Metabase disclosed a critical-severity SQL injection vulnerability exploited in the wild as a zero-day that permits unauthenticated remote attackers to inject arbitrary SQL queries into the application database. This allows attackers to escalate privileges to administrative level, change application settings, steal credentials for connected databases, and access or export data accessible through those connections. The vulnerability was discovered following exploitation attempts targeting Metabase Cloud. Metabase promptly blocked attack endpoints, patched the vulnerability, and updated cloud instances. Self-hosted users are urged to apply patches for versions 58.24, 59.21, 60.17, 61.11, 62.9, and 63.5 or block the /api/session/reset_password endpoint as a temporary workaround. Indicators of compromise include specific API call patterns in logs. Additional recommended post-compromise actions include revoking active sessions, reviewing and deleting unrecognized API keys, auditing administrative accounts, rotating database credentials, and monitoring logs for suspicious activity.

The vulnerability allows unauthenticated remote attackers to gain full administrative access to Metabase instances. This access enables attackers to alter application configurations, steal credentials for connected databases, and read or export any data accessible through those database connections. Such compromise can lead to significant data exposure and loss of control over the affected Metabase environment.

Metabase has released official patches for versions 58.24, 59.21, 60.17, 61.11, 62.9, and 63.5 that address this vulnerability. Metabase Cloud instances have already been updated and patched by the vendor. Self-hosted users should apply these patches immediately. Where patching is not possible, users should block the /api/session/reset_password endpoint to mitigate risk temporarily. If compromise is suspected, users should revoke all active user sessions, review and delete unrecognized API keys, audit administrative accounts, rotate credentials for all connected databases, and review logs for suspicious activity, particularly the specified API call patterns.