OpenCode hit by hackers, AI safety filters hinder defense

opencode, an open-source ai coding agent, was breached overnight, affecting 1.4% of its online platform users. attackers used exposed api keys to access ai services, resulting in unauthorized charges ranging from hundreds to over $1,000. the incident highlighted issues with ai safety filters, which hindered opencode's ability to defend against the attack. anthropic, google, and openai responded by blocking exposed keys and refunding affected customers. the breach occurred as opencode made glм-5.3-flash free for users.