Content Security Policy (CSP)

The modern browsers are able to interpret the Content-Security-Policy HTTP header that defines which dynamic assets are allowed to load on a given website. Alternatively, the CSP content can be sent using meta HTML tags.

Example:

// allow everything but only from the same origin:
default-src 'self'; 
// allow JS but only from example.eu and from the same origin:
script-src 'self' https://example.eu/myapp; 
// allow XMLHttpRequests but only from example.eu and the same origin:
connect-src 'self' https://example.eu/myapp; 

You can find all the possible directives here and a tool that can generate your CSP header here.

The amazing thing about CSP is the Report-URI attribute, which will report the deviations from the policy to the specified URL:

Content-Security-Policy: default-src 'self'; 
report-uri http://reportcollector.example.com/collector.cgi 

One of the services collecting such reports is report-uri.com.

Subresource_Integrity (SRI)

SRI is a very simple and effective concept: the modern browsers load a given asset only if its hash matches the one defined in the ‘integrity’ attribute.

So instead of doing this:

<script src="//www.example.com/script.js" type="text/javascript"></script>;

it’s recommended to do this:

<script src="//www.example.com/script.js" type="text/javascript"
integrity="sha256-Abhisa/nS9WMne/YX+dqiFINl+JiE15MCWvASJvVtIk="
crossorigin="anonymous"></script>;

or even better, link each version of the remote asset with its own URL and hash:

<script src="//www.example.com/1.0.1/script.js" type="text/javascript"
integrity="sha256-Cng8gUe98XCqh5hc8nAM3y5I1iQHBjzOl8X3/iAd4jE=" 
crossorigin="anonymous"></script>

Cross-Origin Resource Sharing (CORS)

Problem:
– a script on client.com wants to access some data from server.com (ex. XMLHttpRequest)
– by default, the same-origin browser policy blocks this request

Solution:
– but by adding some special response headers, server.com can allow the script client.com to access the data.

The modern browsers have implemented a mechanism allowing scripts (like XMLHTTPRequest) to make cross-domain requests. This is Cross-Origin Resource Sharing and it uses a relatively less used HTTP request method (OPTIONS) plus several response headers (Access-Control-Request-Method, Access-Control-Request-Headers, etc)

Resources from Mozilla Development Network (MDN):

Glossary: CSP, SRI, CORS

Technical details: CSP, SRI, CORS

Context

Over the weekend, hackers injected thousands of websites—including UK and US government sites—with code that hijacked visitors’ computers to mine cryptocurrency.

The attack, noticed on Sunday by security researcher Scott Helme, was pulled off by compromising a single plugin that was used by all of the affected sites: Browsealoud, a reputable suite of accessibility and translation tools. According to Helme, the plugin was edited by attackers to embed a script that uses a site visitor’s computer to do the complex math that generates new digital coins (in this case, Monero). This process, known as “mining,” can slow down the victim’s computer.

The attack loaded malicious Javascript onto visitors’ computers. The hackers behind the attack chose to mine cryptocurrency, but they had the power to do almost whatever they wanted.
Cryptocurrency Mining Hack That Compromised Thousands of Sites ‘Could Have Been a Catastrophe’

Scott Helme: Protect your site from Cryptojacking with CSP + SRI
Troy Hunt: Trust in Third Party Libraries

Convenience

In one of the most watched reviews of the iPhone X, Marques Brownlee was showing in a side-by-side comparison that the new FaceID is still slower than the TouchID, but it will eventually get better with software improvements. I think that in the most cases, FaceID is faster: by tapping the screen or pressing the power button to unlock your iPhone, you’re actually creating an extra step when you don’t need one. If you simply pick up the phone, then swiping up will already unlock it. Definitely faster than TouchID.
But what’s even more important, FaceID represents an important milestone in the history of biometric authentication: it’s for the first time when this process is done seamlessly, without even thinking about it. Even Brownlee admits that “facial recognition on a phone is closer to secure, seamless authentication than a separate fingerprint reader will ever be“.
After using FaceID to unlock my phone for more than two weeks I can feel that it works so well that it becomes easy to forget that the authentication actually happens. The closest to this feeling is the Apple Watch authentication model: once you put it on your hand and unlock it, it’s easy to forget that you’re authenticated (by keeping it tied to your hand) and you can receive notifications on it. The fact that the FaceID authentication is now tied to something that you’re doing with the phone anyway – that is, looking at it – makes the experience feel close to magic.

This experience can be technically described as ‘continuous authentication’: you no longer have to make a conscious gesture in order to authenticate (ex. type a PIN or put your finger on the TouchID); by simply looking at the phone you are already authenticated. Some examples:
– expand notifications and show the the control center on the lock screen
– apps using FaceID for authentication, like LastPass: after opening them, you’re already authenticated (without the need to type a password or put your finger).
– Safari browser auto-filling the passwords

Face ID is the most compelling advancement in security I have seen in a very long time. It’s game-changing not merely due to the raw technology, but also because of Apple’s design and implementation. […] The real Face ID revolution: since you’re almost always looking at your iPhone while you’re using it, Face ID enables what I call “continuous authentication.”
Rich Mogull: Face ID’s Innovation: Continuous Authentication

Security

I described above the convenience of using FaceID and how it slowly moves us to a new era where we no longer realize we’re using biometric authentication. But how secure is it?
Before we dive into the main security concerns, it’s worth noting that before TouchID (that is a mere 4 years ago), most of us did not use a passcode on our smartphones. “Before Touch ID, about half of our users had a passcode set. Now, 9 out of 10 do.” (WWDC June 2016). With the introduction of TouchID and now FaceID, the number of people enjoying the security benefits of using a passcode is much higher.

Here are the main concerns related to FaceID:

– accidental in-app buy or download; accidental use of Apple Pay: not possible, you have to double tap the power button to enable the FaceID in order to download an app or to make a purchase
– false positive rate (security concern) / false negative rate (convenience concern): according to Apple, both fare better than TouchID
– somebody clones your face: still possible, and a valid concern if you have a high profile. But FaceID is more secure than TouchID: fabricating a 3D mask is arguably more complicated than a 3D finger.
– law enforcement forces you to unlock your iPhone using your face: the security level here is the same as for Touch ID. If you are concerned about this aspect, then simply don’t use FaceID/TouchID and rely on a strong passcode instead. If you are a just a little bit concerned, then know that holding the power and one of the volume buttons together (‘squeeze’ your iPhone) will immediately require the passcode to unlock
– somebody unlocks your phone while you are asleep: you can enable ‘require attention’. But if you are concerned about this, you have bigger issues that the security on your smartphone 🙂
More about these concerns here: Face ID, Touch ID, No ID, PINs and Pragmatic Security

The only valid concern for me is the following: a thief steals your iPhone, makes you look at it to unlock, then runs away with the unlocked phone.
In this scenario, FaceID is actually worse than TouchID – because it requires no physical contact between you and the thief. The only solution for this is to ‘Find your iphone’ and remotely wipe it, but this can take time during which valuable data can be extracted from your stolen phone. A better solution would be to immediately lock the iPhone from your Apple Watch.

Conclusions

There is always a trade-off between security and convenience. Sometimes compromises need to be made: security requires pragmatism. But FaceID just proved how it can address the most security concerns while providing a seamless user experience. With time, software and hardware updates, it will get even better, and we will see FaceID on other computing devices like tablets or laptops.
And from there it’s easy to imagine a keyless future. How long until you unlock your car by looking at it? Or you unlock your house by simply approaching the door and disarm your house alarm system by stepping into the hallway?
FaceID is not only an important milestone for biometric authentication, but it has the potential to change the way we interact with technology forever.

Some insights about web security from the excellent ‘Bulletproof SSL and TLS‘ book from Ivan Ristic.

– the Internet was not designed with security in mind
– the first efforts to improve this were in 1995, when SSL3 was released (by Netscape). Then TLS followed up.
– TLS 1.2 – the most recent version – was released in August 2008 (its successor – TLS 1.3 – is still under development, as of September 2017)
– the goal back in 1995 was to enable e-commerce. Today we have that (‘commercial security’), but we want much more.

One of the biggest problem with the Web PKI is the approach to certificate validation. The following improvements aim to fix that:
1. Public Key Pinning:
– addresses the fact that any CA can issue a certificate for any domain name without the owner’s permission.
– with pinning, site owners can select (pin) one or more CAs that they trust, creating their own (small) trust ecosystem
– delivered via HTTP headers (Public-Key-Pins)

2. HTTP Strict Transport Security (HSTS)
– allows web servers to declare that web browsers should only interact with it using secure HTTPS connections, and never via the insecure HTTP protocol.
– also, when an invalid certificate is encountered, it instructs compliant browsers to replace warnings with errors, disallowing bypass.
– delivered also via HTTP headers (Strict-Transport-Security)

3. OCSP stapling
– There are two standards for certificate revocation (revocation is used when the private key is compromised or no longer needed)
– CRL (Certificate Revocation List) is a large list of revoked certificate serial numbers maintained by the CA. Because of the size, the lookup is slow
– OCSP (Online Certificate Status Protocol) allows browsers to obtain the revocation status of a single certificate.
– the problem with OCSP is that the individual lookup can slow down browsing and can affect the performance of the OCSP responder for high traffic websites
– to address this, OCSP stapling allows each server to embed an OCSP response directly into the TLS initial handshake

In the end, do not forget that the current Web PKI is controlled by 2 actors: Certificate Authorities (interested in profit) and browser vendors (interested in market share).

Neither group has strong security as its top priority, but they are not necessarily to blame—at least not entirely. They won’t give us security until we, the end users, start to demand it from them.

Sunday afternoon reading, from @ivanristic pic.twitter.com/jlD6W86GtZ

— Dorin Moise (@dorinmoise) September 24, 2017

Update 10 October 2017: I have just enabled HSTS here. Inspired by Troy Hunt.