Evidence, risk and history

What each finding carries: the evidence behind it, a risk level, a confidence score, a history of changes, and a way to set it aside.

Updated

What a finding is

Every lookalike of a domain you watch is a finding. Each one has evidence from DNS and the lookalike’s home page, a risk level, a confidence score and a history of what changed. They are listed in the findings list and opened in the finding view.

Free
Every lookalike finding, evidence, risk, history, dismissing
Pro
All of that, plus WHOIS, screenshots, takedown reports and exports
Business
All of that, plus WHOIS, screenshots, takedown reports and exports
MSP
Same as Pro

Evidence labels

Each registered lookalike is checked in DNS and, when it resolves, its home page is fetched. The finding view shows the result as dated evidence rows; the daily email, alert messages and the welcome page use these short labels:

Not registered
The domain has no nameservers. "Registry wildcard" means the registry answers for names nobody registered, so DNS cannot tell.
Registered
The domain exists but none of the labels below apply.
Mail
It publishes mail servers (MX), so it can receive mail.
Live site
Its home page answered, after up to 3 redirects, and is not a parking page.
Parked
Its nameservers or page belong to a parking or for-sale service.
Login form / Asks for a password
The fetched page has a password field or a login-style title.
Went live
It moved from parked or empty to serving a site or accepting mail, in the last 7 days.
Listed as phishing
The domain, or a name under it, is on a public phishing domain list: CERT Polska's Warning List or PhishDestroy's DestroyList. The evidence row names the list and the listed name, and dates the entry by the list's own date ("Listed on CERT Polska's Warning List since 12 Aug 2026"), or, for a list that gives no dates, by the day Greyphish first downloaded it with the entry ("On PhishDestroy's DestroyList since at least 1 Nov 2026"). An entry that leaves a list and comes back is dated from its return. A name on a hosting platform such as pages.dev, vercel.app or github.io counts as its own domain: a listing there is never one against the platform.
Bad netblock
Its address is on a public list of abusive networks.
Known C2 TLS fingerprint
Its TLS server matches a fingerprint published for malware command servers.
Hosting, Netblock, Certificate
Who runs the network, its address block, and other names on the same certificate.
Checked
When the evidence was last gathered. Findings are re-checked about once a day.

Risk levels

high
It accepts mail, or it is live with a login page, or it is live and was registered in the last 30 days. Also high: a known C2 fingerprint, a bad netblock or a phishing list entry on a domain that is live or has mail, or a page that looks like your site or uses its icon.
medium
It resolves and is not parked. Also medium: a registered domain that would be low but is on a bad netblock or a phishing list.
low
Registered but parked, with no records, or under a registry wildcard.

Risk is separate from the threat level shown in the feeds (critical, high, medium, low), which comes from the kind of lookalike: see kinds of lookalike. Risk is about what the domain is doing now, the level about how it imitates you.

History

Changes found by each re-check are recorded in the finding’s history, on every plan:

Went live
Moved from low risk to medium or high with a site or mail.
Started accepting mail
Mail servers appeared.
A login page appeared
A password field appeared on the page.
A certificate was issued
The first certificate seen for a lookalike that was found before it had one.
Its page looks like your site
A screenshot comparison matched. See screenshots and similarity.
Moved to a different address
Every address changed.
Nameservers changed
The nameserver set changed.
Stopped resolving, Started resolving again
It went down, then came back.

A failed lookup is never recorded as a change. Address and nameserver changes are recorded once a week at most. History is kept for 180 days. Which of these alert is on alerts and channels.

Dismissing findings

On any plan, choose Dismiss on a row of the findings list, or Not a threat or This is mine in the finding view. The finding moves to the Dismissed tab, out of the Needs review counts, and is left out of the daily email, alert channels and exports.

Undo in the notice after Dismiss, or Move back to Needs review in the finding, brings it back. A dismissal is yours alone: other accounts with the same lookalike are not affected.

Clear is not dismiss

Clear, next to the filters, only removes the filters. It does not dismiss anything.

Limits

  • Evidence is what the last check saw. A site can change between checks.
  • The home page is fetched, not every page. A lookalike that only serves a login form at a deeper path may not show a login page.
  • Risk is a summary of the evidence, not a verdict that a domain is malicious.
  • Risk is kept per domain, so two accounts with the same finding see the same risk.
  • Phishing lists cover only part of the phishing on the internet, so most findings are on none. CERT Polska’s list is downloaded every 6 hours and DestroyList’s once a day; an entry that leaves a list stops showing within an hour of the next download, and a list that has not downloaded in 3 days stops showing at all.
  • A listing is the list maintainer’s judgement, shown with their name. Greyphish is not affiliated with CERT Polska, NASK or PhishDestroy. DestroyList is used under its MIT licence, copyright the PhishDestroy contributors.