OpenCode breach may have exposed users’ AI-provider API keys

A breach at OpenCode may have exposed external AI-provider API keys linked to affected Console workspaces, according to a report published by Cyera on 17 October. Cyera says OpenCode confirmed the incident affected 1.4% of Console users; affected developers should revoke and replace linked keys, then check provider usage and bills.

Watch Desk analysis

What happened

Cyera says OpenCode attributed the breach to a legacy Console system it was retiring and emailed affected users. The report says keys linked to providers including OpenAI, Anthropic and Google may have been exposed. Some users have reported charges they did not recognise, but the report says OpenCode has not confirmed whether exposed keys were misused.

The report also says the method of access and how the keys were stored remain unknown. It attributes the phrase “large scale distillation attack” to OpenCode, but says no attacker has been publicly identified. Those unknowns matter: the breach is reported, while its cause and consequences are not yet fully established.

Why it matters

A provider API key can let someone run workloads on the account it belongs to, potentially creating charges. Depending on the key’s permissions, it may also provide access to data or models. Console budgets are not a substitute for provider-side controls: Cyera says they do not stop a stolen key being used directly with the provider.

Our read

If your team received an OpenCode breach notice, treat the linked provider keys as exposed: revoke and replace them in each provider’s own dashboard, then review usage and billing. Cyera advises checking provider activity from 16 October onwards and setting spend limits and alerts. If a key is shared with other tools or jobs, identify those dependencies so its replacement does not quietly break something important. Credential rotation is dull work, which is precisely why stolen credentials make such an attractive target.

What to watch

• Whether OpenCode clarifies how attackers gained access and how keys were stored.
• Whether it confirms misuse, affected services or further consequences.
• Whether affected users and providers report additional details or remediation steps.

Discussion spark: Should teams rotate every provider key linked to a breached service as a precaution, or wait until the service confirms which credentials were exposed?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.