Anthropic has warned Claude users about a security breach in which hackers gained unauthorized access to subscriber accounts and stole API tokens to consume quota without authorization. The incident was first identified when a user noticed unexpected token depletion on an inactive account.

The breach highlights vulnerability in how API credentials are managed and secured in production environments. Anthropic has not disclosed the full scope of affected accounts or the mechanism by which tokens were compromised, but the company is actively investigating and working with affected users.

What This Means for Your Business

For enterprises using Claude API for production workloads, this incident underscores the importance of token rotation, IP whitelisting, and monitoring unusual consumption patterns. Teams should immediately audit their API key storage and access controls, implement automated alerts for abnormal usage, and consider segregating development and production tokens. This also serves as a reminder that managed API services, while convenient, require the same security rigor as self-hosted infrastructure.