Methodology
How the DomainGuard scan works
Every DomainGuard exposure report is built from public, deterministic data — Certificate Transparency logs and live DNS answers, read at scan time. There is no black-box classifier and no guessed finding: if it is in your report, the scan saw the record itself. See it applied to a real domain in our sample report for getingeniousai.com — our own domain, scanned by the same code path as yours.
What the scan checks
1. Certificate Transparency logs
Every publicly trusted certificate is published to Certificate Transparency (CT) logs. The scan queries crt.sh for every certificate issued for your domain and its subdomains, and lists the issuer, the names covered and the validity window. A certificate you do not recognise can be the first visible sign of mis-issuance or impersonation — which is why the fix list starts with reviewing the issuers.
2. CAA posture
The scan reads your domain's CAA DNS records live and reports the three flags that matter: issue (which certificate authorities may issue for you), issuewild (the same, for wildcard certificates) and iodef (where mis-issuance reports should go). No CAA record at all means any CA on earth may issue a trusted certificate for your domain — the single highest-weighted finding in the score.
3. Lookalike (typosquat) sweep
The scan generates lookalike permutations of your domain with its own dnstwist-style engine — omission, transposition, keyboard-adjacent replacement and insertion, repetition, hyphenation, vowel swaps, homoglyphs (look-alike Unicode characters) and bitsquatting — plus trust-word variants (login, verify, account, support…) and common alternate TLDs. Each permutation is then checked against live DNS. A permutation counts as registered only when DNS actually answers for it (an A, AAAA, NS or MX record); its MX records are shown, because a lookalike that can receive mail can also send it as that name.
How findings are verified
- Live evidence only. A finding exists because a CT entry or a DNS answer existed at scan time. The report prints the underlying data — certificate issuers and names, CAA records, registered lookalikes with their MX — so you or your IT provider can re-check every line independently.
- Honest sweep size. The sweep is time-boxed so a scan never hangs. The report always states how many permutations were actually checked out of how many were generated, and says plainly when a sweep was cut short.
- A reproducible score. The exposure score starts at 100. No CAA issue record costs 25 points; each registered lookalike costs 5 points, capped at 30. Same inputs, same score — re-scan after a fix and the score moves only because the records did.
- Nothing inferred about you. The scan reads public records about a domain name. It never touches your servers, your inbox or anything behind a login.
Data sources
- Certificate Transparency logs, via crt.sh
- Live DNS (CAA, A, AAAA, NS and MX records), resolved at scan time
- DomainGuard's own permutation engine — no third-party typosquat feed
What this scan does NOT check
- the content actually hosted on the lookalike domains (phishing-kit scanning is a Pro check)
- your own DMARC / SPF / DKIM enforcement
- anything behind a login, your inbox security, or malware already inside your site
Those gaps are stated on every report too — a report that hides its blind spots is not evidence. Continuous monitoring (hourly CT checks, daily lookalike sweeps, rogue-MX detection and takedown packets) is what DomainGuard Pro adds on top of this one-time snapshot.