CVE-2026-64859 in New API
Summary
by MITRE • 08/17/2026
New API is a large language mode (LLM) gateway and artificial intelligence (AI) asset management system. Prior to 1.0.0-rc.7, the admin user list and user lookup APIs, including GET /api/user/, return User.AccessToken as access_token because User model objects are serialized after queries use Omit("password"), allowing an authenticated administrator to obtain the root user's bearer token and access root-only system configuration APIs. This issue is fixed in version 1.0.0-rc.7.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 08/17/2026
The vulnerability identified in New API, a large language model gateway and artificial intelligence asset management system, represents a critical information disclosure flaw within its administrative interface. Specifically affecting versions prior to 1.0.0-rc.7, this issue stems from improper serialization of user data objects during the execution of admin-level queries. The core technical failure lies in the application's handling of User model instances when retrieving lists or performing lookups via endpoints such as GET /api/user/. Although the development team implemented a mitigation strategy by using an Omit function to exclude password fields from the serialized output, this approach was insufficiently comprehensive. It failed to account for other sensitive attributes embedded within the user object structure, most notably the User.AccessToken field which is exposed in the response payload as access_token.
This architectural oversight allows any authenticated administrator with valid credentials to retrieve the bearer tokens of all users listed by these endpoints. The severity escalates significantly because this capability extends to retrieving the token associated with the root or superuser account. In many system architectures, including New API's design, administrative APIs are often protected by role-based access controls that rely on the validity and privilege level of the provided bearer token. By obtaining the root user's active session token through these lookup endpoints, an attacker can bypass authentication checks entirely for any endpoint restricted to administrators or superusers. This effectively grants the attacker full control over the system configuration without needing to exploit a separate credential theft mechanism such as phishing or brute force attacks against password hashes.
The operational impact of this vulnerability is severe, encompassing both confidentiality and integrity compromises. An adversary leveraging this flaw can access sensitive system configurations, potentially altering AI model parameters, modifying asset management settings, or exposing proprietary data stored within the platform's backend systems. Furthermore, possession of a valid root bearer token allows for persistent unauthorized access as long as the token remains active, facilitating lateral movement if the gateway integrates with other internal services that trust tokens issued by New API. This scenario aligns closely with CWE-200, which categorizes exposure of sensitive information to an unauthorized actor, and specifically relates to CWE-798 regarding the use of hardcoded or easily guessable credentials when considering how token management failures can lead to privilege escalation. From a tactical perspective, this vulnerability supports ATT&CK technique T1530, Data from Cloud Storage Object Discovery, as it enables enumeration of high-value assets within the cloud-native environment managed by New API.
Mitigation for this issue requires immediate patching to version 1.0.0-rc.7 or later where the serialization logic has been corrected to exclude all sensitive fields including access tokens during administrative list operations. Beyond applying the vendor-provided fix, organizations should implement strict input validation and output encoding practices across all API endpoints that handle user data. It is also recommended to enforce short-lived token expiration policies for root accounts to limit the window of opportunity for exploitation even if a token is compromised. Additionally, implementing comprehensive logging and monitoring for administrative queries can help detect anomalous patterns such as frequent lookups targeting specific high-privilege users, providing an early warning system against active reconnaissance attempts leveraging this flaw.