Nov 28, 2025Ravie LakshmananEmail Safety / Enterprise Safety
Cybersecurity researchers have make clear a cross-tenant blind spot that permits attackers to bypass Microsoft Defender for Workplace 365 protections by way of the visitor entry characteristic in Groups.
“When customers function as friends in one other tenant, their protections are decided completely by that internet hosting surroundings, not by their house group,” Ontinue safety researcher Rhys Downing mentioned in a report.
“These developments enhance collaboration alternatives, however additionally they widen the accountability for guaranteeing these exterior environments are reliable and correctly secured.”
The event comes as Microsoft has begun rolling out a brand new characteristic in Groups that permits customers to talk with anybody by way of e mail, together with those that do not use the enterprise communications platform, beginning this month. The change is predicted to be globally obtainable by January 2026.
“The recipient will obtain an e mail invitation to affix the chat session as a visitor, enabling seamless communication and collaboration,” Microsoft mentioned in its announcement. “This replace simplifies exterior engagement and helps versatile work situations.”
Within the occasion the recipient already makes use of Groups, they’re notified by way of the app instantly within the type of an exterior message request. The characteristic is enabled by default, however organizations can flip it off utilizing the TeamsMessagingPolicy by setting the “UseB2BInvitesToAddExternalUsers” parameter to “false.”
That mentioned, this setting solely prevents customers from sending invites to different customers. It doesn’t cease them from receiving invites from exterior tenants.
At this stage, it is price mentioning that visitor entry is completely different from exterior entry, which permits customers to seek out, name, and chat with individuals who have Groups however are exterior of their organizations.
The “elementary architectural hole” highlighted by Ontinue stems from the truth that Microsoft Defender for Workplace 365 protections for Groups could not apply when a person accepts a visitor invitation to an exterior tenant. In different phrases, by getting into the opposite tenant’s safety boundary, the person is subjected to safety insurance policies the place the dialog is hosted and never the place the person’s account lives.
What’s extra, it opens the door to a situation the place the person can change into an unprotected visitor in a malicious surroundings that is dictated by the attacker’s safety insurance policies.
In a hypothetical assault situation, a risk actor can create “protection-free zones” by disabling all safeguards of their tenants or avail licenses that lack sure choices by default. As an example, the attacker can spin up a malicious Microsoft 365 tenant utilizing a low-cost license equivalent to Groups Necessities or Enterprise Primary that does not include Microsoft Defender for Workplace 365 out of the field.
As soon as the unprotected tenant is about up, the attacker can then conduct reconnaissance of the goal group to collect extra info and provoke contact by way of Groups by getting into a sufferer’s e mail handle, inflicting Groups to ship an automatic invitation to affix the chat as a visitor.
Maybe essentially the most regarding side of the assault chain is that the e-mail lands on the sufferer’s mailbox, provided that the message originates from Microsoft’s personal infrastructure, successfully bypassing SPF, DKIM, and DMARC checks. E-mail safety options are unlikely to flag the e-mail as malicious, because it’s legitimately from Microsoft.
Ought to the sufferer find yourself accepting the invitation, they’re granted visitor entry within the attacker’s tenant, the place all subsequent communication takes place. The risk actor can ship phishing hyperlinks or distribute malware-laced attachments by making the most of the dearth of Protected Hyperlinks and Protected Attachments scans.
“The sufferer’s group stays fully unaware,” Downing mentioned. “Their safety controls by no means triggered as a result of the assault occurred exterior their safety boundary.”
To safeguard in opposition to this line of assault, organizations are advisable to limit B2B collaboration settings to solely enable visitor invites from trusted domains, implement cross-tenant entry controls, limit exterior Groups communication if not required, and practice customers to be careful for unsolicited Groups invitations from exterior sources.
The Hacker Information has reached out to Microsoft for remark, and we are going to replace the story if we hear again.
