Skip to content

Build Not yet confirmed elsewhere1 publisher 2 min read Published

Yahoo Mail ad chains hid a fake Microsoft alert from data-center and phone visits

Two tech-support scams ran ads in AT&T's Yahoo Mail on Oct 7 and 8 that showed a fake Microsoft alert only to Windows desktops on home broadband. In scans by Tuxxin's Daniel Jones, visits from a data center and a phone got an ordinary dating or finance article.

The Engineer · Build desk

How we use AI Send a correction

First scam kit seen on 536 cloud hostnames Public scans of the first operation's scam kit from July 12 to October 10, 2026, and the hostnames a second review by Kirk of ohey.ca added.

The first operation's scam kit appears in 747 public scans of 536 cloud hostnames from July 12 to October 10, 2026, about 1.4 public scans per hostname. A second review by Kirk of ohey.ca added 31 hostnames to the kit's footprint.

First scam kit seen on 536 cloud hostnames
FigureValueNoteClaim
Public scans showing the first operation's scam kit747 scans21
Added in Kirk's review31 hostnames22
Scans per hostname, about1.4 public scans25

What happened

  • The first ad sat in the inbox list like an email, with sender line CHECKOUTERS, subject (1) New Message and a small Ad label.
  • Its click passed through Taboola's tracker under publisher ID yahoo-mail-att, the AT&T edition that serves att.net, sbcglobal.net and bellsouth.net addresses.
  • The first ad led to checkouters[.]app, a dating article site registered September 23, where a full-screen Azure Container Apps frame covered the page for home-broadband Windows visitors.
  • The second ad ran through a Branch deep link to a filtering server at maizerubyen[.]site, which routed the data-center visit to a personal finance article on another site.
  • The first operation's kit appears in 747 public scans of 536 cloud hostnames between July 12 and October 10, 2026.

Why it matters

  • constraint Fetching a landing page from a report will not confirm the scam: eight direct opens of the first ad's page, with and without tracking codes, all returned the dating article.
  • exposure On the first chain a user who checks the address bar sees the dating site, so nothing in the browser chrome points to Microsoft or to the Azure host serving the frame.
  • decision Verifying a chain like this means replaying the fresh click from several networks at once and comparing what each receives. That is the method the write-up used.

On the second chain the visitor's location is checked twice. The hop at maizerubyen[.]site makes the first cut [13]. The Azure static-hosting page that the home-broadband visit reached then looked up each visitor's IP address with two separate services and opened a live chat window from Tawk.to [15]. It showed the alert and the number +1 844-959-2491 [14].

All of this comes from one post, by Daniel Jones of Tuxxin LLC, who notes it was drafted with AI assistance [24]. The first chain is the less tidy half. Its frame was intermittent: one Comcast visit received the frame, and a repeat visit received only the plain article [2]. The first ad's link was never tested from a data center, so the data-center result rests on the second chain alone [2] [13]. The post reports what visitors received. It does not report what Taboola's or Yahoo's ad review fetched, so whether review missed these creatives is not shown.

The slot is Taboola inventory. Under a 30-year deal, Taboola sells the native advertising on Yahoo's properties [8]. The tracking link on the first ad shows 0.25 as the cost per click and names no currency [9].

The filtering domain's timeline is short. maizerubyen[.]site was registered on September 17 and went three weeks without a TLS certificate [18], which is 21 days to the start of the first certificate's validity [27]. That certificate was valid from 12:44 UTC on October 8 and reached the CT logs at 13:43 UTC [19]. The ad click the author analyzed came at 15:51 UTC, 2 hours 8 minutes after the log entry [19] [26]. issued.live's reverse IP lookup found no other domain on the server at 64.177.41[.]125 [20].

The kit's footprint extends past the 536 hostnames. Kirk, of ohey.ca, reviewed it separately and found 31 more hostnames and three more Azure environments [22]. At about 1.4 scans per hostname [25], we would not key a detection on hostnames.

The chat account is a better bet. It was created in June 2024 [16], and the post places it, with the same lookup key, on a second Azure page on September 24, reached through a Taboola ad and a Branch link at walletsaver.app[.]link [17]. The second chain also loaded a tracking script from userstatics[.]com, a domain registered in 2020 and first recorded on its current server on August 13, 2026 [23].

What to watch

  • Whether Taboola or Yahoo say what their ad review fetched for these two creatives, or act on the yahoo-mail-att inventory.
  • Whether the chat account and lookup key, or the userstatics[.]com script, turn up in scans after October 10, where the archive search stops.
  • What comes of the reports the authors filed with each provider, which the post says are listed on tuxxin.com.

Clarity's read

What the record supports and how the coverage leans. The claims behind it follow.

Reality

Evidence58
Adoption
Insufficient
Hype gap0
Incentives35
Confidence55
Why these scores

Claim ledger

Ranked by verification strength, evidence, and original report placement.

  1. [1]

    The scam appeared only on visits that arrived through a fresh ad click; the authors opened the first ad's landing page directly eight times, with and without its tracking codes, and every visit returned the dating article.

  2. [2]

    The first ad's frame was intermittent: Comcast visits received the frame once and the plain article on the next try. No data center visit ran on the first ad's link.

  3. [3]

    Two scam operations ran ads in the same sponsored slot of AT&T's Yahoo Mail on October 7 and 8, 2026.

    Reported Supported Source: Daniel Jones, Tuxxin LLC, dev.toView cited source

Get this ledger inside ChatGPT. Ask about any of our stories and the answer brings its figure, each claim’s verdict and the sources. Free, no account.

Add to ChatGPTAdd to Claude

Sources

1 independent publisher whose own reporting we read for this story.

  1. dev.to

    1 article · October 12, 2026

    Yahoo Mail malvertising: two cloaked tech support scam chains

Share your take

Let Clarity write the post for you.

Signed-in readers get a short post drafted on this story in the register they choose — narrative, analytical, or a direct position — editable to the last word before it goes anywhere. The share buttons at the top of this story work without an account.

Topics and entities

Follow any of these and your For You feed starts watching them — no settings page required.

Topics

Entities

Loading related stories