Multiple input validation flaws in Snowflake Spark Connector versions before 3.2.1 enable attackers to perform SQL injection attacks. These vulnerabilities allow exfiltration of OAuth credentials, execution of arbitrary SQL with the connector's Snowflake role, and redirection of COPY operations to attacker-controlled storage. Attack vectors include crafted OAuth token request URLs, malicious files in ingestion pipelines, SQL injection via staging options in shared Spark environments, and runtime SET commands that inject SQL into the SnowflakeFallbackCatalog option map. The injected SQL runs with cluster admin JDBC credentials, potentially leading to credential theft, unauthorized data access, and privilege escalation. The affected product is a cloud service, and Snowflake manages patching and remediation.

Successful exploitation can lead to theft of OAuth client credentials, unauthorized execution of arbitrary SQL commands with elevated privileges, unauthorized access to Snowflake account data, and privilege escalation within connected infrastructure. This poses a critical risk to confidentiality, integrity, and availability of data and services accessed via the Snowflake Spark Connector.

A patch is available for this vulnerability. Since Snowflake Spark Connector is a cloud service, Snowflake manages remediation server-side. Users should verify with Snowflake's official advisory and ensure they are using version 3.2.1 or later of the connector. No additional user action is required if the service is fully managed and updated by Snowflake.