OpenCode has disclosed a security breach that may have exposed third-party AI API keys saved by users of its hosted Console service.
The company said the incident affected 1.4% of Console users who had connected external AI providers, including OpenAI, Anthropic and Google.
The intrusion occurred on the night of October 16, according to OpenCode. On October 17, the company emailed affected users with the subject line “Action Required: Potential provider key breach,” advising them to revoke and replace their API keys. OpenCode also said that users who did not receive the notification were not affected, according to its public statement. <Cite refs={[“turn454618search0″,”turn454618search5”]}/>
OpenCode Console is an optional hosted service that allows teams to connect their own AI provider accounts. Workspace administrators can save provider API keys so team members can access AI models without seeing the underlying credentials themselves. The breach involved keys stored in a legacy Console system that OpenCode was preparing to decommission.
OpenCode co-founder Dax Raad clarified that the incident affected Console, not the OpenCode application itself. API keys configured directly in local command-line or desktop applications are stored on users’ machines, not submitted to the hosted Console service. <Cite refs={[“turn454618search0″,”turn454618search5”]}/>
The exposure could have financial consequences because AI provider API keys can authorize billable requests. If an attacker obtains a valid key, they may be able to use the associated account until the provider revokes the credential or it expires. Removing a key from a Console workspace does not necessarily invalidate the original credential.
Some users have reported unexpected charges following the incident. The Hacker News reported claims from users who said their OpenAI, Anthropic and Google accounts had incurred unauthorized usage. These reports remain individual claims, and OpenCode has not confirmed that every potentially exposed key was copied or used. <Cite refs={[“turn454618search0″,”turn454618search5”]}/>
Affected users should revoke their potentially exposed keys through the relevant provider’s dashboard and generate replacements. They should also update any applications or scripts that depend on the old credentials and examine their usage records and billing history for unfamiliar activity, particularly from October 16 onward.
OpenCode said the breach was related to what it described as a “large-scale distillation attack” and that it was working with OpenAI, Anthropic and Google to address the incident. In AI development, model distillation involves using the outputs of one model to help train another. However, OpenCode has not publicly explained how the intrusion was connected to such an operation.
The company has not disclosed the precise entry point, the number of API keys that may have been exposed or the identity of the attacker. The incident was still ongoing in OpenCode’s October 17 statement, leaving the full impact unclear. <Cite refs={[“turn454618search0″,”turn454618search3”]}/>
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
Sources
- OpenCode’s security announcement on X
- OpenCode co-founder Dax Raad’s clarification
- OpenCode provider documentation
