Storm-0940 is a China state-aligned hacking group that, according to Microsoft, used ARTEX, an open-source AI agent framework, to automate the ConnectWise cloud backup attack. The agents sprayed passwords, abused trusted app access, and reached admin-level access in some Microsoft 365 tenants in under three hours. The defenses still work, but they have to be in place before the attack.
If Spot On Tech manages your Microsoft 365, you were not affected. We do not use ConnectWise Cloud Backup or Cloud Management for the businesses we support, and we checked every tenant we manage for the affected application IDs.
Current as of October 28, 2026. Sources: Microsoft Threat Intelligence's October 21 report, the ConnectWise Trust Center, the Huntress investigation, and GCN.
Key takeaways
- AI agents are now running real attacks. Microsoft documented a nation-state group using an AI agent framework to automate scanning, password spraying, and the abuse of trusted apps.
- The techniques are old. The speed is new. Work that once took a team days happened in hours, against thousands of organizations at once.
- Basic defenses held. The password spray phase failed completely: lockouts made up nearly 98% of the activity and no sign-in succeeded.
- App identities were the weak point. Trusted vendor apps sign in without MFA, so stolen app credentials skipped the protection people rely on.
- Response time is shrinking. With privileged access in under three hours, problems have to be caught by monitoring that runs around the clock, not a weekly review.
Timeline
| Date (2026) | What happened |
|---|---|
| September | Microsoft sees Storm-0940's ARTEX-driven activity begin. |
| October 10 to 14 | A password spray hits 327 accounts at eight organizations, peaking at about 2,900 failed sign-ins a minute on October 11. No sign-in succeeds. |
| October 15 | Huntress sees ConnectWise application identities signing in to customer tenants from the attacker's addresses, and one application granting itself Global Administrator. |
| October 16 and 17 | ConnectWise publishes its advisory, disables seven application identities, and suspends Cloud Backup. |
| October 21 | Microsoft publishes its report naming Storm-0940 and describing its use of ARTEX. |
| October 28 | ConnectWise begins restoring Microsoft 365 backups through a new application. |
What is Storm-0940?
Storm-0940 is Microsoft's name for a China state-aligned group active since at least 2021 that steals data from organizations across many industries. It has a long record of password spraying, exploiting internet-facing applications, and then adding its own credentials to existing Microsoft 365 applications so it can keep reading email and files through Microsoft Graph. Microsoft assesses that Storm-0940 was responsible for the ConnectWise Cloud Backup (formerly SkyKick) breach.
What is ARTEX?
ARTEX is an open-source penetration testing framework: software built so security teams can test their own defenses. It runs several AI agents, powered by whichever large language model you connect, that carry out a test from reconnaissance through data exfiltration with little human input, mapping what they find as they go. In other words, it was built for defenders, and Storm-0940 pointed it at real victims.
What did the AI agents actually do?
According to Microsoft's report, the agents sped up every stage of the campaign:
- Password spraying. More than 151,000 failed sign-ins against 327 accounts in five days, at a peak of about 2,900 a minute. Account lockouts stopped it cold.
- Abusing trusted apps. With four stolen ConnectWise application credentials, the group attempted sign-ins to more than 2,250 Microsoft 365 tenants, with more than 200 sign-ins in under four minutes. Apps sign in without a person, so there was no MFA prompt to stop them.
- Staying in. Inside affected tenants, the attacker added credentials to applications, changed groups and privileged roles, and in one case had an application grant itself Global Administrator.
- Going after secrets. Microsoft saw bulk reads of secrets from Azure Key Vault, the service many companies use to store passwords and keys for their own systems.
Most attempts failed. But in several cases, the time from first targeting to privileged access in a company's cloud was less than three hours. Microsoft also notes that finding an affected application or a successful sign-in does not by itself prove that data was taken.
What changes when attackers use AI agents?
Speed and scale, not new tricks. Password spraying and abusing trusted apps are years old. What AI agents add is the ability to run those techniques against thousands of organizations at once, react to what they find, and move from one step to the next without waiting for a person. Three things follow for a business:
- The gaps you already have get found faster. An unused app with broad permissions, an account without MFA, or a forgotten admin login is more likely to be tested, and sooner.
- Your response window shrinks. A problem spotted in Monday's report is too late when admin access can be gained in three hours.
- Trusted vendors are the shortcut. Breaking one vendor's app credentials opened a path into thousands of customers at once.
How do you protect your business from AI-driven attacks?
1. Make password attacks fail
The spray phase of this campaign failed because accounts locked out and no password worked. Require phishing-resistant MFA for every person, block legacy sign-in methods that skip MFA, keep account lockout on, and remove shared or service accounts that rely on a password alone.
2. Audit the apps connected to Microsoft 365
List every enterprise application and consent grant in Microsoft Entra, remove the ones nobody uses, and cut permissions down to what each app needs. Require admin approval before anyone can connect a new app. Our ConnectWise SkyKick checklist walks through the review step by step.
3. Watch app identities, not just people
Set alerts for credentials added to applications, new privileged role assignments, and app sign-ins from unusual addresses. These were the signs of this attack, and they do not show up in a normal review of user sign-ins.
4. Monitor around the clock
When an attacker can reach admin access in hours, someone or something has to be watching at 2 a.m. and able to act. That means alerts that reach a person who can respond, or automated responses that disable a suspicious app or account first and ask questions after.
5. Protect stored secrets
Review who and what can read your Azure Key Vault or other password and key stores, limit apps to the secrets they need, and rotate secrets after any suspected exposure.
6. Hold vendors to the same standard
Ask the vendors connected to your Microsoft 365 how they protect their application credentials, whether their partner portals require MFA, and how they would tell you about a breach. If you use ConnectWise Cloud Backup, reconnect only through its new application.
7. Keep recovery independent
Keep at least one backup that an attacker in your Microsoft 365 or your backup vendor cannot change or delete, and test restoring from it. Our guide to securing your firewall and VPN covers the same layered approach for the network side.
Should your business use AI agents too?
The same technology that ran this attack can save a small business real time, from answering routine IT questions to drafting reports. The lesson is to approve agents the way you would approve a new employee with system access: decide what each agent may touch, give it the least access that works, and log what it does. Our posts on AI agents for small business and how to vet an AI tool explain how, and our AI governance and AI implementation services can set it up with you.
What to ask your IT provider
- Which third-party apps can sign in to our Microsoft 365, and what can each one access?
- Can employees connect new apps on their own, or does an admin have to approve them?
- Would we get an alert if an app added a credential or gave itself an admin role?
- Who watches our alerts overnight and on weekends, and what can they do on their own?
- Does every person use phishing-resistant MFA, and are any accounts still password-only?
- Can our backups be changed or deleted by someone who gets into Microsoft 365?
How Spot On Tech can help
For the businesses we manage, Microsoft 365 app reviews, MFA, and security monitoring and response are part of the same service. Our cybersecurity services cover the defenses above, and a technology risk assessment finds the over-permissioned apps and forgotten accounts that automated attacks look for. If you have your own IT team, Compass can add 24/7 coverage or security operations center (SOC) monitoring without replacing them.
If you are a New York or New Jersey business and want to know what an AI-driven attack would find in your Microsoft 365, talk with Spot On Tech.
Frequently asked questions
What is Storm-0940?
Storm-0940 is Microsoft's name for a China state-aligned hacking group active since at least 2021. Microsoft says it was behind the October 2026 ConnectWise Cloud Backup breach and used the ARTEX AI agent framework to automate the attack.
What is ARTEX AI?
ARTEX is an open-source penetration testing framework that runs AI agents through an attack from reconnaissance to data exfiltration with little human input. It was built for security testing, and Storm-0940 used it against real organizations.
How many organizations did Storm-0940 hit?
Microsoft says the group attempted sign-ins to more than 2,250 Microsoft 365 tenants using stolen ConnectWise application credentials. Most attempts failed, and Microsoft has not said how many organizations were successfully compromised.
Does MFA stop AI-driven password attacks?
In this campaign, account lockouts stopped the password spray and no sign-in succeeded, and MFA protects people's accounts the same way. But apps sign in without MFA, so stolen application credentials bypassed it. Protecting app identities matters as much as MFA for people.
Are AI agents being used in cyberattacks?
Yes. Microsoft documented Storm-0940 using AI agents to automate scanning, password spraying, and the abuse of trusted applications in 2026, reaching privileged access in some organizations in less than three hours.
How do I check if my Microsoft 365 was affected?
Check whether any ConnectWise Cloud Backup or Cloud Management application is connected to your tenant, then review its sign-ins, role changes, and added credentials. Our ConnectWise SkyKick checklist lists the steps and application IDs.
