New "LeakyLooker" Flaws in Google Looker Studio Could Enable Cross-Tenant SQL Queries

Why It Matters

LeakyLooker exposes enterprise data across Google Cloud projects, raising the stakes for cloud‑based BI security and compliance. Organizations using Looker Studio must reassess their data‑sharing configurations to prevent cross‑tenant breaches.

Key Takeaways

• Nine cross‑tenant flaws discovered in Looker Studio
• Vulnerabilities enable arbitrary SQL on BigQuery, Spanner
• Zero‑click injection possible via data connector credentials
• Google patched all issues after responsible disclosure
• No evidence of real‑world exploitation yet

Pulse Analysis

The LeakyLooker findings underscore how tightly integrated analytics platforms can become attack vectors when multi‑tenant isolation is weak. Tenable’s research revealed that Looker Studio’s report‑copying logic and connector handling allowed malicious actors to inject SQL without user interaction, effectively turning a visual dashboard into a remote database console. By leveraging stored credentials, hyperlink rendering, and timing oracles, attackers could traverse tenant boundaries and manipulate data across a wide array of Google Cloud services, a scenario that challenges the traditional view of BI tools as passive consumers of data.

For enterprises, the practical risk is profound. A compromised Looker Studio report could grant an adversary read, write, or delete privileges on critical datasets, jeopardizing regulatory compliance, intellectual property, and financial reporting. The ability to exfiltrate entire databases with a single click amplifies the threat, especially for organizations that publicly share dashboards or embed them in external portals. Security teams must now audit report permissions, enforce least‑privilege access for data connectors, and monitor for anomalous query patterns that could indicate abuse of the previously hidden attack surface.

Google’s swift remediation demonstrates the importance of coordinated vulnerability disclosure, yet the episode serves as a cautionary tale for cloud‑first businesses. Moving forward, vendors should embed stronger tenant isolation checks and credential vaulting mechanisms within analytics pipelines. Customers should adopt continuous security testing, such as red‑team exercises focused on data‑visualization layers, and integrate security‑as‑code policies that automatically revoke over‑permissive connectors. As cloud analytics mature, the industry must balance ease of data collaboration with rigorous safeguards to prevent cross‑tenant data leakage.