JadePuffer attacks target Azure and destroy cloud resources

JadePuffer, a ransomware operator that uses AI agents, targeted Azure tenants in attacks that mapped cloud resources, collected credentials and deleted storage accounts, Microsoft Security Research reported. Microsoft tracks the actor as Storm-3168.

Researchers at cloud security company Sysdig said JadePuffer emerged in July and used AI agents to automate steps from reconnaissance and credential theft to lateral movement, persistence and data encryption. Sysdig later reported that the operator expanded its focus to AI assets, training datasets and vector databases, using a tool called EncForge.

Microsoft observed two attacks in June. The actor used two compromised service principals in the same tenant: one for reconnaissance and resource discovery, and the other for discovery, destructive operations and credential collection. The destructive activity lasted seven minutes and targeted more than 100 storage accounts, along with Key Vaults, Function Apps, Virtual Machines and App Services. Azure resource locks and account-level protections left some storage accounts unaffected.

The actor also removed Azure Site Recovery locks, which protect backup and recovery. Attempts to delete Azure SQL databases failed because the attacker used an unsupported API version; attempts to remove recovery protection locks also failed. About half an hour after the deletion attempts, Storm-3168 made more than 30 requests for storage account keys, most of which succeeded.

Microsoft could not determine how the actor first gained access. It said credentials for one service principal had appeared in a public GitHub issue before the attacks. Microsoft did not confirm data theft or report financial demands in the observed cases.