OpenCode Security Breach Exposes Third-Party API Keys of its Console Users

OpenCode has disclosed a security breach affecting 1.4% of its Console users, warning that third-party API keys saved in affected workspaces may have been exposed. In an October 17 statement, the company linked the incident to a large-scale distillation attack and said it was working with Anthropic, OpenAI, and Google to address it.

The affected group was limited to some Console users who had linked external providers. OpenCode said it had emailed everyone affected that morning and stated that users who did not receive an email were not impacted. The percentage refers to Console users, not the entire OpenCode user base.

Legacy System Access Put Provider Keys at Risk

A notification shared in the supplied screenshot says OpenCode identified unauthorized access to a legacy system that it was already preparing to shut down. Third-party provider keys stored in users’ workspaces, including those for OpenAI and Anthropic, may have been exposed.

The notice asks recipients to revoke and replace those keys through each provider’s dashboard, update applications that use them, and review recent usage for unexpected activity. Its wording describes possible exposure rather than confirming that every affected key was stolen or used.

OpenCode’s public statement does not explain how attackers gained access, how long they remained inside, or how many keys were involved. It also does not name an attacker or provide technical evidence showing how the breach supported the reported distillation activity.

Distillation involves training one AI model using responses generated by another. In an unauthorized campaign, attackers can collect large amounts of model output without permission. Cybersecurity News previously covered Anthropic’s warnings about large scale distillation attacks, explaining how repeated requests can help transfer capabilities between models.

However, OpenCode has not explained whether exposed credentials were used to make those requests. Its statement establishes a reported connection, not a complete account of the attack. There is also no confirmed loss figure in the public disclosure.

The wider risk of exposed credentials has appeared in earlier reporting on API keys found in public web data. That separate case illustrates why secret storage and removal matter, but it does not establish a shared attacker or method here.

Affected Users Urged to Replace Keys

For notified users, the immediate response is the key replacement process described in the warning. Reviewing provider usage alongside recent billing records can help identify requests that need investigation. Replacing a key also requires updating every application that depends on it.

OpenCode said the incident remained ongoing when it published its statement and that it could not share more details. A fuller technical report would help establish the entry point, the scope of credential exposure, and whether unauthorized model requests occurred.

An EDR alert reveals an attack. But can it tell the full story? Discover how to reconstruct the attack timeline.