Default Azure Automation Setting Enables Cross-Tenant Identity Takeover
A critical vulnerability in Microsoft's Azure Automation service, stemming from a default public configuration for automation account identities, could have allowed attackers to take over another tenant's identity and access their data and cloud resources. While the default setting has been corrected, researchers warn that organizations should audit their identity scope and avoid exposing identities unnecessarily, highlighting a broader trend of cloud risk centered around compromised identities. The vulnerability, initially discovered by Shay Shavit and his team, is part of a larger research effort focused on identifying attack chains within cloud environments.
A critical vulnerability in Microsoft's Azure Automation service has been addressed, but remains a significant concern due to a default configuration that could have enabled cross-tenant identity takeovers. The issue, discovered by Shay Shavit and his team at Microsoft's Azure Networking Security Research team, stemmed from the default setting of Azure Automation accounts being publicly accessible. This allowed an attacker with access to their own Azure Automation account to breach the trust boundary and assume another tenant's automation identity, granting them access to sensitive configuration data and credentials stored in Azure Automation accounts. The vulnerability could also be used to create, change, or delete resources across an organization's cloud workloads.
Shavit tells Dark Reading that while no known exploits currently exist for this vulnerability, he plans to demonstrate the attack chain at Black Hat USA in Las Vegas. The vulnerability is part of a larger research effort focused on identifying attack chains within cloud environments, and it consists of three distinct flaws: the default configuration making Azure Automation accounts publicly accessible, and two separate code-level bugs.
Previously, in 2021, researchers at Orca Security discovered a related flaw, dubbed AutoWarp, which exploited a code-execution component to establish a reverse shell and exfiltrate managed identity tokens from a shared sandbox server. While AutoWarp received a CVE identifier and was publicly disclosed by Microsoft in March 2022, Shavit’s discovery doesn’t involve code execution.
Despite the differences between the two vulnerabilities, both share the same goal – extracting a victim’s managed identity. Shavit recommends that organizations avoid evaluating vulnerabilities in isolation and instead map the paths that connect them, emphasizing the importance of understanding the entire attack chain. He suggests auditing the scope of identities and tokens assigned to cloud automation accounts and avoiding unnecessary exposure of identities.
Black Hat USA Aug 1, 2026 TO Aug 6, 2026 |Mandalay Bay Convention Center, Las Vegas, USA