QuantumNous new-api, a platform for managing AI assets and serving as a large language model gateway, contains a logic flaw in its universal secure verification flow starting from version 0.10.0 up to 0.11.9-alpha.1. The vulnerability (CVE-2026-32879) arises from improper authentication (CWE-287) where an authenticated user possessing a registered passkey can bypass the WebAuthn assertion step, which is intended to provide a strong second factor during step-up authentication for privileged actions. Instead of completing the WebAuthn challenge-response, the system incorrectly accepts the presence of the passkey alone as sufficient verification. This flaw undermines the security guarantees of WebAuthn, potentially allowing unauthorized privilege escalation within the system. The vulnerability requires the attacker to already be authenticated with a registered passkey, but no additional user interaction is needed to exploit the flaw. The CVSS 3.1 base score is 4.9, reflecting a medium severity with network attack vector, low attack complexity, high privileges required, no user interaction, and impact limited to confidentiality compromise. There are no known patches or mitigations released by the vendor at the time of publication. The recommended interim mitigations include avoiding reliance on passkey-based step-up verification for sensitive operations, enforcing TOTP or other 2FA methods for privileged actions, or restricting access to endpoints protected by the vulnerable verification flow. This vulnerability affects organizations that use QuantumNous new-api for AI and LLM management, potentially exposing sensitive data or assets to unauthorized access by insiders or compromised accounts.

The primary impact of CVE-2026-32879 is the compromise of confidentiality within affected QuantumNous new-api deployments. Attackers who have authenticated access with a registered passkey can bypass the intended WebAuthn assertion step, allowing them to perform privileged secure-verification actions without completing the strong second-factor authentication. This can lead to unauthorized access to sensitive AI assets, data leakage, and potential exposure of proprietary or confidential information managed by the platform. Although the vulnerability does not affect data integrity or system availability, the ability to bypass step-up authentication weakens the overall security posture and trust in the authentication mechanisms. Organizations relying on passkey-based step-up verification for critical operations may face increased risk of insider threats or lateral movement by attackers who have compromised user credentials. Since no patches are currently available, the risk remains until mitigations are applied or a vendor fix is released. The medium severity score reflects the requirement for prior authentication and the limited scope of impact, but the potential for sensitive data exposure makes this a significant concern for organizations managing AI assets and LLM gateways.

1. Do not rely solely on passkey-based step-up authentication for privileged or sensitive operations until a vendor patch is available. 2. Enforce additional multi-factor authentication methods such as TOTP or hardware-based 2FA tokens for step-up verification where operationally feasible. 3. Temporarily restrict or disable access to endpoints protected by the vulnerable secure-verification flow to minimize exposure. 4. Monitor authentication logs for unusual or suspicious activity involving passkey authentications and step-up verification attempts. 5. Implement strict access controls and least privilege principles to limit the impact of any compromised accounts. 6. Engage with QuantumNous support channels to track the release of patches and apply updates promptly once available. 7. Conduct internal security awareness training to highlight the risks of this vulnerability and encourage reporting of suspicious behavior. 8. Consider deploying compensating controls such as network segmentation or enhanced logging around critical AI asset management functions to detect and respond to potential exploitation attempts.