Vercel: Third-party AI tool breached, leading to unauthorized access to internal systems; no sensitive data has been tampered with so far

Vercel announced a security incident analysis, stating that some of its internal systems were accessed without authorization. This was caused by a third-party AI tool, Context.ai, used by an employee being compromised, allowing the attacker to take over their Google Workspace account and access some environment configuration data.

The initial impact is that a small number of environment variables not marked as "sensitive" (such as API Keys, Tokens, etc.) may have been leaked. Relevant users have been notified and advised to rotate their credentials immediately. There is currently no evidence that data marked as "sensitive" or the supply chain (such as npm packages) has been tampered with.

Vercel stated that the attacker possesses a high level of technical skill and has partnered with Mandiant and several security agencies to investigate, and has reported the incident to law enforcement. They also emphasized that platform services are still operating normally. Additionally, users are advised to enable multi-factor authentication, comprehensively rotate potentially leaked environment variables, and check account activity logs and deployment records to prevent further risks.