OpenCode says breach hit 1.4% of Console users

OpenCode, the open-source AI coding agent, said Saturday that it had a security breach “related to a large scale distillation attack” that affected 1.4% of users of OpenCode Console, its hosted service, according to a post on X.

Who was hit: OpenCode said it emailed every affected user and that the breach hit only a subset who had external providers linked. Users can store API keys from AI providers including OpenAI, Anthropic and Google in Console, BleepingComputer reported. OpenCode told affected users the keys may have been exposed, and it said it is working with those three companies.

What users saw: Some users reported unauthorized charges ranging from several hundred dollars to as much as $1,700, and some said Google suspended their cloud projects, BleepingComputer reported. It said OpenCode’s message to affected users mentions possible reimbursement.

The defense: OpenCode co-founder Dax Raad said attackers had targeted the company’s systems for at least 48 hours using AI agents, according to BleepingComputer. Raad said the AI models OpenCode used to investigate “did quite a bit and then started refusing” mid-incident, in a reply on X. BleepingComputer reported that things stabilized after OpenCode began using OpenAI’s Daybreak cybersecurity model for defense.

What to do: OpenCode’s email told affected users to revoke those keys with each provider and replace them, The Hacker News reported.