jhoose

I have updated the module to automatically output the OWASP recommended security headers.

These headers are automatically added to the response but can be configured as required, or even disabled.

Code Configuration

        services.AddJhooseSecurity(_configuration, (securityOptions) => {
            
            // define the XFrame Options mode
            securityOptions.XFrameOptions.Mode = XFrameOptionsEnum.SameOrigin;
            
            // disable HSTS
            securityOptions.StrictTransportSecurity.Enabled = false;
        });

Configuration via appSettings

"JhooseSecurity": {
      "ExclusionPaths": [
        "/episerver"
      ],
      "HttpsRedirection": true,
      "StrictTransportSecurity": {
        "MaxAge": 31536000,
        "IncludeSubDomains": true
      },
      "XFrameOptions": {
        "Enabled": false,
        "Mode": 0,
        "Domain": ""
      },
      "XPermittedCrossDomainPolicies": {
        "Mode": 0
      },
      "ReferrerPolicy": {
        "Mode": 0
      },
      "CrossOriginEmbedderPolicy": {
        "Mode": 1
      },
      "CrossOriginOpenerPolicy": {
        "Mode": 2
      },
      "CrossOriginResourcePolicy": {
        "Mode": 1
      }
    }

Managing the server header

The security module doesn’t remove the ‘server header’, this may seem strange, but the approach differs depending on how you are hosting your site. I have included some examples below.

Another consideration, if you are hosting your solution with Optimizely DXP then the CDN will automatically remove the header.

Kestrel

return Host.CreateDefaultBuilder(args)
  .ConfigureCmsDefaults()
  .ConfigureWebHostDefaults(webBuilder =>
{
   webBuilder.ConfigureKestrel(o => o.AddServerHeader = false);
   webBuilder.UseStartup<Startup>();
});

IIS

<?xml version="1.0" encoding="UTF-8"?>
<configuration>
    <system.webServer>
        <security>
            <requestFiltering removeServerHeader="true" />
        </security>
    </system.webServer>
</configuration>

Installation

dotnet add package Jhoose.Security.Admin  --version 1.1.1.89

It has always been difficult to manage the CSP on a website, this new module for Optimizley aims to make the process easier giving control back to advanced editors.

Features

• Interface to manage policies.
• Global ‘report only’ mode, or specify for each policy.
• Add ‘nonce’ to inline script or style tags.
• Ability to specify paths that are excluded from outputting the policy header.

Administration

Once the module is installed you will see a new ‘Security’ menu item within the top menu.

Settings

This screen gives you access to the global settings of the module, allowing the module to be enabled/disabled or switched into ‘Report Only’ mode.

It is also possible to specify an endpoint for a reporting service.

View Policies

All security policies are listed, with a summary of the policy configuration. A user is then able to click on a policy to view the policy in greater detail or amend it as required.

Edit Policy

This screen allows for an individual policy to be managed by the user, these will be saved when the ‘OK’ button is pressed.

When changes are made it is recommended that they are tested in ‘Report Only’ mode to ensure that nothing is adversely impacted by the new configuration.

Installation

Install the package directly from the Optimizley Nuget repository. This will install the admin interface along with the middleware to add the CSP header to the response.

Github: https://github.com/andrewmarkham/contentsecuritypolicy

dotnet add package Jhoose.Security.Admin

Configuration

Startup.cs

services.AddJhooseSecurity(IConfiguration configuration, Action<SecurityOptions> options = null);

The Action<SecurityOptions> options is optional and if not specified then the default will be used.

  "JhooseSecurity": {
    "ExclusionPaths": [
      "/episerver"
    ]
  }

ExclusionPaths: Any request which starts with a path specified in this property will not include the CSP header.

app.UseJhooseSecurity();

Nonce Tag Helper

It is possible to get a nonce added to your inline <script> and <style> tags.

_ViewImports.cshtml

@addTagHelper *, Jhoose.Security.Core

<script nonce src="/assets/js/jquery.min.js"></script>

Introduction

The ‘Content Security Policy’ response header is used to enhance a website’s security. It allows control of how resources are loaded, whitelisting trusted domains.

Any security audit will highlight this as a key recommendation.

Content-Security-Policy: default-src 'self' ; script-src https://www.google-analytics.com https://ssl.google-analytics.com; img-src https://www.google-analytics.com; connect-src https://www.google-analytics.com;

Read More: https://content-security-policy.com/

Why

Content Security Policy is an extra layer of security that protects against certain types of attack vectors (XSS, Click Jacking), anything you can do to protect the end-user is beneficial.

It is common practice these days for developers to draw in 3rd party packages to deliver a feature and these packages will also draw in other packages. This means that malicious code can get drawn into your solution without your knowledge.

Read More : https://medium.com/@alex.birsan/dependency-confusion-4a5d60fec610

How

It is possible to be very granular when configuring the security policy. This gives you a high level of control of what can be loaded and where data can be sent.

Although the Content Security Policy can be added at a meta tag, it is typically delivered in a response header, additionally, it is also possible to enable specific inline scripts by adding a nonce value.

Testing your policy

Changes to the policy can inadvertently disable features on your site. YouTube videos can fail to load, analytics stops sending data, javascript components can fail. To help mitigate this you should either make any changes as ‘report only’ and monitor the output or configure the policy to report issues. This means that any issues are either reported within the browser console or are sent to a 3^rd party reporting service.

Challenges

You have deployed your policy, it’s been tested and signed off, but then things start to fail. Unfortunately, things do change, a dependency can update the endpoints they use or more typically a new tool is added via Google Tag Manager.

Managing the content security policy can be challenging. A common approach is to create a Url Rewrite outbound rule. This means that when the policy needs to change you have to update your source code and redeploy your site, not very efficient.

Is there an easier way?

Read my next post where I will introduce the new Episerver / Optimizely module that simplifies the management of the Content Security Policy.

Next Post: https://jhoose.co.uk/2021/11/01/introducing-jhoose-security-a-module-to-manage-your-content-security-policy/