In a recent project, I wanted to share as much logic as possible between the server, service worker and the client. The project is essentially a simple RSS feed reader, it takes RSS feeds, parses the data and merges them in to a nice set of columns (much like TweetDeck), and also a single merged list.
Because I am taking RSS feeds and displaying in my page, I need to be as sure as possible that it is not doing anything nefarious. I can sanitize the input as much as I want, however I know my own abilities, and I am certain people could manipulate an RSS feed in such a way that I would end up running scripts, importing images or any other 3rd-party in the context of my site.
The web platform offers ability to lock down a site via Content-Security-Policy (CSP). CSP can lock down the external sources from which we can request context such as script, styles, images etc. You can even lock down the ability for a page to run scripts in-line - which can prevent all manor of XSS types of attacks.
It was pretty simple to add it to the app.
`default-src 'self';`
However…. I had a number of issues.
- I generate styles inline on the page and thus I needed to run scripts inline.
- I needed to include Google Analytics which requires an inline script to be run on the page.
CSP lets you run scripts and styles inline by letting you turn on an option called unsafe-eval
of scripts, however this pretty much by-passes any protections that CSP affords.
To run inline scripts and still have the protections of CSP, CSP offers a couple of tools. The one I used is called a ‘nonce’. The nonce is a random id that you set on the CSP HTTP header and that you tally with an associated inline script.
CSP string on HTTP Header
`default-src 'self'; script-src 'self' https://www.googletagmanager.com https://www.google-analytics.com 'nonce-script-${nonce.analytics}'; connect-src 'self'; img-src 'self' data: https://www.google-analytics.com; style-src 'self' 'nonce-style-${nonce.style}'
Inline script using nonce
<script src="https://www.googletagmanager.com/gtag/js?id=1111"></script>
<script nonce="script-{nonce.analytics}">
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', '{{=it.config.then(config=>config.site.googleAnalytics)}}');
</script>
The above code works well and makes it simple to get analytics working correctly when we are securing the site with CSP.
For every single web request, you need to have a unique ‘nonce’ value and I do this via the {nonce.analytics}
which is a value that I generate on the server and apply via a template. If you re-use a nonce value the browser will refuse to execute the content in the script.
I had a little trouble generating nonce values. I needed something that would create a unique value that wont be re-used by the same user. I felt that a nonce value of the format ‘[source]-[date.now + request-count]’ would suffice.
The ‘source’ allows me to add a namespace to the nonce, and date.now() + an ever increasing request count gives me a relatively stable non-repeatable set of values.
I generate the nonce using the following function:
function generateIncrementalNonce(source) {
let val = 0;
let max = Math.pow(10, 3); // Date + pow 3 gets us close to max number;
const generate = () => {
let now = max * +new Date();
if(val >= max) val = 0;
else val++;
return (source !== undefined ? source : '') + (now + val).toString();
}
return generate;
};
Looks good. However, I cache all my pages in a service worker, which means that if I just simply served the content from the cache the nonce values will be reused and thus not executed.
Luckily, I am share logic between my server and service worker, which allows me to generate anything that I need in one central place of my code. I use the ‘source’ parameter in my generateIncrementalNonce
function to prepend ‘server’ or ‘service-worker’ to the nonce value and I did this in each of the request handlers in both the server and service worker. Using this source parameter means that I can guarantee a nonce value generated via the server will never clash with a page loaded via the service worker.
This pattern has served me well. It has allowed me to allow the required inline scripts for Google Analytics whilst stopping any third party from injecting or running untrusted code in my page.
Below is the code that I used in the project. There are a number of different places in my pages that I need nonce values, I generate them for each request and then apply it to my templating function and the HTTP header at the same time.
common.js - shared logic
function generateCSPPolicy(nonce) {
return `default-src 'self'; script-src 'self' https://www.googletagmanager.com https://www.google-analytics.com 'nonce-script-${nonce.analytics}'; connect-src 'self'; img-src 'self' data: https://www.google-analytics.com; style-src 'self' 'nonce-style-${nonce.style}' 'nonce-style-${nonce.inlinedcss}';`;
};
function generateIncrementalNonce(source) {
let val = 0;
let max = Math.pow(10, 3); // Date + pow 3 gets us close to max number;
const generate = () => {
let now = max * +new Date();
if(val >= max) val = 0;
else val++;
return (source !== undefined ? source : '') + (now + val).toString();
}
return generate;
};
service-worker.js - fetch handler
const generator = generateIncrementalNonce('service-worker');
let nonce = {
analytics: generator(),
inlinedcss: generator(),
style: generator()
};
// Call the route handler with all data needed
let response = all(nonce, {
dataPath: paths.dataPath,
assetPath: paths.assetPath
}).then(r => setHeader(r, 'Content-Security-Policy', generateCSPPolicy(nonce)));;
e.respondWith(response);
server.js - request handler
const generator = generateIncrementalNonce('server');
let nonce = {
analytics: generator(),
inlinedcss: generator(),
style: generator()
};
res.setHeader('Content-Security-Policy', generateCSPPolicy(nonce));
// Call the route handler with all data needed
all(nonce, {
dataPath: `${paths.dataPath}${hostname}.`,
assetPath: paths.assetPath
})
.then(response => {
node.responseToExpressStream(res, response.body)
});
Top comments (0)