It Happens Every Day, Somewhere in the World — One Week of Security News (Nov 9–16, 2026), and How to Protect Yourself
The result first.
Between November 9 and 16, 2026, security incidents were reported every day, in Japan and around the world. They looked unrelated. They were not. Almost all of them entered through one of three doors:
- A trusted third party — not the company you use directly, but a supplier, a contractor, or a connected service
- A postponed update — a fix existed, but the machine had not received it yet
- A connection key — the credentials that link one service to another were targeted, and once taken, they opened the main system too
So the protection is also three things, and none of them is difficult:
- Review who you are connected to, and cut what you no longer use
- Install updates — and restart, because the fix only takes effect after the restart
- Never reuse a password, and add a second step (2-step verification or a passkey)
Everything below is the week’s news that led to that conclusion, what it means for an ordinary person, and why I believe this pattern is exactly what makes tools like Muse and Google apps safe enough to use with confidence.
This roundup will continue weekly. This is the first issue; each week will add the next week’s events, and the next layer of protection.
I am Kurashiki Bouzu, born in 1971, a former funeral director. I live in an old house — a kominka — in Kurashiki, Okayama, and I work together with my AI butler Moka, who researched and counted the figures in this article.
Japan: the names were familiar
The Fugaku supercomputer was breached. On November 16, RIKEN confirmed that an outsider had gained unauthorised access to the system used to manage and operate Fugaku, its supercomputer in Kobe. The machine is offline, with no restart date. Reports say users’ personal folders may have been read and private keys may have been taken; the entry point, the attacker, and the number of users are still under investigation. Even one of the world’s great computers can be stopped. The lesson is small and practical: when a key may have leaked, invalidate it and make a new one. Large institutes and ordinary people follow the same order.
A shared shopping system leaked about 4.37 million people’s order data. GoQSystem, an order-management system used by 1,694 online shops in Japan, disclosed unauthorised access on November 12. Names and addresses in about 4.37 million people’s orders were taken, reportedly through a weakness in an inquiry form. Attackers do the arithmetic too: one shared system is worth more than 1,694 individual shops.
A marketing platform used by many companies stopped everything. On November 13, PLAID detected unauthorised access to KARTE, its customer-experience platform, and halted all functions — tracking, delivery, and admin screens — for every project using it. By November 15 it had isolated the affected environment from the network and asked client companies to invalidate the keys (secret keys and access tokens) used for integrations. No restart date has been announced. Notice what the company chose: stopping its own service is painful, and it stopped anyway. Stopping is not losing. Stopping is a defensive move.
The week’s other Japanese disclosures were names people touch daily: Keihan Holdings (November 14 — a service linked to its official LINE account; display names and profile images for up to about 224,000 people may have leaked, with names, addresses or phone numbers in up to about 64,000 of those cases), Keisei Electric Railway (November 13 — its Skyliner seat-reservation system; up to about 3.37 million records may have been taken), Giken Seisakusho (November 9 — ransomware in a business server; about 3,200 people’s information at risk), and a ransomware attack on one construction consultant in Akita that spread to work it held for the prefecture and four cities.
The world: old systems, shared systems, and very large claims
Expedia — read the claim and the confirmation separately. The extortion group n0n claims it stole more than 830 million loyalty-program records from Expedia Group. Separately, Expedia was reported on November 12 to have confirmed that attackers reached its systems through a compromised third-party data platform provider. Those are two different kinds of statement. The group’s numbers are not Expedia’s confirmed numbers. The bigger the number, the more calmly it should be read.
Maersk was entered through an inherited old system. On November 11, the Danish shipping giant confirmed that systems belonging to LF Logistics, a company it acquired in 2022, were breached and data was stolen. A ransomware group claims about 100 GB and is demanding payment — that amount is the group’s claim. The door was not the newest system. It was the inherited one, like a house where only the annex still has the old lock.
The management tool itself was targeted. More than 30 unpatched flaws in N-able’s N-central, software used to manage other companies’ computers remotely, were publicly disclosed on November 8. N-able took its hosted environments offline across the Americas, Europe and Asia-Pacific on November 9 and released emergency hotfixes on November 11. When the tool that manages everything is the door, everything behind it is at risk — which is why the company chose the heavy move of stopping it all.
November’s updates were unusually important. On November 10, Microsoft’s monthly release addressed more than 1,200 vulnerabilities (CrowdStrike’s count). One of them, in Outlook, is confirmed exploited: opening a crafted message file can connect Outlook to an attacker’s server without its usual warning and leak authentication information. Google also fixed Chrome’s eighth confirmed-exploited flaw of 2026 — and a Chrome fix only takes effect after a full relaunch. A browser that was never closed was never fixed. Check Point security gateways were also attacked through chained flaws, added to the US CISA list of known exploited vulnerabilities with a November 16 deadline.
South Korea’s government email saw abnormal logins on more than 300 accounts, reported November 16, reached through weaknesses in some local-government networks and in the single sign-on system. One key that opens everything is convenient — and, if taken, opens everything.
AI is now reducing the attacker’s workload, too. The FBI has warned that an AI penetration-testing framework called ARTEX is powering real attacks, automating reconnaissance and vulnerability testing, according to reports on November 16. Reporting on AI-accelerated ransomware the same day estimated about 3,000 victim listings on leak sites over the summer, up 20–40% since spring, with AI helping to write phishing messages and even negotiate ransoms. Defenders are not the only ones using AI anymore.
Claims, confirmations, and unknowns — keep them apart
This is the reading skill the week taught most clearly:
- A claim is what an attack group says it took. The numbers tend to be large.
- A confirmation is what the affected company or a public agency has checked and admitted.
- Under investigation means it is not known yet — and should be left as not known.
Expedia’s 830 million and Maersk’s 100 GB are, at this point, claims. KARTE’s shutdown and Fugaku’s shutdown are confirmations. Panic should follow the confirmations, not the claims — because the classic follow-up attack is a fake “please verify your account” message that rides on the panic.
How to protect yourself — seven things, starting today
You do not need to do all of these at once. Start from the top.
- Install updates, and restart. Computer, phone, and browser. A restart is part of the update, not an extra.
- Never reuse passwords, and turn on a second step. A leaked password should open exactly one door, and ideally not even that one without your phone. Passkeys (fingerprint or face) are even harder to phish.
- Review your connections. LINE integrations, Google account connections, permissions given to apps you no longer use. Every connection left open is a door left counted. Once a year is enough.
- If a key may have leaked, invalidate it and make a new one the same day. Keys exist to be replaced.
- Never start a real check from a button inside an email or text message. Open the usual app or page yourself, and check there. This one habit blocks a large share of fakes.
- Keep important things in two places. If ransomware locks one copy, your life does not stop. In my house, the master records live in the cloud, so a broken computer at home loses nothing.
- Do not be afraid to stop. Cut a connection, close the screen, disconnect. This week, large companies and a national research institute did exactly that. For an individual, closing an unknown screen and checking calmly is the same move, at a smaller scale.
Why this is a reason to feel safer with Muse and Google apps — not less
None of this week’s doors was “a mysterious new tool”. They were unpatched machines, reused passwords, connections left open, and old inherited systems. Every one of them is a door that the basic pattern makes smaller: updates applied, separate passwords, a second step, connections reviewed.
Google apps carry the machinery for that pattern from the start — a screen that shows which devices are signed in, 2-step verification, and warnings that stop unknown sign-ins. I have written about each of them on this blog, one at a time.
The same holds for Muse, my Moka. What you may tell it, and whether conversations are used for training, can be checked in the company’s official explanations. And I do not hand Moka my passwords themselves; they are used only through a secure vault mechanism. Knowing the system, and deciding your own line, is what safety actually is.
Not “too frightening to start”, but “I know how to protect myself, so I can start with confidence”. If this article becomes that trigger for even a few readers, it has done its job.
The price of progress
Every gain in convenience adds doors. Every wider connection adds places where your information is entrusted. Every smarter AI reduces someone else’s workload — including an attacker’s.
This is the price of progress.
But a price is not a reason to stop. When cars became faster, seatbelts became normal. Convenience has always arrived with a new layer of protection beside it. The way to live with progress is not to count the price in fear, but to add the protection that matches it — and keep going.
The news will happen again next week. When it does, remember the split (claim, or confirmation?) and the three protections (update, password, connections). That is enough.
Sources checked
- Fugaku: reporting on RIKEN’s confirmation, November 16, 2026 (ITmedia NEWS, NHK and others); entry point and user count under investigation
- GoQSystem: company disclosure, November 12, 2026 — 1,694 shops, about 4.37 million people’s order information
- PLAID / KARTE: company disclosure and status page — detected and fully suspended November 13; network isolation completed and integration credentials to be invalidated, November 15
- Keihan Holdings: company disclosure, November 14 — up to 224,386 people (display names etc.), of which personal details in up to 64,332 cases
- Keisei Electric Railway: company disclosure, November 13 — up to about 3.37 million records
- Giken Seisakusho: company announcement, November 9 — ransomware, about 3,200 people
- Akita contractor case: disclosures by Yuzawa, Yokote, Daisen and Yurihonjo cities, early November 2026, and reporting
- Expedia: the n0n group’s claim (recorded by Ransomware.live and others) and reporting of Expedia’s November 12 confirmation of access via a third-party provider; the claim’s figures are not confirmed by Expedia
- Maersk: company confirmation, November 11, 2026, and reporting; the roughly 100 GB figure is the attack group’s claim
- N-able N-central: company response and reporting — flaws disclosed November 8, hosted environments taken offline November 9, hotfixes November 11
- Microsoft November 2026 updates: Microsoft’s release and CrowdStrike’s analysis, November 10 — more than 1,200 vulnerabilities; Outlook CVE-2026-124194 confirmed exploited
- Chrome: Google’s release — CVE-2026-125350, the eighth confirmed-exploited Chrome flaw of 2026; effective after relaunch
- Check Point: company disclosure and CISA’s Known Exploited Vulnerabilities catalogue — CVE-2026-123039, due November 16
- South Korea government email: reporting of the Ministry of the Interior and Safety announcement, November 16 — abnormal logins on more than 300 accounts
- AI-powered attacks: reporting of the FBI warning on ARTEX, November 16, and WIRED’s reporting on AI-accelerated ransomware growth, November 16 (about 3,000 leak-site listings over the summer, up 20–40% since spring, as an estimate)
Read next
- The basic pattern: If You Know How to Protect Yourself, You Can Begin — Security Basics
- The second step, from real experience: That “Yes” and a Number on Your Phone — Google’s Prompt and 2-Step Verification
- Checking signed-in devices: You Can See Which Devices Are Connected — Checking “Devices” on Your Google Account
- What you may tell an AI: Comparing AI Safety — Are Conversations Used for Training?
- November 17, 2026 ― Draft created — English version (result-first rewrite of the Japanese article, from public reports and company disclosures of Nov 9–16, 2026)(10,941 characters)
- November 17, 2026 ― Published(10,941 characters)
