December 2, 2021

Compromised Sites and Malicious Registrations: Best Practices for the Identification and Mitigation of DNS Abuse


by Graeme Bunton, Director of the DNS Abuse Institute

One of the more interesting revelations from University of Grenoble Associate Professor  Maciej Korczyński during the DNS Abuse Institute’s 2nd Online forum was that a substantial percentage of domains involved in malware and phishing were not registered specifically to cause harm, but were in fact benign domains being used by websites that had been compromised.  

It is important to distinguish between malicious and benign domains because the approaches to mitigation should be different. Removing malicious domains from the Internet rarely causes any collateral damage. However, acting at the DNS level upon a benignly registered domain with a subsequently compromised website may have commercial impacts for a business such as preventing email and other services from working, as well as non-commercial impacts on expression.

In this blog post, we’re going to describe what is meant by malicious registrations, how to manually distinguish between malicious registrations and compromised websites, and what the best practices are for mitigation in each circumstance.  

You can download and share a condensed version of this post as DNS Abuse Institute Best Practice.


Malicious registrations are domain names that are registered for the explicit purpose of causing harm. In the context of DNS Abuse, that means the domain was registered for: 

  • phishing, 
  • the distribution of malware, 
  • the operation of a botnet, 
  • conducting pharming, 
  • engaging in the distribution of spam in the service of one of those previous four harms, or
  • A combination of these harms.

A compromised website is where a domain is registered and a website created for a legitimate or benign purpose that is subsequently exploited by a third party and used for conducting online harms and typically without the registrant’s knowledge. In colloquial terms, the website has been hacked and is being used for undesired activity.  The most common example of this occurs when a website using a content management system such as WordPress, Joomla, or Drupal is left un-updated and becomes insecure, uses insecure plugins, or uses an insecure theme and is vulnerable to compromise.

Distinguishing Between Malicious Registrations and Compromised Websites

It’s important to distinguish between these two alternatives, because the response should be different for each type.  We’ll discuss how to differentiate before sharing remedy best practices for each type of harm.  

Recent work by the COMAR (Classification of Compromised versus Maliciously Registered domains) project identified some 38 different domain or website attributes that could be useful for making this distinction.  These attributes are important for algorithms to make distinctions, but only a subset of the attributes will be readily available or practical for a manual review. We’ve selected the following attributes due to both their relative usefulness and availability.

These attributes can also be grouped into categories. Useful attributes are easily found in:

  • The domain name and any sub-domains
  • The URL aside from the domain or sub-domains
  • The domain registration meta-data
  • The content of the homepage, as well as at reported URLs

It requires noting that neither algorithm nor human review is perfect. The goal here is to be able to quickly and easily identify the majority of cases. There will always be exceptions and edge cases. The rest of this blog presumes that there has been a report of a domain engaged in DNS Abuse, most commonly for phishing or malware distribution. 

Frequent Malicious Registration Attributes

A domain that was maliciously registered, is more likely to have some of the below features:

  • Is more likely to be new, or reported shortly after registration,
  • Contains dashes in the domain name,
  • Contains common brand names or misspellings of brand names in the domain or subdomain(s),
  • Contains common service related words like: login, support, or account in the domain or subdomain(s), or
  • Uses subdomains in a reported URL.

An example of a malicious registration can be found here:

In this case, a domain was registered to appear to belong to PayPal and was used to distribute malware. It uses dashes, a common brand name, and was never renewed.

Frequent Compromised Website Attributes

A domain name that was registered for a legitimate or benign purpose, but had it’s website subsequently compromised will likely have some of the below features:

  • Has viewable content related to the domain name,
  • Has been renewed or registered for longer than 1 year,
  • Has a content management system, most commonly wordpress, and often visible in the harmful url, for example: /wp-includes/, wp-content or /wp-admin/,
  • If there is a brand, it is in the URL and not the domain or subdomain, or 
  • If there are common service related words, like login, support, or account, they are in the URL and not the domain or subdomain.

Generally, the above attributes of a compromised website indicate that someone else has control over the website, but not the domain itself. 

An example of a compromised website looks like this:

In this case, a website with wordpress was compromised and used to distribute malware.  There is content on the website that relates to the domain, it was registered in 2019 and subsequently renewed, and you can see a reference to wordpress (/wp-content/) in the URL.

Lastly, none of the above attributes are definitive alone or in isolation. One must consider all of the indicators and balance the information available. 


Mitigating abuse at the DNS level is always an act of balancing harms with the proportionality of the response and the possibility for collateral damage.  For registrars and registries, the only mitigation technique available is preventing a domain name from resolving, which can have serious consequences for a website that is providing legitimate services. The below are the recommended best practices. 

Mitigating Malicious Registrations

It’s worth noting that maliciously registered domain names are usually in violation of registrar registration agreements, terms of service, or acceptable use policies,  providing a basis for domain name suspension.

Where a domain is reasonably determined to have been maliciously registered, a registrar should:

  • Suspend the domain name, 
  • Notify the registrant of the suspension and provide a mechanism to dispute the suspension,
  • Review other domains owned by the registrant or in their account for domains with malicious registration attributes.
  • Notify the web hosting provider.

Mitigating Compromised Websites

Where a domain is reasonably determined to have been compromised, in order to minimize harm to the registrant, website visitors, and risk to themselves, the registrar should:

  • Notify the registrant that the domain is compromised,
  • If possible, provide resources to secure the website, and
  • If possible, notify the web host.

In the circumstance where the registrar determines the compromised website may cause substantial harm, or is unable to contact the registrant,  it should provide the registrant with a reasonable time frame for remediation or response, typically 72 hours, and then suspend the domain. 


For the majority of cases, distinguishing between a malicious domain registration and a website that has been compromised can and should be done.  Doing so will limit the harm of malicious registrations and reduces the impacts on legitimate domains.  

This best practice points towards further work for the DNSAI. First, education and best practices for registrants and website operators on keeping their content management systems clean and up to date, and how to mitigate a harm when it occurs.  As well, the creation or improvement of systems to detect potentially malicious domains prior to registration has the potential to reduce harms and ease the burden on anti-abuse efforts.

Further Examples

Compromised Website Example 1

Source APWG
Submission Date 2021-10-12 
URL Submitted
Registration Date 2017-06-21

Homepage Screenshot:

Harmful URL Screenshot:

Compromised Website Example 2

Submission Date 2021-09-30
URL Submitted
Registration Date 2020-04-03

Homepage Screenshot:

Harmful URL Screenshot:

Malicious Domain Example 1

Source APWG
Submission Date 2021-10-07
URL Submitted hXXp://
Registration Date 2021-10-06

Homepage Screenshot:

Harmful URL Screenshot:

Malicious Domain Example 2

Source APWG
Submission Date 2021-11-01
URL Submitted hXXp://
Registration Date 2021-11-01

Homepage and harmful URL Screenshot: