OpenCode says attackers got into a slice of user accounts and used linked AI keys in a "distillation attack"

OpenCode, the open-source AI coding agent, said on Saturday, Oct. 17, 2026, that it suffered a security breach overnight tied to a "large scale distillation attack" that affected 1.4% of OpenCode Console users who had linked outside AI providers.

The new target in AI security isn't your password, it's your API key, because a frontier-model key is a prepaid card for compute. A coding tool that stores keys for several labs is a one-stop shop for anyone trying to copy a model on someone else's dime. 1.4% sounds small, but this is a tool with a very large developer user base, the incident is still "ongoing," and one of its founders says his own accounts are under attack. If you parked keys in a third-party agent dashboard, today is a good day to rotate them.

OpenCode's official account posted on Oct. 17, 2026: "We had a security breach last night related to a large scale distillation attack that impacted 1.4% of OpenCode Console users." OpenCode Console is the hosted dashboard where some users manage accounts and linked provider keys. Those lines are OpenCode's.

OpenCode said it notified all affected users by email that morning. Users who did not get an email were not affected. The affected group, it said, was a subset of users "who had external providers linked." Those lines are OpenCode's.

OpenCode said: "The incident is ongoing so we cannot share more details at this point but we are working with Anthropic, OpenAI, and Google to remediate." That quotation is OpenCode's.

OpenCode co-founder Dax Raad posted afterward: "my personal accounts are being attacked now too / whole new world with ai powered hacking." That quotation is his.

OfficeChai, quoting an email OpenCode sent to affected customers, reported that the company said it had "identified unauthorized access to a legacy system" that was being decommissioned, and that "third party provider keys (eg OpenAI, Anthropic) saved in your workspace may have been exposed." Affected users were asked to revoke and replace those keys with each provider and check recent usage. Those email details are OfficeChai's report of OpenCode's mail.

OpenCode is an open-source, provider-agnostic AI coding agent — it runs in the terminal, on the desktop and in an IDE — built by the team behind the SST framework. Users can plug in their own API keys from many model providers. An API key is a secret that lets software spend against a lab's models.

OpenCode did not define "distillation attack." In the industry the term usually means hammering a strong model with queries to harvest its outputs for training another model. Stolen API keys let an attacker do that on someone else's bill.

If you linked provider keys in OpenCode Console, rotate them in the OpenAI, Anthropic and Google dashboards and check usage, especially if you got the email. The statement is about OpenCode Console accounts and a legacy system. It is not a claim that the open-source coding agent itself was compromised. OpenCode has not said who the attackers were, how many keys were stolen, or how much money was lost.

In plain terms, OpenCode said on Saturday that attackers got into 1.4% of OpenCode Console accounts in what it called a large-scale distillation attack, that the hit users had outside AI providers linked, and that the incident is still ongoing. It says it is working with Anthropic, OpenAI and Google. OfficeChai's report of the customer email says a legacy system may have exposed saved provider keys.