Revision as of 19:20, 30 October 2014

DRAFT DOCUMENT - WORK IN PROGRESS

Description

Cross-site scripting attacks exploit vulnerabilities in web page validation by injecting client-side script code. The script code embeds itself in response data, which is sent back to an unsuspecting user. In addition to validating input, any data retrieved from untrusted or shared sources should be encoded on output. For example: data retrieved from a database that may have had malicious input persisted to it.

Validating Input

See the ASP.NET Request Validation article for details on how request validation can be used to protect against malicious input.

Encoding Output Values in Code

Use Server.HtmlEncode to encode untrusted data for use in HTML output:

var encodedHtml = Server.HtmlEncode(untrustedData);

Use Server.UrlEncode to encode untrusted data for use in constructing URLs

var encodedUrl = Server.UrlEncode(untrustedData);

Use Server.UrlTokenEncode to encode untrusted data in byte array form for use as a URL parameter

var encodedUrlToken = Server.UrlTokenEncode(untrustedData);

Encoding Output Values in HTML markup

You can HTML encode the value in markup with the <%: %> syntax, as shown below.

<span><%: untrustedData%></span>

Or, in Razor syntax, you can HTML encode with @, as shown below.

<span>@untrustedData</span>

Preventing Double Encoding

You may run into a scenario where encoding values results in them becoming double encoded on output. ASP.NET provides the HtmlString and MvcHtmlString classes starting with .NET 4.0 which are designed to help. These both implement the IHtmlString interface which will instruct ASP.NET to not apply encoding within HTML markup when using <%: untrustedData %> or @untrustedData. Converting a property on your view model from String to MvcHtmlString will instruct ASP.NET that HTML encoding has already been accounted for.

public class User 
{
    public int Id { get; set; }
    public string Name { get; set; }
    public MvcHtmlString Description { get; set; } // Output encoding is handled manually
}

Enhanced Encoding

By default the ASP.NET encoding methods use a black-listing technique that evaluates the string for a set of character combinations that may indicate presence of malicious script. A superior approach is to use a white-listing technique for validation, which can be achieved using the Anti-Cross Site Scripting Library from Microsoft. Starting with ASP.NET 4.5 you can specify that the AntiXssEncoder from this library be used as the default encoder for you entire application using the encoderType setting in web.config as shown below.

<httpRuntime encoderType="System.Web.Security.AntiXss.AntiXssEncoder" />

If you are using a version of .NET earlier than 4.5, you will need to download and include the library as a reference to your project, and then use the earlier library name for the encodeType setting as shown below.

<httpRuntime encoderType="Microsoft.Security.Application.AntiXssEncoder, AntiXssLibrary" />

In addition to the common HtmlEncode and UrlEncode methods, the Anti-Cross Site Scripting Library provides the following AntiXssEncoder methods for more specialized output encoding needs:

References

• Output Encoding
• New <%: %> Syntax for HTML Encoding Output in ASP.NET 4 (and ASP.NET MVC 2)
• IHtmlString Interface
• What is an MvcHtmlString and when should I use it?
• Microsoft Anti-Cross Site Scripting Library V4.3
• AntiXssEncoder Methods