OpenCode Discovers 1.4% Distillation Attack Hit Console Users

Precedent & Market Context

OpenCode's security issues predate this incident, with problems surfacing as recently as early 2026. The latest breach adds to that track record at an awkward time for the platform.

Main Event

OpenCode made the breach public on October 17, 2026, disclosing that attackers exploited vulnerabilities in a legacy system to gain unauthorized access to API keys stored in user workspaces. Those keys were connected to models from Anthropic, OpenAI, and Google. With the keys in hand, the attackers ran repeated queries against high-end AI models.

Analysis & Market Impact

The term "distillation" describes a technique where one AI model learns by studying the answers of another. In this case, the attack appears designed to harvest high-quality responses from top-tier models. Those outputs can then be used to train competing models, without the attacker paying for the compute or the original research.

Implications & Outlook

For affected users, the immediate priority is containment. Anyone who connected provider accounts to the OpenCode Console should rotate or revoke those keys and review recent usage on each provider's dashboard. The incident also points to a new category of AI-specific breach. Anthropic's February 2026 disclosure suggests the labs are already watching for distillation behavior closely. Restricting direct access is one defense, but this breach shows its limits. When the front door locks, the pressure moves to every intermediary holding a key.