Azure Automation Flaw Allowed Cross-Tenant Privilege Escalation

Critical flaw in Azure's workflow automation service

Microsoft has remediated a severe security vulnerability in Azure Automation that could have allowed attackers to escalate privileges and potentially breach the trust boundaries separating different Azure tenants. The flaw, designated CVE-2025-29827, earned a CVSS severity score of 9.9 out of 10.

Azure Automation is a cloud service organizations use to run automated workflows—called runbooks—for tasks like resource deployment, configuration management, credential rotation, and update orchestration across Azure and hybrid infrastructure. Because these automation accounts frequently handle sensitive administrative operations, a compromise could expose critical cloud resources and protected configuration data.

Microsoft security researcher Shay Shavit discovered the vulnerability last year and reported it through the company's internal security response process. The flaw is exploitable remotely, requires only low-level privileges to trigger, and does not rely on tricking users into taking action.

Why it matters

This vulnerability highlights a particularly dangerous class of cloud security risk: flaws that can break tenant isolation. In multi-tenant cloud environments, strict boundaries between customers are fundamental to the security model. When researchers demonstrated that CVE-2025-29827 could be chained with a problematic default configuration to cross tenant boundaries, it revealed a potential path for attackers to impersonate another organization's automation identity and access their credentials, workflows, and cloud assets. For enterprises running sensitive workloads in Azure, such cross-tenant access scenarios represent a worst-case breach of trust in shared infrastructure.

How the vulnerability worked

The core issue involved weaknesses in how Azure Automation validated and enforced access permissions. While an attacker would need legitimate access to an Azure Automation account as a starting point, the flaw allowed them to move beyond their authorized scope and gain control over identities with elevated privileges.

Researchers found that when combined with a risky default configuration setting and other application weaknesses, the vulnerability could be exploited to cross trust boundaries between separate Azure tenants. In a successful attack, a malicious actor could impersonate automation identities belonging to other organizations, creating unauthorized access to their automated processes and protected resources.

Recommended security measures

Microsoft has addressed both the underlying authorization flaw and modified the default configuration that contributed to the risk. Organizations should verify that their Azure Automation deployments have received the latest security updates and that automation accounts align with current security guidance.

The company recommends implementing the principle of least privilege across automation environments. Administrators should regularly audit managed identities, permissions, and access tokens tied to automation accounts to ensure they hold only the minimum rights necessary for their intended functions. Organizations should also avoid exposing cloud services or endpoints externally unless required and continuously monitor for misconfigurations that could combine with other vulnerabilities to create expanded attack surfaces.

Petri.com first reported the details of this vulnerability and Microsoft's response.