui.min.js"></script>
<script src="csp_test.js"></script>
csp_test.js
alert(jQuery)
alert(jQuery.ui)
This extension loads jQuery and jQuery UI from remote domains. If you remove either origin from the CSP, you will see an "undefined" alert signifying that one of the libraries failed to load.
I have face the same problem to whitelist the secure resources URL's with the below warning.
There were warnings when trying to install this extension:
« Ignored insecure CSP value "object-src" in directive 'script-src'.
« CSP directive 'object-src' must be specified (either explicitly, or implicitly via 'default-src') and must whitelist only secure resources.
To resolve HTTP Content-Security-Policy use below key value in manifest.json file.
{
"content_security_policy":
"script-src 'self' https://yash.test.com:8443 'unsafe-eval'; object-src 'self' https://yash.test.com:8443"
}
I'm trying to use Firebase in a Chrome extension background page, but it looks like it's executing inline-scripts, which isn't allowed because of security concerns.
I've currently set the CSP to:
{"content_security_policy":
"script-src 'self' https://cdn.firebase.com https://<my-subdomain>.firebaseio.com; object-src 'self'"}
I'm able to load the initial Firebase script, but upon calling new Firebase('my-firebase-url'), I get the following error:
Refused to execute inline script because it violates the following
Content Security Policy directive: ". Uncaught
ReferenceError: pRTLPCB is not defined
Is there any work around or advice the Firebase team (or anyone) can give, and maybe an explanation of why scripts are being executed inline?
At the time the question was asked, there was a bug preventing Firebase from working in Chrome extensions, but this has now been fixed.
The correct CSP is:
"content_security_policy": "script-src 'self' https://cdn.firebase.com https://*.firebaseio.com; object-src 'self'"
(Note that the wildcard in the domain is important, since Firebase may connect to other subdomains internally.)
For a sample chrome extension using Firebase, see: https://github.com/firebase/firebase-chrome-extension.
I'm having a similar problem; you see, Firebase's constructor seems to perform some dom manipulation in order to do some scripting (vague, I know), which triggers Chrome CSP because well, you are not supposed to do that.
I even trying to wrap the constructor through the Sandbox Pages, but without success (I get an DOM ERR 18, even though my manifest has all permissions). Same happens if you try to do it in a Background Page or a Popup Page for a Page Action/Browser Action.
Alternative? You can inject Firebase as a content script (do it from your manifest), and through Message Passing send the callbacks as Chrome.extension.sendMessage. I'm exactly doing this at the moment, so I can tell you how that goes, so far, at least the Firebase constructor works.
Solution? James Tampling reads this and prompts the Firebase team to look up after this :)
UPDATE: Injecting Firebase.js as a Content Script doesn't work neither, but the good news is that the Firebase team (reach Andrew Lee) is checking it out.
UPDATE 2 Firebase team fixed it, and now it does work from a Popup page (both in a Browser Popup or a Page Action one). You need to add the following CSP in your manifest.json though "content_security_policy": "script-src 'self' 'unsafe-eval' https://cdn.firebase.com https://*.firebaseio.com https://*.firebaseio.com; object-src 'self'; " It works wonders after that.
I have a cordova js app, only the below one worked:
<meta http-equiv="Content-Security-Policy" content="
default-src 'self' data: gap: https://ssl.gstatic.com 'unsafe-eval';
script-src 'self' https://www.gstatic.com https://cdn.firebase.com https://*.firebaseio.com; object-src 'self';
" />
hope this helps.