Researchers show BluePhishing can abuse Microsoft Entra AI agents to take over Microsoft 365 tenants
- Severity
- Medium
- Sources
- 1
What happened
BluePhishing abuses Microsoft Entra Agent ID: an attacker creates a blueprint with signInAudience set to AzureADMultipleOrgs and a client secret, then lures a victim into user consent. The resulting blueprint principal holds AgentIdentity.CreateAsManager, which Proofpoint says cannot be revoked, enabling creation of up to 250 agent identities per blueprint. Test agents held User.Read.All and Group.Read.All and were used via GraphRunner. Hunt for "Add application" and "Add service principal" audit events, review multi-tenant blueprints by blueprint ID, apply user consent limits to blueprints and agent identities, and enforce Conditional Access and Entra ID Protection for agents. No CVE is assigned.
For security analysts · Written by AI from 1 source · Wrong?
What to do
- Train staff to be cautious when a third-party agent asks for permissions.
- Apply the same user consent limits to agent blueprints and agent identities.
- Watch the audit log for Add application and Add service principal events that indicate new blueprints or agent identities.
- Apply least privilege, Conditional Access policies and Entra ID Protection to agents.
Written by AI from the sources, checked against them
Full story timeline · 1 update
Who reported · 1 outlet
Who and what is in it
Companies