Credential Attacks

Credential attacks are attempts to steal, guess, or misuse the usernames, passwords, access keys, and API tokens that let people and software log into systems. In AI deployments these tokens often grant access to model platforms, training data, and cloud accounts, so a single stolen key can expose far more than one application. For governance teams, this category matters because it drives requirements for secret management, key rotation, and monitoring of unusual logins.

1 item

Corporate PolicyGlobal2026-10-23

State-Linked AI Attack Hit 2,250 Microsoft 365 Tenants via Trusted App Identities

Microsoft Threat Intelligence linked a large-scale cloud intrusion to Storm-0940, a China-state-aligned hacking group. The attackers used an open-source agentic AI tool called ARTEX to automate credential attacks at 2,900 attempts per minute across more than 2,250 Microsoft 365 tenants. They entered through a breached cloud backup provider and gained privileged access within three hours by impersonating trusted third-party application identities.

agentic AI non-human identity cloud security Microsoft 365 Storm-0940 credential attacks third-party risk China