Grok-Built Apps Expose 35,000 Unauthenticated Entry Points

A study by security firm Claroc reveals that over 35,000 web apps built with xAI's Grok are publicly accessible without authentication, with nearly 95% of online apps exposing unauthenticated entry points. Leaked data includes internal sales materials from SpaceX, X, and Tesla, employee performance and bonus tables, ICU patient lab results and ventilator records, and even plaintext passwords. Claroc tested 43,551 Grok Build app addresses and found 94.7% of online apps could be opened directly, with about 650 to 800 apps containing private data, affecting six Fortune 500 companies and 26 enterprises with over 1,000 employees.

The root cause lies in Grok's share link design: random identifiers are placed before the domain rather than in the URL path, allowing links to be indexed by passive DNS services and publicly searchable. Developers relied on hard-to-guess URLs as access control instead of true authentication. Access control flaws in Grok-generated code and default public publishing settings further amplified the risk. Apps created by SpaceX and X employees have been taken down, but many third-party apps remain online, and it is unclear whether the platform has notified affected customers.

This incident, following the Meta Muse and Anthropic shared conversation leaks, once again exposes systemic risks in default security configurations of AI-generated apps. As vibe-coding tools rapidly proliferate, sensitive enterprise data is exposed without authentication, sounding an alarm for developers and platforms and highlighting the urgency of security design in AI programming tools.