|
||
|
||
Consider a phishing report that arrives with a timestamped screenshot. The image shows a page asking a customer to sign in to a bank. An analyst opens the reported URL twenty minutes later and gets a 404. What should the registrar do with those two observations?
Sending the researcher a form reply saying “unable to reproduce” leaves the first observation unexplained. Suspending the entire domain on the strength of the screenshot leaves a different question unanswered. The page could sit on a legitimate site whose owner can remove it, or the domain itself could have been registered to steal credentials. The decision I want an abuse team to be able to explain is what it knew at the time, what harm was still occurring, and why the action it chose was likely to stop it.
At NiceNIC, our August 2026 report shows how easily the workload can be misread. Recorded complaints rose from 23,316 in July to 30,450 in August. Over the same period, the number of domains involved edged down from 9,644 to 9,539. Several reports can concern one domain. A later report may repeat the first one, or it may contain the only image that captured what a victim actually saw. A reviewer has to read it to know which.
September and October offer a different comparison. Looking only at generic top-level domains in both months, complaints fell from 26,608 to 25,551, while the count of domains recorded as paused rose from 5,012 to 6,129. Our October 2026 abuse report also gives a broader count. Including country-code domains, 6,433 domains were recorded as paused. The report groups complaints by creation date and shows the status of the associated domains at the time of the snapshot. Those figures describe the work in the queue. Judging a particular suspension requires the evidence in the case file.
Returning to the bank example, I would first want to know whether the screenshot identifies the exact URL and when the researcher observed it. The bank can help establish whether the sign-in page it was shown was genuine. An independent observation of the same URL can tell us whether other users were exposed. Someone who has already supplied those details deserves a reply that identifies the specific question still open.
A failed visit matters, too. The page may already have been removed. What visitors saw may have depended on their location, their device, or how they reached the page. The time and scope of each observation give the reviewer a way to weigh both accounts. Researchers can send the affected URL, the time and time zone, and any capture they can safely share through our public abuse form. Sensitive details such as victim credentials should be redacted.
For a domain we sponsor under a generic top-level domain, Section 3.18 of the Registrar Accreditation Agreement requires prompt investigation and, once we have actionable evidence of DNS abuse, appropriate action to stop or disrupt it. A registrar-level hold can interrupt the website and email that depend on the domain. If a hosting provider can promptly remove a phishing page from a compromised business site, that can protect visitors while keeping the business reachable. Where the evidence shows that a domain is actively being used to steal credentials and a hold is needed to disrupt the harm, waiting for every analyst to see the same screen can leave users exposed.
Our abuse handling manual calls for immediate first review of an emergency involving active harm. It sets an internal target of mitigation normally within 24 hours when the evidence is actionable. The same urgency applies to a well-documented report from an individual as to one from an established security provider.
After ICANN87 in Bali, NiceNIC signed a memorandum of understanding with the Trusted Notifier Network on November 16. TNN announced that NiceNIC was its 17th partner. Under the model, an accredited notifier verifies a case with the victim and stands behind the notice. NiceNIC retains the decision on what action to take. Work to connect the channel to our case-handling system is underway. Our July–October figures predate the agreement.
ChainPatrol’s account of its work with NiceNIC illustrates another contribution. A researcher can preserve the impersonating page, the URL and the time it was observed, giving the next person to review the case something concrete to examine even after the page changes. A reseller contacting the site owner needs to know which URL was reported, what a visitor saw and what still needs attention.
The question I would put to fellow registrars is what the record shows when a case closes. Can a second reviewer see what the reporter observed, what the registrar could verify, and why the chosen action made sense? If the domain was later restored, can that reviewer tell whether the operator fixed the problem or new evidence changed the assessment? Those are questions worth publishing alongside suspension counts. They are also what I would want answered if it were my company’s domain, or my bank’s customers, on the other side of the report
Sponsored byDNIB.com
Sponsored byWhoisXML API
Sponsored byIPv4.Global
Sponsored byVerisign
Sponsored byCSC
Sponsored byVerisign
Sponsored byRadix