Researchers expose Looker flaws risking cross-tenant data

Cybersecurity researchers have uncovered a series of critical vulnerabilities affecting Google’s Looker data analytics platform that could have enabled attackers to execute malicious code, move between cloud environments and extract sensitive corporate data. The flaws highlight growing security concerns surrounding widely used cloud-based analytics tools that sit at the centre of many organisations’ data infrastructure.

Security specialists from the cybersecurity firm Tenable identified two major vulnerabilities—collectively dubbed “LookOut”—that could potentially allow attackers to take control of Looker servers or access internal databases containing sensitive operational information. Looker, a business intelligence and analytics platform owned by Google, is used by tens of thousands of organisations worldwide to visualise and analyse data through dashboards and automated reporting systems.

The most serious issue involves a chain of vulnerabilities that could lead to remote code execution. By exploiting weaknesses in how Looker processes certain project files and dependencies, an attacker with developer-level access could run arbitrary commands on the server hosting the platform. That level of control could allow intruders to steal authentication secrets, manipulate datasets or pivot deeper into internal networks connected to the system.

Researchers demonstrated that the exploit relied on manipulating LookML projects—the modelling framework used within Looker to define relationships between data sources. By crafting malicious dependencies and exploiting path-traversal behaviour in Git configuration files, the attack chain could override internal safeguards and trigger the execution of malicious scripts. A race condition within the system allowed the attackers to bypass protections intended to reset unsafe configuration changes before execution occurred.

The implications extend beyond individual corporate deployments. According to the researchers, the vulnerability could potentially enable cross-tenant access in cloud environments, meaning attackers might have been able to move between different organisations’ data environments hosted on the same infrastructure. Such a scenario is particularly concerning in multi-tenant cloud platforms where multiple customers share underlying computing resources.

A second flaw identified by Tenable involved an authorisation bypass that allowed access to an internal management database used by Looker to store metadata, user permissions and configuration details. By manipulating requests within the application interface, researchers found they could connect to the internal database and extract information through carefully crafted SQL error messages. Through repeated queries, attackers could potentially retrieve the entire contents of the database, including sensitive credentials and configuration secrets.

Looker plays a pivotal role in many enterprises because it acts as a central data layer connecting multiple databases and analytics tools. Companies rely on it to consolidate data from operational systems, cloud warehouses and third-party applications. Because of that central position, any compromise could give attackers broad visibility across corporate data ecosystems.

Security experts say the discovery illustrates how analytics platforms have become critical infrastructure within modern organisations. Business intelligence tools increasingly handle sensitive operational data such as financial records, customer behaviour metrics and supply-chain analytics. Breaches involving these platforms could therefore expose strategic information or enable manipulation of decision-making dashboards.

Google has addressed the vulnerabilities by releasing security updates across affected versions of the Looker platform. Cloud-hosted deployments managed by Google were patched automatically, while organisations running customer-hosted or on-premises versions must apply the updates themselves to mitigate risk. Security advisories recommend upgrading to patched versions released in late 2025 and early 2026 to prevent exploitation.

The company stated that it found no evidence indicating that the vulnerabilities had been exploited in active attacks. Nonetheless, the incident has prompted renewed scrutiny of cloud-based analytics systems that operate with extensive permissions across enterprise data environments.

Industry analysts note that security challenges in business intelligence platforms stem partly from their design. Tools like Looker must integrate with numerous databases and data pipelines while providing developers the flexibility to run complex queries and manage modelling code. Those capabilities increase the attack surface if input validation or access controls are not carefully implemented.

Enterprises using analytics platforms are being urged to follow several defensive practices, including limiting developer privileges, isolating analytics infrastructure from core systems and maintaining strict patch-management policies. Security professionals also recommend monitoring for unusual activity in analytics environments because attackers targeting data platforms often attempt to blend malicious queries with normal analytical workloads.

The discovery also reinforces the role of coordinated vulnerability disclosure programmes between technology providers and independent security researchers. Tenable reported the flaws through Google’s vulnerability reward framework, enabling patches to be developed before the details were publicly disclosed.

Cybersecurity specialists warn that vulnerabilities affecting data-analysis platforms may become more frequent as organisations expand their use of cloud analytics and artificial intelligence tools. Data warehouses, analytics dashboards and machine-learning pipelines increasingly form interconnected ecosystems, meaning weaknesses in one layer could expose a wide range of operational data.