Phishing hops through six Google domains to beat filters
KnowBe4 Threat Lab found a campaign chaining Google Meet, Search, DoubleClick, Tag Manager and Analytics redirects, hiding targets in the URL fragment your logs never see.
Domain reputation is the control most email gateways lean on hardest, and this phishing campaign never leaves a domain your gateway trusts until the last hop. The KnowBe4 Threat Lab documented operators chaining open redirects across Google's own infrastructure to move a victim from an inbox to a credential harvester — or to a working remote-access install.
What actually happened
The chain uses redirect endpoints on meet.google.com/linkredirect, www.google.com/url?q=, adservice.google.com DoubleClick click tracking, cse.google.com/url?q=, regional images.google.com variants, googletagmanager.com/debug/clearcookies?url=, and Analytics with a dl= parameter. A single lure can move through three Google hops before it reaches attacker infrastructure. Every one of those is a domain your proxy allowlist, your URL scanner and your firewall already trust.
The targeting trick is better than the redirect trick. The victim's email address is encoded in the URL fragment — everything after the #. Browsers strip fragments before sending the request, so the pre-targeted address never appears in server logs or in scanner telemetry. The link looks generic to your tooling and arrives pre-filled to the victim.
Lures impersonate DocuSign and SafeSend document reviews, Microsoft 365 credential-expiry notices, FedEx tracking, QuickBooks and OneDrive payment alerts, Social Security benefit notices, and Microsoft voicemail. Targets span manufacturing, government, finance and non-profits. Two payload tracks follow: a high-fidelity Microsoft sign-in page (or device-code interception), or a fake identity-verification prompt that installs ScreenConnect — legitimate remote monitoring software, used here as a persistent backdoor.
Why it matters for your business
This is the second ScreenConnect story this week, after the rogue-client worm ConnectWise is still patching. Different entry point, same endgame: a signed, legitimate RMM binary is a cheaper backdoor than malware, and it survives your EDR because it is supposed to be there.
The practical shift is that "the link goes to google.com" is now worthless as a safety signal, for your filters and for your staff. Three things are worth doing this week. Hunt your endpoints for ScreenConnect installs nobody authorized — an inventory query, not a project. Block the specific Google redirect endpoints KnowBe4 lists at the proxy; almost no business workflow needs googletagmanager.com/debug/clearcookies. And check whether your gateway logs the full URL including fragments, because if it does not, your incident timeline will be missing who was targeted.
Key takeaways
- Redirect chains run through Google Meet, Search, DoubleClick, Custom Search, Image Search, Tag Manager and Analytics before touching attacker infrastructure
- Victim addresses hide in the URL fragment, which browsers strip — so pre-targeting is invisible to server logs and scanners
- Payloads split between a Microsoft sign-in clone and a fake verification prompt that installs ScreenConnect for persistent access
- KnowBe4 advises blocking the listed redirect endpoints, hunting Telegram Bot API exfil traffic, and searching for unauthorized ScreenConnect installs
- Trusted-domain allowlisting is not a phishing control when the trusted domain is the delivery mechanism
Your filter trusts Google. The attacker read the same documentation you did. We audit what your email and proxy layers actually block, inventory the remote-access tools running on your machines, and write the detections your vendor's defaults leave out. Book a stack audit or see how we work.
- #phishing
- #email-security
- #screenconnect
- #credentials
Tommy Rush — Founder, Rush Commerce
Operator turned builder. 15+ years running operations — now shipping the systems businesses run on. More
Get The Rush Report weekly — one email, zero fluff.
Keep reading
VMware Workstation VM escape: get to 26H1u1
CVE-2026-59346 is a CVSS 9.3 integer overflow in the VMXNET3 adapter that lets a VM admin run code on the host. No workaround. Your sandbox is not a boundary.
Read itSonicWall SMA1000: CVSS 10 SSRF chained to pre-auth RCE
Two SonicWall SMA1000 zero-days are under active attack. A CVSS 10.0 pre-auth SSRF chains with command injection for unauthenticated RCE on your VPN box.
Read it