A high-severity SQL injection vulnerability in Spring AI allows authenticated attackers to execute unauthorized database commands, posing a significant risk to data integrity and system confidentiality.

The flaw exists within the MariaDBFilterExpressionConverter due to insufficient input sanitization. An attacker with low-level privileges can manipulate filter expressions to perform SQL injection, enabling unauthorized interaction with the underlying MariaDB instance.

Successful exploitation allows an attacker to bypass metadata-based access controls, potentially leading to unauthorized data exfiltration, modification, or deletion. Given the CVSS score of 8.8, this vulnerability represents a high risk to the business, as it grants attackers the ability to execute arbitrary database queries, which may compromise the integrity and availability of the entire application backend.

Immediate Action: Update the affected Spring AI components to version 1.0.4 or 1.1.3 immediately to incorporate the necessary input sanitization fixes.

Proactive Monitoring: Review database audit logs for anomalous or unauthorized SQL queries that deviate from standard application-generated traffic patterns.

Compensating Controls: Deploy or update Web Application Firewall (WAF) rules to detect and block common SQL injection patterns targeting the application's search or filtering parameters.

Public Exploit Available: Yes, a public proof-of-concept exists via the GitHub repository: https://github.com/NULL200OK/CVE-2026-22730-Scanner.

The severity of this vulnerability necessitates an immediate response. Organizations utilizing Spring AI must prioritize the transition to the corrected versions (1.0.4 or 1.1.3) to prevent potential database compromise. Until updates are applied, ensure strict access control policies are enforced for all users who interact with the application's filtering features.