OpenCode Key Leak Sparks Billing Fraud, Distillation Claim Met With Skepticism

Open-source AI coding tool OpenCode disclosed that its legacy system suffered unauthorized access, leaking plaintext API keys for approximately 1.4% of users. This breach triggered widespread billing fraud, with some victims facing losses up to $1,700 in a single day across major providers like Anthropic, OpenAI, and Google.

While OpenCode attributed the incident to a "model distillation attack," community scrutiny highlights fundamental security failures as the root cause. The compromised system stored keys in plaintext without rotation mechanisms, allowing attackers to bypass quotas and execute high-frequency abuse. Although cloud vendors have intervened to revoke keys and process refunds, delayed risk control responses underscore significant supply chain vulnerabilities.

The incident serves as a critical warning for the AI ecosystem, demonstrating how poorly secured AI applications can become vectors for automated attacks. It emphasizes the urgent need for stricter key management standards and real-time monitoring to prevent infrastructure flaws from cascading into severe financial and operational damages.