Simply add the middleware to your ASP.NET Core application by configuring it as part of your normal Startup pipeline. Note that the order of middleware matters, so to apply the headers to all requests it should be configured first in your pipeline.
To use the default security headers for your application, add the middleware using:
public void Configure(IApplicationBuilder app)
{
app.UseSecurityHeaders();
// other middleware e.g. static files, MVC etc
}
This adds the following headers to all responses that pass through the middleware:
X-Content-Type-Options: nosniff
Strict-Transport-Security: max-age=31536000; includeSubDomains - only applied to HTTPS responses
X-Frame-Options: Deny - only applied to "document" responses
X-XSS-Protection: 1; mode=block - only applied to "document" responses
Referrer-Policy: strict-origin-when-cross-origin - only applied to "document" responses
Content-Security-Policy: object-src 'none'; form-action 'self'; frame-ancestors 'none' - only applied to "document" responses
"Document" responses are defined as responses that return one of the following content-types:
text/html
text/javascript
application/javascript
Customising the security headers added to responses
To customise the headers returned, you should create an instance of a HeaderPolicyCollection and add the required policies to it. There are helper methods for adding a number of security-focused header values to the collection, or you can alternatively add any header by using the CustomHeader type. For example, the following would set a number of security headers, and a custom header X-My-Test-Header.
public void Configure(IApplicationBuilder app)
{
var policyCollection = new HeaderPolicyCollection()
.AddFrameOptionsDeny()
.AddXssProtectionBlock()
.AddContentTypeOptionsNoSniff()
.AddStrictTransportSecurityMaxAgeIncludeSubDomains(maxAgeInSeconds: 60 * 60 * 24 * 365) // maxage = one year in seconds
.AddReferrerPolicyStrictOriginWhenCrossOrigin()
.RemoveServerHeader()
.AddContentSecurityPolicy(builder =>
{
builder.AddObjectSrc().None();
builder.AddFormAction().Self();
builder.AddFrameAncestors().None();
})
.AddCrossOriginOpenerPolicy(builder =>
{
builder.SameOrigin();
})
.AddCrossOriginEmbedderPolicy(builder =>
{
builder.RequireCorp();
})
.AddCrossOriginResourcePolicy(builder =>
{
builder.SameOrigin();
})
.AddCustomHeader("X-My-Test-Header", "Header value");
app.UseSecurityHeaders(policyCollection);
// other middleware e.g. static files, MVC etc
}
The security headers above are also encapsulated in another extension method, so you could rewrite it more tersely using
public void Configure(IApplicationBuilder app)
{
var policyCollection = new HeaderPolicyCollection()
.AddDefaultSecurityHeaders()
.AddCustomHeader("X-My-Test-Header", "Header value");
app.UseSecurityHeaders(policyCollection);
// other middleware e.g. static files, MVC etc
}
If you want to use the default security headers, but change one specific header, you can simply add another header to the default collection. For example, the following uses the default headers, but changes the max-age on the Strict-Transport-Security header:
public void Configure(IApplicationBuilder app)
{
var policyCollection = new HeaderPolicyCollection()
.AddDefaultSecurityHeaders()
.AddStrictTransportSecurityMaxAgeIncludeSubDomains(maxAgeInSeconds: 63072000);
app.UseSecurityHeaders(policyCollection);
// other middleware e.g. static files, MVC etc
}
There is also a convenience overload for UseSecurityHeaders that takes an Action<HeaderPolicyCollection>, instead of requiring you to instantiate a HeaderPolicyCollection yourself:
public void Configure(IApplicationBuilder app)
{
app.UseSecurityHeaders(policies =>
policies
.AddDefaultSecurityHeaders()
.AddStrictTransportSecurityMaxAgeIncludeSubDomains(maxAgeInSeconds: 63072000)
);
// other middleware e.g. static files, MVC etc
}
RemoveServerHeader
One point to be aware of is that the RemoveServerHeader method will rarely (ever?) be sufficient to remove the Server header from your output. If any subsequent middleware in your application pipeline add the header, then this will be able to remove it. However Kestrel will generally add the Server header too late in the pipeline to be able to modify it.
Luckily, Kestrel exposes it's own mechanism to allow you to prevent it being added:
var host = new WebHostBuilder()
.UseKestrel(options => options.AddServerHeader = false)
//...
In Program.cs, when constructing your app's WebHostBuilder, configure the KestrelServerOptions to prevent the Server tag being added.
AddContentSecurityPolicy
The Content-Security-Policy (CSP) header is a very powerful header that can protect your website from a wide range of attacks. However, it's also totally possible to create a CSP header that completely breaks your app.
The CSP has a dizzying array of options, only some of which are implemented in this project. Consequently, I highly recommend reading this post by Scott Helme, in which he discusses the impact of each "directive". I also highly recommend using the "report only" version of the header when you start. This won't break your site, but will report instances that it would be broken, by providing reports to a service such as report-uri.com.
Set the header to report-only by using the AddContentSecurityPolicyReportOnly() extension. For example:
public void Configure(IApplicationBuilder app)
{
var policyCollection = new HeaderPolicyCollection()
.AddContentSecurityPolicyReportOnly(builder => // report-only
{
// configure policies
});
}
or by by passing true to the AddContentSecurityPolicy command
public void Configure(IApplicationBuilder app)
{
var policyCollection = new HeaderPolicyCollection()
.AddContentSecurityPolicy(builder =>
{
// configure policies
},
asReportOnly: true); // report-only
}
You configure your CSP policy when you configure your HeaderPolicyCollection in Startup.Configure. For example:
The permissions-policy is a header that allows a site to control which features and APIs can be used in the browser. It is similar to CSP but controls features instead of security behaviour.
With Permissions-Policy, you opt-in to a set of "policies" for the browser to enforce on specific features used throughout a website. These policies restrict what APIs the site can access or modify the browser's default behaviour for certain features.
By adding Permissions-Policy to headers to your website, you can ensure that sensitive APIs like geolocation or the camera cannot be used, even if your site is otherwise compromised, for example by malicious third-party attacks.
For more information about the permissions, I recommend the following resources:
You configure your CSP policy when you configure your HeaderPolicyCollection in Startup.Configure. For example:
public void Configure(IApplicationBuilder app)
{
var policyCollection = new HeaderPolicyCollection()
.AddPermissionsPolicy(builder =>
{
builder.AddAccelerometer() // accelerometer 'self' http://testUrl.com
.Self()
.For("http://testUrl.com");
builder.AddAmbientLightSensor() // ambient-light-sensor 'self' http://testUrl.com
.Self()
.For("http://testUrl.com");
builder.AddAutoplay() // autoplay 'self'
.Self();
builder.AddCamera() // camera 'none'
.None();
builder.AddEncryptedMedia() // encrypted-media 'self'
.Self();
builder.AddFullscreen() // fullscreen *:
.All();
builder.AddGeolocation() // geolocation 'none'
.None();
builder.AddGyroscope() // gyroscope 'none'
.None();
builder.AddMagnetometer() // magnetometer 'none'
.None();
builder.AddMicrophone() // microphone 'none'
.None();
builder.AddMidi() // midi 'none'
.None();
builder.AddPayment() // payment 'none'
.None();
builder.AddPictureInPicture() // picture-in-picture 'none'
.None();
builder.AddSpeaker() // speaker 'none'
.None();
builder.AddSyncXHR() // sync-xhr 'none'
.None();
builder.AddUsb() // usb 'none'
.None();
builder.AddVR() // vr 'none'
.None();
// You can also add arbitrary extra directives: plugin-types application/x-shockwave-flash"
builder.AddCustomFeature("plugin-types", "application/x-shockwave-flash");
// If a new feature policy is added that follows the standard conventions, you can use this overload
// iframe 'self' http://testUrl.com
builder.AddCustomFeature("iframe") //
.Self()
.For("http://testUrl.com");
});
app.UseSecurityHeaders(policyCollection);
// other middleware e.g. static files, MVC etc
}
Using Nonces and generated-hashes with Content-Security-Policy
The use of a secure Content-Security-Policy can sometimes be problematic when you need to include inline-scripts, styles, or other objects that haven't been whitelisted. You can achieve this in two ways - using a "nonce" (or "number-used-once"), or specifying the hash of the content to include.
To help with this you can install the FSH.AspNetCore.SecurityHeaders.TagHelpers package, which provides helpers for generating a nonce per request, which is attached to the HTML element, and included in the CSP header. A similar method helper exists for <style> and <script> tags, which will take a SHA256 hash of the contents of the HTML element and add it to the CSP whitelist.
To use a nonce or an auto-generated hash with your ASP.NET Core application, use the following steps.
1. Install the FSH.AspNetCore.SecurityHeaders.TagHelpers NuGet package, e.g.
Configure your security headers in the usual way. Use the WithNonce() extension method when configuring ContentSecurityPolicy directives to allow whitelisting with a nonce. Use the WithHashTagHelper() extension methods on script-src and style-src directives to allow automatic generation of whitelisted inline-scripts
public void Configure(IApplicationBuilder app)
{
var policyCollection = new HeaderPolicyCollection()
.AddContentSecurityPolicy(builder =>
{
builder.AddUpgradeInsecureRequests();
builder.AddDefaultSrc() // default-src 'self' http://testUrl.com
.Self()
.From("http://testUrl.com");
builder.AddScriptSrc() // script-src 'self' 'unsafe-inline' 'nonce-<base64-value>'
.Self()
.UnsafeInline()
.WithNonce(); // Allow elements marked with a nonce attribute
builder.AddStyleSrc() // style-src 'self' 'strict-dynamic' 'sha256-<base64-value>'
.Self()
.StrictDynamic()
.WithHashTagHelper(); // Allow whitelsited elements based on their SHA256 hash value
})
.AddCustomHeader("X-My-Test-Header", "Header value");
app.UseSecurityHeaders(policyCollection);
// other middleware e.g. static files, MVC etc
}
3. Add a using directive for the TagHelpers
Add the following to the _ViewImports.cshtml file in your application. This makes the tag-helper available in your Razor views.
Add the NonceTagHelper to an element by adding the asp-add-nonce attribute.
<script asp-add-nonce>
var body = document.getElementsByTagName('body')[0];
var div = document.createElement('div');
div.innerText = "I was added using the NonceHelper";
body.appendChild(div);
</script>
This will use a unique value per-request and attach the required attribute at runtime, to generate markup similar to the following:
<script nonce="ryPzmoZScSR2xOwV0qTU9mFdFwGPN+gy3S2E1/VK1vg=">
var body = document.getElementsByTagName('body')[0];
var blink = document.createElement('div');
blink.innerText = "And I was added using the NonceHelper";
body.appendChild(blink);
</script>
Note that some browsers will hide the value of the nonce attribute when viewed from DevTools. View the page source to see the raw nonce value
While the CSP policy would look something like the following:
To use a whitelisted hash instead, use the HashTagHelper, by adding the asp-add-content-to-csp attribute to <script> or <style> tags. You can optionally add the csp-hash-type attribute to choose between SHA256, SHA384, and SHA512:
