LeakyLooker Vulnerabilities Expose Google Looker Studio Data

Tenable researchers have identified a series of security vulnerabilities in Google Looker Studio that could allow attackers to execute arbitrary SQL queries and extract sensitive data from victims’ databases within Google Cloud environments.

The vulnerabilities, collectively named LeakyLooker, include nine cross-tenant security flaws that could potentially expose data across organisations using the analytics platform. According to the research findings, the weaknesses affected multiple connectors commonly used in cloud analytics deployments.

These connectors include services such as Google Sheets, BigQuery, Cloud Spanner, PostgreSQL, MySQL, and Google Cloud Storage.

Live Data Architecture Creates New Attack Surface

Looker Studio is designed to provide live analytics by allowing users to connect dashboards directly to multiple data sources. This flexibility enables real-time updates but also introduces architectural challenges in maintaining strict isolation between users and data environments.

Tenable researchers found that the platform’s “Live Data” architecture created an exploitable pathway. Attackers could potentially exploit the vulnerabilities using either zero-click attacks, which require no user interaction, or one-click attacks, which only require a victim to open a malicious web page controlled by the attacker.

According to Liv Matan, Senior Research Engineer at Tenable, the vulnerabilities compromised a fundamental expectation of the platform: that users with viewer-level permissions should not be able to manipulate or control underlying data sources.

Sticky Credential Flaw Enables Unauthorized Access

One of the vulnerabilities involved a logic flaw referred to as “Sticky Credential.” This issue was identified in the platform’s Copy Report feature.

The flaw allowed unauthorized users to clone reports while retaining the original owner’s stored credentials. As a result, attackers could potentially gain access to connected databases and perform actions such as modifying or deleting tables.

This type of flaw demonstrates how credential handling within shared reporting environments can create unintended security exposures.

Data Exfiltration Through Malicious Reports

Another attack path involved a one-click data exfiltration technique. In this scenario, an attacker could share a specially crafted report with a target.

When the victim opened the report, their browser could be forced to execute malicious code that interacted with an attacker-controlled project. The attacker could then reconstruct database contents from logs generated by these interactions.

This method allowed the extraction of large datasets from connected cloud databases without directly accessing the systems themselves.

Nine Vulnerabilities Identified

The LeakyLooker research identified nine separate cross-tenant vulnerabilities affecting Looker Studio environments. These included:

• Zero-click SQL injection vulnerabilities on database connectors

• SQL injection through stored credentials

• SQL injection paths affecting BigQuery and Spanner

• Data source leaks through hyperlinks and image rendering

• Cross-tenant data exposure using frame counting and timing methods

• A denial-of-wallet scenario targeting BigQuery usage

Together, these flaws created multiple potential pathways for attackers to access or manipulate data across cloud tenants.

Patches Released Following Responsible Disclosure

Following responsible disclosure by Tenable researchers, Google addressed all nine vulnerabilities globally.

The remediation included security fixes designed to close the identified attack vectors and strengthen isolation between tenants within Looker Studio environments.

Security Recommendations for Organizations

To reduce the risk of similar exposures, security experts recommend that organisations review and limit access to analytics reports and dashboards. Special attention should be given to users with viewer-level permissions in both public and private reports.

Organizations are also advised to treat business intelligence connectors as critical entry points to cloud infrastructure and regularly revoke access to unused connectors or services.

The LeakyLooker findings highlight how analytics platforms integrated with multiple data sources can introduce new security risks when live data connections and credential sharing mechanisms are not fully isolated.