CVE-2026-91510: Unauthenticated Pre-Auth Compromise in Oracle Access Manager Authentication Engine (CVSS 10.0)
What the vulnerability is
CVE-2026-91510 is a critical, unauthenticated, network-exploitable flaw in the Authentication Engine component of Oracle Access Manager (OAM). An attacker with no credentials and network access to the OAM authentication endpoints can fully compromise the OAM authentication service — effectively taking over the identity and access management layer that other applications rely on to make authentication decisions. With a CVSS score of 10.0, this is as severe as the scoring model allows: no privileges required, no user interaction, and full impact on confidentiality, integrity, and availability.
Affected software
The vulnerability affects the Authentication Engine within Oracle Access Manager, which ships as part of Oracle Fusion Middleware 12.x. The detection record does not list specific vendor/product/version entries beyond this, so organizations running OAM on Fusion Middleware 12.x should treat this as in scope and consult Oracle's advisory for exact affected releases and patch availability.
Exploitation status
Exploit status is currently theoretical — there is no confirmed public proof-of-concept or evidence of in-the-wild exploitation at this time. However, given the unauthenticated, pre-auth nature of the flaw and its maximum severity score, defenders should not wait for weaponization before deploying detection and hardening measures. Theoretical status can change quickly once a working exploit surfaces.
How our detection catches it
Because no confirmed exploit chain exists yet, the detection logic targets the vulnerable attack surface and anomalous OAM access patterns rather than a specific payload signature. At a high level, it watches for:
- Anomalous or malformed requests to OAM authentication endpoints, including
/oam/server/,/oam/CredCollectServlet, and/oamfed/ - Crafted or malformed authentication tokens presented to these endpoints
- Unexpected process or child-process spawning originating from the WebLogic/OAM managed server, which would indicate post-compromise activity following a successful authentication-engine bypass
This logic is shipped across seven SIEM dialects — Microsoft Sentinel (KQL), Splunk (SPL), Elastic (EQL), IBM QRadar (AQL), Sumo Logic, Google Chronicle (YARA-L), and CrowdStrike (CQL) — so teams can deploy consistent coverage regardless of their stack, and tune thresholds as real-world exploitation data emerges.
Get the full detection
For the complete, ready-to-deploy queries, investigation guidance, and atomic tests across all supported SIEM platforms, see the full detection page for CVE-2026-91510.