JadePuffer Uses Compromised Azure Credentials in Destructive Cloud Attack
Summary
Microsoft says JadePuffer compromised two Azure service principals, then deleted storage accounts and other resources. Exposed credentials may have been a factor, but Microsoft could not confirm their use and found no confirmed data exfiltration or ransom note.
Key points
- The attack targeted one Azure tenant in early June and was attributed by Microsoft to JadePuffer, also tracked as Storm-3168.
- Two compromised service principals were used for reconnaissance, destructive actions, and credential collection.
- The attacker made more than 100 storage-account deletion attempts, most of which succeeded, and deleted a Key Vault, Function App, and App Service plan.
- Attempts to delete Azure SQL databases failed because the attacker used an unsupported API version.
- The attacker later made more than 30 successful requests to retrieve storage-account access keys.
- Microsoft found the service principal's client ID, secret, and tenant ID had been exposed in a public GitHub issue, but could not confirm those credentials were used in the attack.
- Microsoft reported no observed ransom note or confirmed data exfiltration; its mitigations include rotating exposed credentials, protecting secrets, and applying least privilege to service principals.
Article Details
- Victim Organization
- Not disclosed
- Incident Type
- Compromised Azure service principals followed by reconnaissance and attempted destructive cloud-resource operations consistent with ransomware or extortion activity
- Incident Date
- 2026-06-01
- Operational Impact
- More than 100 attempts to delete storage accounts, most of which succeeded; an Azure Key Vault, Function App, and App Service plan were also deleted. Attempts to delete multiple Azure SQL databases failed. Successful data exfiltration was not confirmed.
- Ransom Or Extortion
- Activity was consistent with ransomware or extortion operations, but no ransom note was observed.
- Claim Status
- confirmed
People
Nick TausekLead security automation architect at Swimlane; said the Azure evidence showed coordinated automation but did not prove AI directed each step.Ross FilipekChief information security officer at Corsica Technologies; commented on the exposed credentials.Tushar MudiMicrosoft researcher who co-authored the account of the attack.Yossi WeizmanMicrosoft researcher who co-authored the account of the attack.
Threat Actors
Vendors
MicrosoftAn agentic AI attacker turned compromised Microsoft Azure credentials into a weapon for destroying cloud resources in minutes in what appeared to be a ransomware or extortion attack.SysdigMicrosoft attributed the attack to JadePuffer, which it tracks as Storm-3168. Sysdig researchers identified the group in July as the first documented large language model (LLM)-driven ransomware operation.
Products
AzureAn agentic AI attacker turned compromised Microsoft Azure credentials into a weapon for destroying cloud resources in minutes in what appeared to be a ransomware or extortion attack.MDASHthan requiring analysts to manually follow each individual action, efforts such as Project Perception and MDASH are intended to support a model in which defenders can investigate and respond across increasingly largeMicrosoft Defender for Cloudof activity similar to that conducted by JadePuffer. This advice includes enabling appropriate Microsoft Defender for Cloud plans for critical Azure workloads; protecting and continuously assessing applicationProject Perception"Rather than requiring analysts to manually follow each individual action, efforts such as Project Perception and MDASH are intended to support a model in which defenders can investigate and respond across increasingly