Tenable Research recently uncovered “LeakyLooker,” a critical set of nine novel cross-tenant vulnerabilities within Google Looker Studio that enabled attackers to silently exfiltrate or modify sensitive data across various Google Cloud Platform services.
Following responsible disclosure by security researchers, Google has successfully patched all nine vulnerabilities globally, neutralizing the threat without requiring any manual updates from end users.
These vulnerabilities exposed highly sensitive data across numerous enterprise environments, potentially compromising any organization that actively utilizes Looker Studio data connectors.
The affected cloud integrations spanned a massive attack surface, including widely adopted services such as Google Sheets, BigQuery, Spanner, Cloud Storage, PostgreSQL, and MySQL databases.
By breaking the core isolation guarantees between different cloud tenants, malicious actors could leverage the visualization platform as a stealthy backdoor directly into the heart of a victim’s backend infrastructure.
Looker Studio dynamically generates charts by relying on live data connections governed by two distinct authentication models: Owner credentials, which use the report creator’s permissions, and Viewer credentials, which strictly require the viewer to possess their own database access.
Security researchers from tenable discovered that by decoupling and manipulating these two isolated mechanisms, they could forge entirely new attack paths that bypassed standard security checks.
This architectural oversight effectively transformed a business intelligence platform meant strictly for safe data visualization into a powerful weapon for cross-tenant database hijacking.
Zero-Click Database Exploits
The first attack path generated devastating zero-click vulnerabilities by targeting Owner credentials, notably through an “Alias Injection” flaw that allowed attackers to bypass Google’s security filters using SQL comment tags (/**/) to execute arbitrary backend commands.
Furthermore, a severe “Sticky Credential” vulnerability allowed unauthorized users to clone existing reports and inadvertently copy the hidden database credentials of JDBC-connected data sources into their newly created instance.
Because the attacker automatically became the owner of the cloned report, they gained full administrative privileges to read, write, update, or completely delete data from the victim’s database without ever needing the original password.
The second attack path leveraged one-click vulnerabilities against Viewer credentials by manipulating background network requests to secretly connect a seemingly harmless report to a victim’s active BigQuery project.
Attackers abused the platform’s native calculation features to hide a complex, multi-statement SQL injection script that immediately executed when the victim unknowingly clicked and opened the compromised report.
To successfully extract the stolen information, the script read the victim’s data character by character and pinged corresponding empty tables in the attacker’s environment, allowing the cybercriminals to reconstruct the entire database simply by monitoring their own cross-tenant access logs.
Remediation and Mitigations
Google has fully remediated the LeakyLooker vulnerabilities across all cloud environments, highlighting the severe and often overlooked risks associated with live-data integrations in modern business intelligence tools.
While direct patching is not required by customers, security teams must proactively audit who possesses viewing rights for both public and private reports to minimize future exposure.
Organizations are strongly advised to treat all data connectors as critical components of their external attack surface and immediately revoke Looker Studio’s access to any unused or legacy databases.
Follow us on Google News, LinkedIn, and X to Get Instant Updates and Set GBH as a Preferred Source in Google.
