OpenCode Credential Leak: Don’t Panic—First, Understand the Difference Between Console and the Coding Agent
OpenCode experienced a security incident, but one important distinction needs to be made clear: the incident affected some users of OpenCode Console, not everyone who uses the open-source OpenCode coding agent.
OpenCode Console is a standalone hosted model-calling service. It is not the same product as the CLI, TUI, or desktop coding tools people use day to day.
According to information disclosed later, about 1.4% of OpenCode Console users were affected. Third-party API keys stored in some affected workspaces may have been exposed, including keys for services such as OpenAI, Anthropic, and Google. The company also said it had notified affected users and released a fix.
There are three things to keep in mind:
First, don’t confuse the product names.
Using only the local OpenCode coding agent does not mean you used OpenCode Console. The similar names do not mean that all local users were affected by this incident.
Second, if you’ve used Console, check your credentials.
If you received an official security notice, follow the instructions to revoke and replace the relevant third-party API keys, update any applications that depend on those keys, and check recent usage logs and bills for anything unusual.
If a third party misuses your API key, the risk may extend beyond data security to unauthorized model calls and extra charges.
Third, this incident is a reminder that key management matters for AI tools, too.
Many people focus on how powerful a model is or how cheap its API is, while overlooking where keys are stored, which services can access them, and whether they can be revoked quickly if exposed.
When using a third-party hosted service, include its credential management and security practices in your own risk assessment.
Security incidents should not be exaggerated, but they should not be ignored either.
OpenCode experienced a security incident, but one important distinction needs to be made clear: the incident affected some users of OpenCode Console, not everyone who uses the open-source OpenCode coding agent.
OpenCode Console is a standalone hosted model-calling service. It is not the same product as the CLI, TUI, or desktop coding tools people use day to day.
According to information disclosed later, about 1.4% of OpenCode Console users were affected. Third-party API keys stored in some affected workspaces may have been exposed, including keys for services such as OpenAI, Anthropic, and Google. The company also said it had notified affected users and released a fix.
There are three things to keep in mind:
First, don’t confuse the product names.
Using only the local OpenCode coding agent does not mean you used OpenCode Console. The similar names do not mean that all local users were affected by this incident.
Second, if you’ve used Console, check your credentials.
If you received an official security notice, follow the instructions to revoke and replace the relevant third-party API keys, update any applications that depend on those keys, and check recent usage logs and bills for anything unusual.
If a third party misuses your API key, the risk may extend beyond data security to unauthorized model calls and extra charges.
Third, this incident is a reminder that key management matters for AI tools, too.
Many people focus on how powerful a model is or how cheap its API is, while overlooking where keys are stored, which services can access them, and whether they can be revoked quickly if exposed.
When using a third-party hosted service, include its credential management and security practices in your own risk assessment.
Security incidents should not be exaggerated, but they should not be ignored either.