One Number, One Domain: A Brand-Less Settlement-Mill Smishing Ring
One Number, One Domain: A Brand-Less Settlement-Mill Smishing Ring
Since February 2026, one operator has run an elder-fraud smishing ring that assigns each burner number its own disposable cheap-TLD domain. The lures name no company, law firm, or agency. Each first-name-personalized text mentions a legal settlement, a "senior inflation relief" payment, or a state-treasury asset recovery and includes a bare link to a newly registered throwaway domain. Over roughly 180 days, the operator burned about 1,200 US-mobile numbers and about 1,200 disposable apex domains, one number to one domain, to send tens of thousands of scam texts to a recycled list of senior-skewed first names. The vagueness is deliberate: without a named brand, an elderly target has nothing to call and verify against.
Key Takeaways
- The operator pairs one burner US-mobile number with one freshly registered cheap-TLD apex per blast. Taking down one number or domain reveals none of the remaining fleet.
- Every lure URL is bare-root: a random five-character subdomain on a short coined wordmark, with no path, query string, or tracking token. This removes most substring and path-reuse indicators.
- The domains come mainly from NameSilo: PrivacyGuardian.org-protected exotic-TLD registrations in rolling weekly batches, plus a small reused stash of aged 2025 domains.
- The pretexts deliberately impersonate no brand. This deprives victims of a verification path and removes the trademark owner who might otherwise press for a fast takedown.
- Because we found no shared hosting, backend, or CDN across the pool, attribution depends on registrar cohort, domain-generation grammar, a recycled victim name list, and a shared pretext toolkit.
Background
Smishing aimed at older adults is now one of the costliest consumer-fraud vectors in the United States. The FTC reported $470 million in losses from scams that began with a text message in 2024, roughly five times the 2020 figure. Imposter scams remained the single largest fraud category at about $2.95 billion. The damage is concentrated among seniors: the FBI's Internet Crime Complaint Center recorded roughly $4.9 billion in losses from victims aged 60 and older across about 147,000 complaints in 2024, and around 7,500 of those victims lost more than $100,000 each. Government-imposter, treasury, asset-recovery, and settlement or refund pretexts delivered by SMS are central to that trend.
This operator applies that pretext family at scale using cheap, disposable infrastructure. New generic top-level domains such as .rest, .click, .autos, .world, and .top are among the hundreds of ICANN name spaces beyond legacy .com. Many cost under a few dollars a year, and several registries and resellers have historically responded to abuse more slowly than .com. A domain can therefore survive a short smishing burst before anyone acts. Independent measurement supports this assessment: Interisle's Phishing Landscape 2024 found that 42 percent of all domains reported for phishing were registered in new gTLDs, up from 25 percent the prior year. Among the 35 gTLDs with the highest phishing rates, 33 cost under $5. The same study reported a 51 percent rise in phishing at subdomain providers, matching this operator's use of random subdomain labels to multiply hostnames from each apex.
A discount registrar and its in-house privacy proxy support the rest of the model. NameSilo is a high-volume discount registrar that supports self-service, API-driven bulk registration, and its WHOIS privacy service, PrivacyGuardian.org, replaces the registrant's real identity in public records. This setup allows an operator to create hundreds of anonymous domains and replace them as soon as the old ones burn.
Discovery and Infrastructure
Analysts initially mapped the ring from 14 seed pairs. Each pair consisted of a US-mobile number tied to one cheap-TLD apex, and all were first seen during a two-day burst in late April 2026. We expanded from that infrastructure pattern: a number matching +1 followed by a US-mobile pattern and sending a bare-root link to a four-to-eight-character wordmark on a cheap TLD. That process surfaced about 1,200 sibling numbers and about 1,200 sibling domains across the 180-day window.
One structural rule defines the operation: each burner number sends to a single apex domain. We found no shared hosting provider, resolved backend, or CDN overlap across the pool, which points to deliberate compartmentalization. Removing one number or domain ends only one victim thread and provides no pivot into the rest of the fleet because there is no common host or reused domain to cluster on. The operator accepts the cost of roughly 1,200 numbers and 1,200 domains in exchange for takedown resilience.
Most domains follow one naming grammar: consonant-heavy coined pseudo-words of four to eight characters on exotic TLDs, such as orbora[.]rest, gigosi[.]rest, coobey[.]click, biogix[.]autos, and souragu[.]pro.
Across the 180-day sibling footprint, .rest and .click account for most of the cheap-TLD distribution:
| TLD | Domains | Senders |
|---|---|---|
| .rest | 510 | 531 |
| .click | 341 | 356 |
| .autos | 76 | 77 |
| .link | 55 | 60 |
| 37 | 40 | |
| .top | 32 | 32 |
| .pro | 30 | 31 |
| .world | 26 | 27 |
| .biz | 11 | 11 |
| .xyz / .live / .icu | 5 combined | 5 combined |
How It Works
A target on the operator's list receives a short text that begins with their first name and a piece of officialese. The message might refer to a case that "remains active," a "path sync" tied to a ten-digit reference, a "Senior Inflation relief" payment awaiting confirmation, or an asset that the state treasury is about to reclaim. It ends with a bare link to a random subdomain on a throwaway apex and often includes a countdown, such as "You have 7 minutes to verify your status."
The text never names a real institution. That omission is the operation's main evasion mechanism. A recipient who tries to verify the message has no company to call because none is named, while the linked domain is an unfamiliar coined word. The page sits on a domain registered only days earlier, and the URL has no path or parameter that a filter could have encountered before. If the recipient does not respond, another message arrives the next day from a fresh number and domain, using a slightly different pretext from the same toolkit.
The personalization follows a pattern. First names repeat across otherwise unrelated senders, as do fixed first-name and surname pairings. This points to a purchased or leaked senior-demographic lead list containing roughly 20 to 30 recurring entries rather than a per-recipient generator. Ten-digit reference numbers embedded in the messages also recur across independent burners, consistent with victim-phone fragments carried on the same list.
Sample Lures
The samples below are attacker-produced SMS bodies. We removed every recipient identifier: placeholders replace first names, embedded reference numbers, and subdomain labels, and all domains remain defanged. Only the attacker's pretext structure remains.
SMS, case-active and legal-threat pretext:
[first name] | Still no reply on the [reference #] path fix. If you stay silent to...
hxxps://[label].spelum[.]rest
SMS, senior inflation relief pretext:
Hi [first name], pardon this interruption today, however, your Senior Inflati...
hxxps://[label].eleganix[.]email
SMS, priority-access and digital-twin pretext:
[first name], your priority access registered under [name] is pending confirm...
hxxps://[label].souragu[.]pro
SMS, transfer and last-phase pretext:
[first name] - We are in the last phase of the process. Your input is essent...
hxxps://[label].plihod[.]rest
SMS, religious-providence pretext:
God opened this for you. This wasn't random...
Technical Analysis
No individual domain or number provides a durable fingerprint because both rotate daily. The production machinery does. Four signals separate this traffic from legitimate messaging, and each persists through the operator's daily infrastructure churn.
Domain-Generation Grammar
Most apexes are consonant-heavy coined wordmarks that are pronounceable but meaningless, using CVCVCV or CVCCVC shapes of four to eight characters: orbora, gigosi, plihod, coobey, souragu, muniri, kurare, croshe. A smaller share are short vowelless strings or simple word pairs, mostly in the aged stash (uhtnk, vrmjh, logicpic). None carries a brand token, so no apex points a recipient to a company they could look up.
Registration Cohorts
Public WHOIS data ties the pool to a small group of registrar and privacy-proxy factories. Creation dates cluster in rolling 2026 batches, with a thin tail of aged assets reused from 2025.
| Registrar | Privacy proxy | Cohort | Examples (defanged) | Created |
|---|---|---|---|---|
| NameSilo | PrivacyGuardian.org | Fresh exotic-TLD batches | orbora[.]rest, gigosi[.]rest, coobey[.]click, biogix[.]autos |
Feb to Apr 2026 |
| NameSilo | none | Aged stash, reused | uhtnk[.]biz, vrmjh[.]biz, logicpic[.]rest |
Apr to Dec 2025 |
| registrar.eu | WHOIS Privacy Protection | Marginal | azotipa[.]top |
Dec 2025 |
The newest registrations precede observed activity by only days. The exotic-TLD apex biogix[.]autos was created April 14, consistent with burst activation within a day or two of registration. This registrar cohort is the highest-specificity connection across the operation, and the underlying data is public. A fresh-registration stream from this factory, filtered for cheap-TLD targets matching the wordmark shape, could flag siblings before the first text goes out.
Subdomain Grammar
Every CTA host adds a random lowercase five-character, consonant-heavy label before the apex. Each label remains confined to one apex, with no token reuse across domains. When a label repeats, it does so only under the same domain. The hosts are bare-root and contain no path, query string, or tracking token.
Representative hosts (defanged):
gjqvt[.]spelum[.]rest
atiwi[.]orbora[.]rest
wlueo[.]biogix[.]autos
gbuui[.]eleganix[.]email
lbyce[.]coobey[.]click
vpyfh[.]gigosi[.]rest
nobsr[.]souragu[.]pro
sbdnl[.]animum[.]click
What Ties the Clusters Together
Although we found no shared DNS or backend infrastructure across the registrar cohorts, five signals connect their naming and registration patterns to one operator: the registrar-plus-privacy-proxy factory registered in tight batches; the coined-wordmark-on-cheap-TLD grammar; the reused senior-skewed first-name and surname pairings; ten-digit reference fragments repeated across independent burners; and a common pretext toolkit containing eleven template families. Explicit email and Facebook sweeps for the seed apexes found nothing, which limits the observed operation cleanly to SMS.
The same pretext toolkit appears across the registrar cohorts:
| Template family | Representative fragment |
|---|---|
| Case-active / path-sync | "this case remains active" / "path sync for [reference]" |
| Priority-access / digital-twin | "priority access registered under [name]" |
| Senior inflation relief | "Senior Inflation relief" / "benefit adjustment" |
| Transfer / last phase | "the transfer was made to" / "last phase of the process" |
| Check returned / state deposit | "check made out to [name] was returned" |
| Religious providence | "God opened this for you" / "Grace moves in quiet ways" |
| State-treasury reclaim | "the state treasury is stripping" / "recovery of your assets" |
| Case closure / abandonment | "CASE ABANDONMENT" / "CASE CLOSURE" |
| Urgency / balance | "You have 7 minutes" / "your balance remains unaffected" |
| Formal notice | "Dear Mr or Mrs [name]" |
| Access code | "Enter [4 digits] to access" |
Detection Observations
This traffic differs from legitimate messaging through its overall shape rather than one repeated artifact. A US-mobile number sends a bare-root link whose hostname combines a random five-character label, a four-to-eight-character coined word, and a cheap gTLD, with no path or query string. Almost no legitimate sender uses that combination. Legitimate branded links include a real brand token, usually a path, and a stable registrable domain. This operator uses none of them.
The registrar cohort provides the strongest early pivot. Cheap-TLD apexes that match the wordmark shape and are freshly registered in weekly batches through the NameSilo-PrivacyGuardian.org factory form a high-specificity cluster visible in public WHOIS before a domain appears in a message. The pretext toolkit provides a second, content-side pivot. Phrases such as "path sync," "Senior Inflation relief," and "priority access registered under" recur across otherwise unrelated senders while the infrastructure changes.
Shape-based rules for this pattern require one caution. Legitimate brands also use short labels on cheap TLDs for branded shortlinks, so a rule based on TLD and label shape needs an allow-list of known-legitimate branded-shortlink services before it acts.
Indicators of Compromise
All indicators below are defanged and come from the verified-malicious set for this campaign. We removed recipient data. Each type includes a representative subset.
Sender Numbers
| Value | Role | Notes |
|---|---|---|
+1-210-347-3914 |
Sender | Burner, coined-wordmark apex |
+1-423-208-6898 |
Sender | Burner |
+1-713-542-2490 |
Sender | Burner, shared apex orbora[.]rest |
+1-346-406-6541 |
Sender | Burner |
+1-561-546-1915 |
Sender | Burner, senior-relief pretext |
+1-352-727-0173 |
Sender | Burner |
+1-917-245-8930 |
Sender | Burner |
+1-279-253-6314 |
Sender | Burner |
+1-312-350-5197 |
Sender | Burner, priority-access pretext |
+1-956-820-2389 |
Sender | Burner |
+1-405-435-9531 |
Sender | Burner |
+1-850-741-8236 |
Sender | Burner |
| ... (representative subset; 50+ verified-malicious sender numbers) |
Domains
| Value | Role | Notes |
|---|---|---|
orbora[.]rest |
CTA apex | NameSilo, coined wordmark |
gigosi[.]rest |
CTA apex | NameSilo, path-sync pretext |
plihod[.]rest |
CTA apex | NameSilo, last-phase pretext |
coobey[.]click |
CTA apex | NameSilo |
animum[.]click |
CTA apex | Coined wordmark |
biogix[.]autos |
CTA apex | Registered days before use |
eleganix[.]email |
CTA apex | Senior-relief pretext |
souragu[.]pro |
CTA apex | Priority-access pretext |
reemit[.]world |
CTA apex | Coined wordmark |
| ... (representative subset; 100+ verified-malicious domains) |
Hosts
| Value | Role | Notes |
|---|---|---|
gjqvt[.]spelum[.]rest |
CTA host | Random 5-char label |
atiwi[.]orbora[.]rest |
CTA host | Random 5-char label |
wlueo[.]biogix[.]autos |
CTA host | Random 5-char label |
ccumm[.]sippihri[.]email |
CTA host | Random 5-char label |
gbuui[.]eleganix[.]email |
CTA host | Random 5-char label |
lbyce[.]coobey[.]click |
CTA host | Random 5-char label |
wtyds[.]plihod[.]rest |
CTA host | Random 5-char label |
vpyfh[.]gigosi[.]rest |
CTA host | Random 5-char label |
nobsr[.]souragu[.]pro |
CTA host | Random 5-char label |
dxfcm[.]vrmjh[.]biz |
CTA host | Aged-stash apex |
sbdnl[.]animum[.]click |
CTA host | Random 5-char label |
plbfi[.]reemit[.]world |
CTA host | Random 5-char label |
| ... (representative subset; 25+ verified-malicious hosts) |
MITRE Fight Fraud Framework Mapping
The mapping below follows the MITRE Center for Threat-Informed Defense Fight Fraud Framework (F3). F3 reuses ATT&CK technique IDs for existing behaviors and assigns fraud-specific IDs to the rest. Two consumer-side social-engineering stages used by this campaign, trust-building through feigned authority and manufactured urgency, lack dedicated technique IDs in the public F3 matrix. We therefore describe them behaviorally rather than map them to invented IDs.
| Tactic | Technique | ID |
|---|---|---|
| Reconnaissance | Gather Customer Information (recycled senior lead list) | F1029 |
| Resource Development | Acquire Infrastructure: Domains (bulk cheap-TLD apexes) | T1583.001 |
| Resource Development | Establish Accounts (burner sending numbers) | T1585 |
| Resource Development | Create Fake Materials: Fake Website | F1020.002 |
| Initial Access | Phishing, smishing variant (unsolicited SMS lure) | T1660 |
| Initial Access | Impersonate Official (state-treasury / legal authority) | F1032 |
| Monetization | Electronic Funds Transfer | F1025 |
Conclusion
The operator relies on disposability rather than sophistication. Each blast uses one number, one domain, and one pretext, mostly registered anonymously days before use and then abandoned. That leaves defenders with almost no reusable artifact to pursue. The production machinery is harder to replace: one registrar pipeline, a wordmark generator with a recognizable grammar, and a recycled victim list that keeps the same names circulating across thousands of burners. Those signals persist even when the next batch of numbers and domains appears new.
Related research
Midterm Texts Promise Fake Checks and Open Recurring Donation Pages
Ahead of the November 3 US midterms, political texts claiming fake checks and voter-record errors led to donation forms built around repeat gifts.
How a Government-Impersonation Smishing Kit Signs Its Own Registration Batches
A smishing kit impersonating toll agencies, tax offices and national police stamped the registration date into its own WHOIS records.
The Supplement SMS Kit That Borrows Real Business Identities
Since April 2026, one operation has run 26 supplement storefront domains by text message, most carrying the legal identity of a real registered business.