Vercel has disclosed a security incident stemming from an exploited OAuth grant, allowing unauthorized access to internal systems. The breach began when an employee integrated a third-party AI tool that had been compromised by an infostealer malware. Attackers used an unreviewed OAuth token from the affected vendor to bypass authentication safeguards. This event highlights vulnerabilities introduced by rapid adoption of AI tools without thorough evaluation of third-party security practices and permission scopes.
The attack circumvented standard security measures, complicating detection and response efforts. Vercel is working with industry partners to investigate the extent of the breach and confirm whether customer data was exposed. Experts stress the importance of strict vendor risk assessments, enforcing minimal access rights for OAuth tokens, and implementing ongoing monitoring. As digital ecosystems grow more interconnected, maintaining visibility and control over external integrations becomes critical to safeguarding enterprise environments against emerging threats.