It starts with an ordinary email. A message file lands in your inbox, you double-click it, and Outlook does what it always does — it opens. Nothing looks wrong. No warning pops up, no red banner, no reason to hesitate.
That missing warning is the whole story. Behind the scenes, Outlook quietly reaches out to a server controlled by an attacker and hands over a piece of your identity: your Windows login hash. You never clicked a link. You never typed a password. You just opened a file.
Microsoft says attackers are already doing exactly this in the wild. And the fix arrived on Tuesday, November 10, as part of the biggest Patch Tuesday the company has shipped in months — a release carrying more than a thousand fixes, one of which was found with help from an AI model.
Microsoft November Patch Tuesday by the Numbers
The headline figure depends on who is counting, but every count lands in the same place: enormous. Microsoft itself assigned 1,093 vulnerabilities in the November release, with 63 rated Critical. Tenable’s analysis matched that total, breaking it down as 63 Critical, 1,029 Important and a single Moderate flaw.
BleepingComputer counted higher — 1,235 vulnerabilities, including 66 Critical — because its tally adds 142 flaws in the Chromium-based Microsoft Edge. Either way, this is the second month in a row that Microsoft has crossed the 1,000-flaw line, after more than 1,400 fixes in October and roughly 900 in September.
The year’s total has now passed 5,160 disclosed vulnerabilities, according to Tenable’s Satnam Narang, with one more monthly update still to come on December 8.
Nearly half of November’s flaws — 48.9 percent — were elevation-of-privilege bugs, the kind that let an attacker who is already inside a system gain more power over it. Remote code execution flaws made up another 27.1 percent.
The Zero Day Initiative judged 30 of the flaws wormable, meaning they could spread from machine to machine without human help.
The Outlook Zero-Day That Skips the Warning
Out of more than a thousand fixes, only one flaw was already being exploited. That is the one that matters most, and it is the only Moderate-rated entry in the entire release: CVE-2026-124194, a spoofing flaw in Outlook, carrying a CVSS score of 6.5.
The attack is simple enough to explain in one sentence. An attacker sends a specially crafted message file, and when the victim opens it, Outlook connects to an attacker-controlled network location without showing its usual external-content warning. That connection discloses the user’s NTLMv2 hash.
An NTLMv2 hash is not a password, but it is close enough to be dangerous. Attackers can try to crack it offline to recover the password, or use it directly to authenticate as the user in certain attacks.
Microsoft’s own advisory puts it bluntly: the attacker “could use this to authenticate as the user.”
There are two details that make this flaw stranger than its modest severity score suggests. The first is that the Preview Pane is not an attack vector — you have to actually open the file.
The second is that a Windows update alone does not fix it. The flaw lives in Office, so the Microsoft 365 Apps or Office update has to be installed separately.
Affected products are Microsoft 365 Apps for Enterprise and Office LTSC 2021 and Office LTSC 2024. The US Cybersecurity and Infrastructure Security Agency added the flaw to its Known Exploited Vulnerabilities catalog on the same day Microsoft disclosed it, ordering federal civilian agencies to patch by November 24.
Dustin Childs, head of threat awareness at the Zero Day Initiative, wrote in his monthly review that “the impact is bigger than the name suggests.” A spoofing flaw sounds harmless. A stolen login hash is not.
The Flaw Claude Helped Find
The most surprising line in Microsoft’s advisory is not in the technical description. It is in the credits. Microsoft lists the discovery of the Outlook zero-day to Benoît Sevens “using Claude with Anthropic,” alongside a researcher who goes by the name “nono.”
An AI model helped find a flaw in the world’s most widely used email client, and attackers found it useful enough to exploit before Microsoft could ship a fix. It is a neat summary of where security research is heading — and of how fast the other side moves.
The timing is hard to miss. Just a day earlier, Anthropic had expanded Claude into a full office suite, adding spreadsheets, whiteboards and meeting tools to its workplace push — a story we covered here on Wasl Tech. The same family of models is now showing up in Microsoft’s own security credits.
Microsoft has not said who is exploiting the Outlook flaw, or at what scale. The flaw was not publicly disclosed before the fix, which means the attackers who found it were keeping it quiet.
What Else to Patch First
Once the Outlook flaw is closed, the next priorities are the flaws rated 9.8 out of 10 — the highest severity in the release. A Windows Search remote code execution flaw, CVE-2026-116928, can be triggered with no clicks at all through Exchange indexing. An LDAP flaw, CVE-2026-119958, carries the same score.
A Schannel flaw, CVE-2026-119952, scored 8.1 and was marked “Exploitation More Likely” by Microsoft, which is the company’s way of saying attackers will probably weaponise it soon. An Exchange Server remote code execution flaw, CVE-2026-118207, rounds out the shortlist.
Security teams reading the release will also note a quieter milestone buried in it: this is the final monthly security update for the Enterprise and Education editions of Windows 11 version 23H2. Organisations still running those editions need to move on, or they will stop receiving fixes altogether.
The wider week has been busy for defenders everywhere. Google fixed Chrome’s eighth exploited zero-day of 2026, CVE-2026-125350, on November 9 — a day before Microsoft’s release. SAP fixed four flaws rated a perfect 10.0 on the same day. It has been a week where the phrase “patch now” actually means now.
Why the Count Keeps Climbing
Two straight months above 1,000 fixes is not normal, even for Microsoft. October’s release topped 1,400. September sat near 900. The question security researchers keep asking is whether Microsoft is finding more bugs, or whether more bugs are being reported — and Microsoft has not answered it.
Part of the answer may be AI itself. As AI-assisted research becomes standard, more flaws get found, reported and fixed — including flaws found with the help of models like Claude. The same tools that find flaws for defenders can find them for attackers, too.
The other part of the answer is scale. Windows, Office, Exchange, Azure and Edge are among the most attacked software products on the planet. Every month brings a fresh crop of researchers probing them, and a fresh crop of fixes.
That is worth remembering the next time a security update looks boring. Behind every Patch Tuesday is a list of flaws that somebody, somewhere, was hoping nobody would find.
This month, one of them was already being used — and the only thing standing between an opened email and a stolen login was a warning that never appeared.
The same lesson runs through the year’s other big breaches. Fireworks AI’s October breach exposed customer datasets and model weights after attackers slipped through a cloud build system.
OpenAI built a kill switch that can halt a rogue model in 35 seconds. The pattern is the same every time: a small gap, a fast attacker, and a fix that arrives just in time — or just too late.
What to Do Now
For home users, the advice is simple: install the November updates and do not stop at Windows Update. Open your Office apps and check for updates there too, because the Outlook fix ships through the Office channel. On Microsoft 365 Apps, the update should arrive automatically, but it is worth confirming it landed.
For IT teams, the priority order is clear. Patch the Outlook flaw first through the Office update, then the 9.8-rated Windows Search and LDAP flaws, then Schannel and Exchange. Federal agencies have until November 24 under CISA’s order; everyone else should treat that date as a deadline, not a suggestion.
And for everyone: be careful with unexpected message files, even from senders you recognise. The Outlook flaw worked precisely because nothing looked wrong. The warning that should have appeared did not — and in a month of more than a thousand fixes, that missing warning was the most dangerous bug of all.
