Attackers are impersonating Microsoft 365 outage alerts in a high-volume phishing campaign, telling recipients that held emails will disappear within 24 hours unless they act to recover them, Microsoft warns. Microsoft Threat Intelligence detected the coordinated activity on November 13 across organizations in multiple industries worldwide. Microsoft considers credential theft the likely objective, but it has not confirmed stolen credentials, malware delivery or subsequent account compromise.
One sample invoked a supposed November 12 Exchange Online incident and told the recipient that four emails had been held back. Its recovery button sent users into a chain of tracking links that ultimately reached a website outside Microsoft. Microsoft advises checking service problems through the Microsoft 365 admin center, saying genuine service-health messages do not send people to outside domains to retrieve held email.
Read more: Threat Groups Hijack Microsoft 365 Accounts Using OAuth Device Code Exploit
The emails combined outage-related subjects with random character strings and display names using a three-character prefix intended to resemble an organization. Links were tailored to individual recipients, while several samples converged on the same destination outside Microsoft, making simple pattern-based detection harder, according to the company. Microsoft also found a fabricated email exchange placed where recipients would not normally see it, which it assessed as an attempt to evade mail filters.
Despite the suspected credential-theft objective, Microsoft lacked the destination page's source code, evidence of password fields or telemetry showing credentials being submitted. The samples relied on links, with no observed QR codes or malware payloads, and Microsoft has not assigned the activity to a named attacker group. Microsoft says Defender blocks the malicious domains and Defender for Office 365 stops and quarantines the phishing messages.
For organizations investigating possible exposure, Microsoft recommends locating delivered messages and checking who clicked through, using browser evidence, proxy records and click telemetry. If someone may have supplied credentials, it recommends password resets, session and refresh-token revocation, and checks of multifactor authentication registrations. Investigators should also examine suspicious sign-ins, forwarding rules and unusual access to cloud services, according to Microsoft.
Earlier this month, Microsoft described a separate campaign in which Storm-3358 used the ARTEX agentic AI framework against financial organizations, with activity beginning in October. That activity included exploiting exposed systems and deploying remote access malware to retain access and move between systems.













