“Related to a large-scale distillation attack that impacted 1.4% of OpenCode Console users,” OpenCode wrote — a concise admission that, in practice, means some teams’ centrally stored API keys for Anthropic, OpenAI and Google Gemini may have been exposed.
What OpenCode says happened and who was affected
On October 17, OpenCode confirmed a security breach that occurred the night before and affected its Console service. Console is the part of OpenCode that "allows teams to share access to AI models through centrally managed API credentials." Rather than exposing provider keys to individual team members, Console stores each workspace’s credentials and uses them to authenticate requests to the respective providers. OpenCode said the incident was “related to a large-scale distillation attack that impacted 1.4% of OpenCode Console users.”
The company added that the impact was limited to "a subset of users who had external providers linked." OpenCode notified the affected users by email; the subject line OpenCode used was "Action Required: Potential provider key breach." Copies of that email were made public on an OpenCode-dedicated subreddit forum and the project's official Discord channel.
Which credentials could be exposed and the scope inside the product
OpenCode’s notification warned that it had "identified unauthorized access to a legacy system that we were in the process of decommissioning" and that "third party provider keys (e.g. OpenAI, Anthropic) saved in your workspace may have been exposed." The company also notified Anthropic, OpenAI and Google to remediate the issue.
Dax Raad, one of OpenCode’s main contributors and a co-founder of Anomaly, posted on X that the breach involved only "the OpenCode Console, not OpenCode itself." Anomaly is the company that develops and maintains OpenCode; Y Combinator is named in reporting as a financial backer of Anomaly. Separately, OpenCode has 217 stars on GitHub and, per a July 2026 interview cited by the project, founder and CEO Jay V said OpenCode had reached approximately 13 million monthly active users and 4.6 million weekly active users.
Audit-ready is a season. It shouldn't be.
Evidence in spreadsheets, controls drifting between audits, frameworks multiplying on flat headcount. Nubivance runs continuous compliance on Rapid7 Cyber GRC - SOC 2, HIPAA, ISO 27001, PCI, CMMC.
End the scrambleSafeguard’s incident-response checklist for affected teams
Software supply chain security firm Safeguard published a set of recommended actions for teams using OpenCode Console. Key steps they listed include:
- Check whether you were affected by searching your inbox for the subject line “Action Required: Potential provider key breach.” If you received that email, follow its guidance; if not, review your keys as a precaution.
- Identify all potentially exposed keys: enumerate every OpenAI, Anthropic and Google Gemini API key ever stored in an OpenCode Console workspace, including keys added by other workspace admins.
- Review API usage for suspicious activity: check provider usage and billing reports from October 16 onward, and further back if possible for unexpected usage spikes, unfamiliar models or unexplained charges.
- Create replacement keys before revoking the old ones; use project- or workspace-specific keys with restrictions and spending limits.
- Revoke potentially exposed keys through the provider dashboards and verify that revoked keys return authentication errors rather than successful responses.
- Update all systems that used the old keys — CI/CD secrets, environment files, local agent configurations, production apps and team members’ machines — and ensure consumers use the new keys before removing dependencies on the old ones.
- Apply additional security controls: restrict keys to required projects and APIs, set spending limits and budget alerts, and review workspace access and permissions.
- If continuing to use Console, add a dedicated, restricted API key rather than reusing a production key.
Responses inside the OpenCode ecosystem: disclosure and remediation
OpenCode disclosed the incident publicly and sent direct notices to impacted users, while also informing the named providers. Users posted the notification email on OpenCode’s subreddit and in the official Discord, making the company’s outreach visible to the community. OpenCode and Anomaly have been contacted for comment, according to reporting.
How developers, providers and workspace administrators should react
Developers and workspace administrators: they should immediately enumerate any provider keys stored in Console, follow Safeguard’s checklist to create and roll keys, and search billing and usage reports for anomalies dating from October 16 onward.
Providers (Anthropic, OpenAI, Google): they were notified by OpenCode and will be the destination for any revocation and replacement workflows; account-level monitoring and alerting on unexpected model usage or billing spikes will be central to confirming whether exposed keys were abused.
The breach highlights a narrow but concrete operational risk: when a convenience layer centrally stores third‑party credentials, a failure in that layer — even a legacy component being retired — can widen the blast radius to multiple AI providers. OpenCode’s admission that a decommissioning legacy system was accessed and that remediation notices have already been sent frames the immediate task for teams: inventory, rotate, revoke and verify. The unanswered detail is how many of the affected keys showed malicious use; users and the notified providers will be watching billing and usage records closely in the days and weeks after October 16.
Original reporting: Infosecurity Magazine — OpenCode Breach Could Affect OpenAI, Anthropic and Google
