Device Code Phishing: Hornetsecurity Sees a Return of Microsoft 365 Attacks Where Victims Pass MFA on a Real Microsoft Page, and What to Tell Staff
Hornetsecurity saw device-code phishing against Microsoft 365 accounts return in November 2026. Victims sign in on a genuine Microsoft page and pass MFA themselves, but the sign-in belongs to the attacker.
Where Worldwide
On this page 9 sections
Hornetsecurity says its teams saw device-code phishing aimed at Microsoft 365 accounts come back in November 2026. Any business on Microsoft 365 should pay attention: the person being tricked signs in on a real Microsoft page and passes any MFA check themselves, so having MFA turned on will not stop it.
This post covers what the researchers confirmed, what is still unknown and what to tell staff this week.
What is device code phishing?
Hornetsecurity explains that in a phishing scenario, the client asking for the code is controlled by the attacker, not the victim. The victim is talked into entering the code, so the sign-in they finish belongs to the attacker's client.
Hornetsecurity's Threat Intelligence Lab states the hard part plainly: "The recipient completes the sign-in on a genuine Microsoft page, including any MFA requested by the service."
That one sentence explains why this attack keeps coming back. Nothing on the screen is fake. The page is Microsoft's, and the MFA prompt is real. The only thing wrong is who asked for the sign-in.
Confirmed vs claimed: what Hornetsecurity actually reported
| Point | Status |
|---|---|
| Device-code phishing against Microsoft 365 accounts came back in November 2026 | Confirmed: Hornetsecurity's own teams observed it |
| The victim completes the sign-in on a genuine Microsoft page, including any MFA | Stated by Hornetsecurity |
| In a phishing scenario, the client requesting the device code is controlled by the attacker | Stated by Hornetsecurity |
| How many accounts or companies were targeted or affected | Not stated in the sources we verified |
| The exact lures, emails or pages used | Not confirmed in the sources we verified, so we leave them out |
What is not known yet
The sources we verified do not say:
- How many Microsoft 365 accounts were targeted, or how many attacks worked.
- Which industries, company sizes or countries were hit.
- What attackers did after a sign-in succeeded.
- Whether Microsoft has changed anything in response.
If you see figures or named victims elsewhere, treat them as unconfirmed until a primary source backs them up.
Does MFA stop device code phishing?
Not on its own. Many people search for this as an "MFA bypass" for Microsoft. Going by Hornetsecurity's description, MFA is not actually broken. The victim completes the MFA step themselves on Microsoft's own page. MFA proves the right person is there. It does not prove that person started the sign-in they are approving.
Keep MFA switched on. It still protects you against plenty of other attacks. Just don't assume it covers this one. The defence that works here is a clear habit for staff, backed up by your IT team.
Why device code phishing matters for a business on Microsoft 365
For many businesses, a Microsoft 365 account is the front door to company email and documents. Two things make this attack harder to catch:
- The page looks right because it is right. Training that teaches people to spot fake sign-in pages does not cover a real Microsoft page.
- The MFA prompt looks normal. Staff who are used to approving MFA prompts will not notice anything odd.
The practical fix is cheap, but you have to tell people about it.
What to do this week: device code phishing checklist
| Action | Who | How to check it is done |
|---|---|---|
| Tell every staff member Hornetsecurity's rule: only enter a device code for a sign-in or device setup you started yourself and understand | Owner or manager | Send a short note or say it in a team meeting, and ask people to confirm they read it |
| Make it clear that a request to type a code into a Microsoft page, from an email, chat or caller, should be reported, not followed | Everyone | Name one person or inbox that takes these reports |
| Ask whoever runs your Microsoft 365 to list the people and devices that really need device code sign-in in your company | IT lead or provider | A written list, so anything outside it stands out |
| Decide in advance what happens if someone entered a code they did not start: treat it as a possible account compromise and tell IT straight away | IT lead | A short written plan that staff know about |
| Read Hornetsecurity's write-up and the related Microsoft post in full | IT lead | Notes on anything that applies to your setup |
The sources we verified give no fixed version or patch to install. This is a phishing method, not a software flaw with an update.
Device codes and AI coding agents
Microsoft's security blog has a related post on the risks of device code authentication for AI coding agents. We have not summarised it here and make no claims about what it found. If your developers use AI coding tools that sign in with a device code, read it alongside Hornetsecurity's write-up. Our post AI Can Build Your App. Who Is Going to Keep It Running? covers the wider question of who looks after AI-built software.
Device code phishing questions people are asking
What is device code phishing in Microsoft 365?
An attacker's client asks for a device code, and the attacker gets a victim to enter it on a real Microsoft page. The sign-in the victim finishes belongs to the attacker's client.
Can device code phishing get past MFA?
MFA is not broken, but it does not stop this attack. Hornetsecurity says the victim completes the sign-in, including any MFA the service asks for, on a genuine Microsoft page.
How do I know if a device code request is real?
Ask yourself one question: did I start this sign-in or device setup myself, and do I understand what it is for? If not, don't enter the code. Report it.
Has Microsoft 365 been hacked?
The sources we verified describe phishing that uses a genuine Microsoft sign-in page. Nothing in them says Microsoft's own systems were breached.
What should I do if I entered a device code I did not start?
Tell whoever runs your IT straight away, and treat the account as possibly compromised until they have checked it.
How many companies were affected?
Nobody knows from the sources we verified. Hornetsecurity reports a resurgence of the technique but gives no count.
The bigger lesson
The hardest attacks to spot don't fake anything on screen. They use the real page, the real prompt and a real person, and all they change is who started the request. Security that relies on spotting a fake will miss them. What catches them is a simple rule that everyone knows, plus someone who knows what normal looks like on your accounts.
Want help tightening how your company signs in? See security hardening or request a quote.
Get security alerts by email, plus the free checklist, at /security-alerts.
Sources: Hornetsecurity: Microsoft 365 device-code phishing; Microsoft Security Blog: the risks of device code authentication for AI coding agents.
Have a question about this article?
Ask it here and we will answer you directly by email. Questions are private, never posted.
Security alerts by email
Hear about the next one first.
When attackers start using a flaw in software businesses run on, we email what it is, whether it touches you and what to do that week, with a round-up of our latest security writing. Free, one-click unsubscribe. You get the checklist straight away: 12 things to secure this week.
Want a result like this?
Security alerts, breaches and the technology a business runs on: what happened, who it affects and what to do.
Get a Tech QuoteKeep reading
VMware vCenter CVE-2026-123890
A critical VMware vCenter flaw, CVE-2026-123890 (CVSSv3 9.8), lets attackers open a vCenter administrator sess...
ARTEX AI Hacking Framework: Storm-0940 Tried Sign-Ins on More Than 2,250 Tenants and Reached Privileged Access in Under 3 Hours, What Is Confirmed and What to Check
Microsoft tied a campaign to Storm-0940, which tried sign-ins on more than 2,250 tenants and in some cases rea...
AI Agents and Employee Credentials: Nudge Security Says Many Agents Inherit Their Builder's Access, What Is Confirmed and What to Check This Week
Nudge Security research says many AI agents run on the credentials of the employee who built them, so they get...