QuantumNous new-api, an AI asset management and LLM gateway system, contains a vulnerability identified as CVE-2026-25591, classified under CWE-943 for improper neutralization of special elements in data query logic. The vulnerability affects versions prior to 0.10.8-alpha.10 and resides in the /api/token/search endpoint. This endpoint accepts user-supplied parameters 'keyword' and 'token' which are directly concatenated into SQL LIKE clauses without escaping wildcard characters such as '%' and '_'. Because these characters have special meanings in SQL LIKE queries, an attacker with valid authentication can inject complex wildcard patterns that cause the database to perform highly expensive pattern matching operations. This leads to resource exhaustion, resulting in denial of service (DoS) conditions. The vulnerability requires authentication but no additional user interaction. The CVSS v4.0 score is 7.1 (high severity), reflecting network attack vector, low attack complexity, no privileges beyond authentication, no user interaction, and a high impact on availability. The vendor patched this issue in version 0.10.8-alpha.10 by implementing proper escaping of wildcard characters in user inputs to prevent injection of malicious patterns. No public exploits have been reported yet, but the vulnerability poses a significant risk to availability for organizations using affected versions.

The primary impact of CVE-2026-25591 is denial of service through resource exhaustion of backend databases. Organizations running vulnerable versions of QuantumNous new-api may experience degraded performance or complete service outages when attackers exploit the SQL LIKE wildcard injection to trigger expensive queries. This can disrupt AI asset management workflows and LLM gateway operations, potentially halting critical AI-driven processes. Since the vulnerability requires authentication, insider threats or compromised credentials increase risk. The attack does not compromise data confidentiality or integrity but can severely impact system availability and operational continuity. Large-scale or repeated exploitation could strain infrastructure resources, increase operational costs, and damage organizational reputation. Given the growing adoption of AI management platforms, affected organizations worldwide could face significant operational disruptions if unpatched.

Organizations should immediately upgrade QuantumNous new-api to version 0.10.8-alpha.10 or later, where the vulnerability is patched. If immediate upgrading is not feasible, implement input validation and sanitization on the keyword and token parameters to escape or reject SQL wildcard characters ('%', '_') before they reach the database layer. Employ database query parameterization or prepared statements to avoid direct concatenation of user inputs into SQL queries. Monitor and restrict authenticated user access to the /api/token/search endpoint, applying the principle of least privilege. Implement rate limiting and anomaly detection on search requests to identify and block patterns indicative of resource exhaustion attacks. Regularly audit logs for unusual query patterns and resource usage spikes. Finally, ensure robust credential management and multi-factor authentication to reduce the risk of credential compromise that could enable exploitation.