Vernabilities, Threat Assessment - no fear prep
20 Nov 2026, 10:00 UTC

SITREP · 12–20 NOVEMBER 2026

Vernabilities, Threat Assessment - no fear prep

A week of failures. No evidence of a coordinated attack.

Microsoft, the FAA, Chase, United and an NSA network all went down within days. Each published a different, unrelated cause. The real risks to watch now sit elsewhere: a faulted Gulf undersea cable and a confirmed campaign against cold-chain controllers.

6.4M

US Downdetector reports in one week — ~4× normal

~10,000

Flights canceled after the FAA ground stop

70+

Countries hit by the Microsoft 365 outage

0

Verified links to a coordinated cyberattack

DEFENSE / RESEARCH UPDATE

Defender strength & susceptibility

Reviewed 20 Nov 2026, 10:27 UTC · Qualitative judgments, not measured scores or collapse predictions.

DEFENDER STRENGTH

Improving, unevenly

Frontier AI is helping equipped defenders find flaws earlier. Anthropic reported more than 10,000 high- or critical-severity findings with Glasswing partners in May; that is a vendor-reported cumulative count, not an annual rate or a count of deployed fixes.

SYSTEM SUSCEPTIBILITY

Persistent, uneven

Finding bugs faster does not guarantee that every affected system is patched. Exposure remains greatest where inventories, staff, tested fixes and recovery capacity lag. This assessment does not establish that the outages above were cyberattacks.

COUNTEREVIDENCE / NOVEMBER 4

An observational research preprint found that Glasswing-access firms more than doubled vulnerability disclosures, kept fixing reported flaws as quickly as before, and saw a lower share exploited within a month. The number exploited did not change. This is evidence of defensive gains—not proof that all defenders are falling. The study is not peer-reviewed and does not establish universal protection.

Who is giving defenders a head start?

Anthropic

Glasswing → Cyber Verification Program

Glasswing launched in April; October’s expanded CVP moves existing partners into Specialized Access and offers verified defensive-access tiers.

Google DeepMind

Fairwind Program

Early frontier-model access for high-priority defenders, with Gemini for cyber and CodeMender for vulnerability research and patching.

Four pressures that stronger tools alone cannot remove

01

Legacy code & asymmetric exposure

Attackers need a viable entry point; defenders must prioritize many systems and dependencies. Perfect bug elimination is not realistic—segmentation, least privilege and recovery plans can still contain damage.

02

A shrinking response window

Google reports that agent-assisted flaw discovery and attacks can compress weeks of specialist work into minutes. Discovery is not the finish line: validation, safe patching and deployment must keep pace. A zero-day is a flaw exploited before a fix is available to defenders.

03

More code, more review pressure

Faster software production increases what teams must review. AI-written code is not automatically insecure—or secure. The decisive measures are exploitable defects, test coverage and time to deploy a verified fix, not lines of code alone.

04

Access, funding & policy gaps

Restricted access can leave smaller teams behind. Shared open-source dependencies need maintainers, review capacity and sustained patch support; reporting rules do not supply these by themselves. This is a structural exposure assessment, not proof that every volunteer project is insecure.

Claim checks & reporting limits
  • Google’s verified program name is Fairwind, not “Fairplay.”
  • The supplied “175,000 bugs a year” figure and claim of a post-open-model zero-day surge are not substantiated by the sources checked; neither is used as a fact or score.
  • Japan’s official record dates passage of its cyber-response laws to May 16, 2025, not October 2026. Passage and commencement are different milestones.
  • The $63 million initiative, South Korea’s stated 10% revenue fine, the UK reporting deadline, the US private-offensive-operations claim and the attributed Tim Stevens quote remain unverified here and are excluded from the assessment.

HOUSEHOLD / FOUR PRACTICAL STEPS

No fear preparation

Prepare calmly. Look after one another. These steps are useful for ordinary outages as well as cyber incidents. No evidence here predicts a nationwide catastrophe in the coming weeks.

Practical considerations before you buy

  • Start with what you have. Check your cupboard, flashlight, contact list and chargers. Choose one affordable improvement rather than buying everything at once.
  • Plan around real needs. Account for dietary restrictions, pets, limited storage, mobility and medical equipment. Choose supplies you can carry, open and use.
  • Keep the plan usable. Rotate food and water, check expiry dates and keep paper contacts where everyone can find them. Agree on a meeting place if phones fail.
  • Stay informed, not alarmed. Check local emergency notices and your provider’s official updates. Verify claims before sharing them; a forecast is not a certainty.

YOUR HOUSEHOLD / 14 DAYS

Readiness planner

Needs appear on your card; they do not determine medical doses or dietary calories. Choose suitable meal portions and agree any medicine, special nutrition or device plan with a clinician. Notes stay in this browser; avoid sensitive details on shared devices.

14

gallons water

42

person-meals

0

pet portions

14 gallons for people + 0 for pets. Water includes drinking and sanitation; some needs require more. Food uses 3 meals per person per day—not a calorie prescription.

Ready when you need it

0/7 ready

Loading saved checklist…

STEP 01

Build a food & water cushion

14 days · 14 gallons per person

Build gradually, without panic buying. Aim for two weeks of shelf-stable food: beans, oats, canned fish and other foods your household can eat without refrigeration. Ready.gov recommends at least one gallon of water per person per day for drinking and sanitation; some people and hot climates need more. The 14-day cushion is this briefing’s planning target, not a prediction or a universal FEMA minimum.

STEP 02

Keep health & payments resilient

Paper records · lawful refills · modest cash

Ask your doctor or pharmacist about permitted refills and an interruption plan—do not change doses or stockpile beyond their advice. Print medications, doses, allergies and provider contacts. Keep essential bills current where practical and a modest amount of small-denomination cash securely accessible at home; an inaccessible bank or safe-deposit box is not an outage cash plan.

STEP 03

Protect devices & family calls

Safe backup power · verify the caller

Never operate fuel-burning generators indoors or in a garage; use them outdoors, away from doors, windows and vents, following safety instructions. Ask the device supplier or clinician to verify battery compatibility, runtime and medicine refrigeration needs; a consumer battery is not a guaranteed life-support solution. Agree on a private family phrase, but also hang up and call a known number before sending money. A code word alone cannot guarantee identity.

STEP 04

Check on older & vulnerable neighbors

A named contact · a practical check-in plan

With their permission, agree who will check in and when. Help with water, food, accessible transport and contact lists. Ask about medical-device backup and cooling needs; arrange support rather than assuming they can lift supplies or manage alone. If someone is in immediate medical danger, contact emergency services—not an informal check-in network.

01 / CHRONOLOGY

Outage timeline

1213141516171819
High severityModerateOngoingBar length ≈ reported duration. Sources: AppleScoop timeline ↗TechRepublic ↗
“We have zero evidence or any reason to believe that any of these isolated or individual instances have anything to do with nefarious actors.”

02 / EXPOSURE

Sector pressure map

Banking & Finance

2

All resolved · peak sev 4/5

Aviation

2

All resolved · peak sev 5/5

Cloud & AI

4

All resolved · peak sev 5/5

Telecommunications

1

Active issue · peak sev 4/5

Government & Security

1

All resolved · peak sev 4/5

Food, Water & Health

1

Active issue · peak sev 4/5

03 / FORECAST

Early-warning board

ELEVATED

60%

Gulf submarine cable disruption widens

SMW5 is faulted off Pakistan with no ETA. Iranian officials have publicly called Hormuz cables 'legitimate targets' and threatened to block repairs. Cause of SMW5 fault is NOT established.

Watch → Additional faults on Gulf cables; repair-ship access; latency in India and Gulf states.

ELEVATED

55%

Cold-chain attacks spread to North America

A joint CISA/FBI/HHS/USDA/DOT warning follows two European waves. Physical consequences — spoiled food and medical stock — are already confirmed abroad.

Watch → US grocery or pharmacy spoilage reports; new CVEs in temperature controllers.

GUARDED

40%

Change-management failures keep producing outages

Most November incidents trace to upgrades, router deployments, or tests — not attacks. Rapid patching of AI-discovered flaws is raising change volume.

Watch → Year-end freeze discipline; vendor post-incident reports; CISA KEV additions (4 added 18 Nov).

GUARDED

35%

OT / PLC targeting in water and industry

CISA's Iranian-affiliated PLC advisory was expanded to Schneider and Siemens; new water-sector ICS advisory (InfoScan) issued 17 Nov.

Watch → Utility operational disruption notices; HMI tampering reports.

LOW

20%

Cascading payment-system failure

Chase's bug reached other banks via payments, proving the dependency path — but recovery was same-day and contained.

Watch → Simultaneous Fedwire/ACH/card-network notices.

WHAT IF / CONTINGENCY LAB

Scenario impact simulator

Illustrative cascades, not verified incident effects or predictions. Outcomes depend on redundancy, local safeguards and outage duration.

Undersea cable severance

A regional cable route becomes unavailable; alternative capacity may be constrained.

Assumed duration

Sustained focus: reconcile delayed work, verify stock and medicine safety, and check vulnerable people.

POSSIBLE CASCADE

  1. 01Telecommunications: rerouting and slower connectivity
  2. 02Cloud & business: delayed remote access and support
  3. 03Payments & logistics: transaction timeouts and delayed coordination

Contingency action matrix

WhoBefore disruptionDuring disruption
HouseholdsSave contacts and essential documents offline; agree a meeting place.Use available SMS or voice; follow provider updates and avoid unverified outage claims.
BusinessesTest an independent carrier or route; queue essential work offline.Prioritize critical traffic, record queued work, and reconcile after service returns.
Care & logisticsKeep paper delivery schedules and care contacts.Confirm critical deliveries by available channels; contact care providers early.

04 / REGISTRY

Incident registry

11 of 11 incidents

GS-2611-001RESOLVEDRESOLVED

Microsoft 365 global outage

Microsoft · Global · 70+ countries

Start
12 Nov
Duration
~24–36 h
Impact
Outlook, Exchange Online, Teams, Copilot; ~7.8M Downdetector reports worldwide
Cause
Issue from an earlier network capacity upgrade surfaced after a network change; bug dating to July
Cyber
Ruled out
GS-2611-002RESOLVEDRESOLVED

Claude M365 connector failure

Anthropic · Global

Start
13 Nov
Duration
Hours
Impact
Microsoft 365 connector broken — knock-on effect of Microsoft outage
Cause
Dependency on Microsoft 365; Anthropic has not published a cause
Cyber
Unknown
GS-2611-003RESOLVEDRESOLVED

Workday service disruption

Workday · Global customers

Start
13 Nov
Duration
Hours
Impact
Payroll / HR systems for enterprise customers
Cause
Backup test
Cyber
Ruled out
GS-2611-004RESOLVEDRESOLVED

FAA nationwide ground stop

FAA / L3Harris · United States

Start
13 Nov · 20:30 UTC
Duration
~4 h 45 m ground stop
Impact
~10,000 flights canceled; longest US shutdown since 9/11
Cause
Data flood (1,000 connection attempts / 4–5 s) on FTI network traced to new router deployment at El Dorado, AR and Amarillo, TX radar site
Cyber
Ruled out
GS-2611-005RESOLVEDRESOLVED

AWS us-east-1 disruption

Amazon Web Services · US East

Start
15 Nov
Duration
Hours
Impact
Services hosted in us-east-1
Cause
Not published
Cyber
Unknown
GS-2611-006RESOLVEDRESOLVED

Chase banking interruption

JPMorgan Chase · United States

Start
16 Nov
Duration
Several hours
Impact
Chase services; delayed payments reached Wells Fargo and other banks' customers
Cause
Software bug in Chase's own hardware systems
Cyber
Ruled out
GS-2611-007RESOLVEDRESOLVED

United Airlines systems outage

United Airlines · United States

Start
16 Nov
Duration
Hours
Impact
Flight operations
Cause
Pilots' union pointed to weight-and-balance system
Cyber
Ruled out
GS-2611-008CORROBORATEDRESOLVED

NSA classified network outage

National Security Agency · United States

Start
17 Nov
Duration
Several hours
Impact
Key classified network unavailable
Cause
Not publicly disclosed
Cyber
Unknown
GS-2611-009CORROBORATEDRESOLVED

Apple, Wells Fargo, BofA disruptions

Multiple · United States

Start
17 Nov
Duration
Hours
Impact
Consumer access to banking & Apple services
Cause
Not detailed; officials say not connected
Cyber
Ruled out
USA Today ↗Fortune ↗
GS-2611-010CONFIRMEDNEW

SEA-ME-WE 5 submarine cable fault

SMW5 consortium / TWA · Pakistan · Gulf of Oman

Start
18 Nov
Duration
Ongoing — no restoration ETA
Impact
Intermittent connectivity and degraded quality in Pakistan; traffic rerouted
Cause
Fault ~11 km offshore; root cause under investigation
Cyber
Unknown
GS-2611-011CONFIRMEDWORSENING

Refrigeration / cold-chain controller attacks

Grocers, cold storage, blood banks · Europe (FR, SE, BE, LU, IE, EE) · US warning

Start
Oct – 5 Nov
Duration
Ongoing campaign
Impact
Hundreds of French stores (Oct), 100+ European stores (early Nov); temps reportedly to 50°C
Cause
Exploitation of internet-exposed temperature controllers
Cyber
Confirmed (campaign)

05 / METHOD

How to read this briefing

CONFIRMEDAuthoritative statement or direct technical evidence.
CORROBORATEDMultiple credible independent reports.
UNVERIFIEDEarly reports or complaints only.
RESOLVEDRestoration confirmed.
DISPUTEDConflicting evidence.

Limitations

This is a point-in-time snapshot compiled 20 Nov 2026, 10:00 UTC from public reporting — not live telemetry. Downdetector counts are user complaints, used only as a preliminary signal. No financial-loss figures are shown because no credible source has published them. Forecasts are analytic judgment.