Quantumnous
Monthly
Revoked API token enforcement is absent in QuantumNous new-api up to 1.0.0-rc.15, allowing a holder of a disabled token to query read-only endpoints such as /api/usage/token/ and retrieve usage log data. The root cause (CWE-613, Insufficient Session Expiration) lies in the TokenAuthReadOnly middleware in middleware/auth.go, which skipped the disabled-token status check entirely before the patch. A public POC is available; no active exploitation is confirmed in CISA KEV, and the CVSS 4.0 score of 2.1 reflects the narrow, low-impact attack surface.
Authorization bypass in QuantumNous new-api versions up to 0.12.1 allows remote attackers to access Midjourney image relay endpoints without proper authentication. The vulnerability resides in RelayMidjourneyImage and GetByOnlyMJId functions within relay-router.go. Despite high attack complexity (CVSS AC:H) and CVSS score of only 3.7, a publicly available proof-of-concept exploit exists (disclosed via GitHub Gist), reducing the technical barrier. The vendor did not respond to early disclosure attempts. EPSS data not provided, but the combination of public exploit and unauthenticated network access (PR:N) warrants attention for organizations using this API gateway for Midjourney integration.
SQL injection in QuantumNous new-api versions up to 0.12.1 allows authenticated remote attackers to manipulate database queries through the SearchUserTopUps and SearchAllTopUps functions in the self endpoint. The vulnerability exists in model/topup.go with confirmed public exploit code available on GitHub. With EPSS data unavailable and CVSS 6.3 (medium severity), the primary risk stems from the low-complexity exploitation requiring only low-level authentication, enabling attackers to exfiltrate sensitive data, modify records, or potentially execute denial-of-service attacks against the database layer.
A logic flaw in New API's universal secure verification flow allows authenticated users with registered passkeys to bypass WebAuthn assertion completion, effectively circumventing step-up authentication for privileged actions. This affects New API versions 0.10.0 and later, enabling authenticated attackers with passkey enrollment to access sensitive functionality without completing proper cryptographic verification. No patched versions are currently available, making this an unresolved authentication bypass affecting all current deployments.
An Insecure Direct Object Reference (IDOR) vulnerability exists in New API versions prior to 0.11.4-alpha.2, a large language model gateway and AI asset management system. Authenticated users can bypass authorization checks on the video proxy endpoint (GET /v1/videos/:task_id/content) to access video content belonging to other users and cause the server to authenticate to upstream AI providers (Google Gemini, OpenAI) using credentials derived from tasks they do not own. The vulnerability stems from a single unguarded function call that queries tasks by task_id alone without validating user ownership, contrasting sharply with all other task-lookup functions in the codebase that properly enforce ownership checks.
Revoked API token enforcement is absent in QuantumNous new-api up to 1.0.0-rc.15, allowing a holder of a disabled token to query read-only endpoints such as /api/usage/token/ and retrieve usage log data. The root cause (CWE-613, Insufficient Session Expiration) lies in the TokenAuthReadOnly middleware in middleware/auth.go, which skipped the disabled-token status check entirely before the patch. A public POC is available; no active exploitation is confirmed in CISA KEV, and the CVSS 4.0 score of 2.1 reflects the narrow, low-impact attack surface.
Authorization bypass in QuantumNous new-api versions up to 0.12.1 allows remote attackers to access Midjourney image relay endpoints without proper authentication. The vulnerability resides in RelayMidjourneyImage and GetByOnlyMJId functions within relay-router.go. Despite high attack complexity (CVSS AC:H) and CVSS score of only 3.7, a publicly available proof-of-concept exploit exists (disclosed via GitHub Gist), reducing the technical barrier. The vendor did not respond to early disclosure attempts. EPSS data not provided, but the combination of public exploit and unauthenticated network access (PR:N) warrants attention for organizations using this API gateway for Midjourney integration.
SQL injection in QuantumNous new-api versions up to 0.12.1 allows authenticated remote attackers to manipulate database queries through the SearchUserTopUps and SearchAllTopUps functions in the self endpoint. The vulnerability exists in model/topup.go with confirmed public exploit code available on GitHub. With EPSS data unavailable and CVSS 6.3 (medium severity), the primary risk stems from the low-complexity exploitation requiring only low-level authentication, enabling attackers to exfiltrate sensitive data, modify records, or potentially execute denial-of-service attacks against the database layer.
A logic flaw in New API's universal secure verification flow allows authenticated users with registered passkeys to bypass WebAuthn assertion completion, effectively circumventing step-up authentication for privileged actions. This affects New API versions 0.10.0 and later, enabling authenticated attackers with passkey enrollment to access sensitive functionality without completing proper cryptographic verification. No patched versions are currently available, making this an unresolved authentication bypass affecting all current deployments.
An Insecure Direct Object Reference (IDOR) vulnerability exists in New API versions prior to 0.11.4-alpha.2, a large language model gateway and AI asset management system. Authenticated users can bypass authorization checks on the video proxy endpoint (GET /v1/videos/:task_id/content) to access video content belonging to other users and cause the server to authenticate to upstream AI providers (Google Gemini, OpenAI) using credentials derived from tasks they do not own. The vulnerability stems from a single unguarded function call that queries tasks by task_id alone without validating user ownership, contrasting sharply with all other task-lookup functions in the codebase that properly enforce ownership checks.