CVE-2026-72649: CWE-502 Deserialization of Untrusted Data in Elastic Elasticsearch

This vulnerability (CWE-502) in Elasticsearch's machine learning component arises from unsafe deserialization of untrusted data, enabling remote code execution via object injection (CAPEC-586). An attacker with authenticated access and adequate privileges can craft a trained model artifact that executes attacker-controlled logic with an expanded system-call surface. The vulnerability affects specific versions of Elasticsearch including 8.0.0, 9.0.0, and 9.5.0. No vendor advisory or patch information is currently provided, and no known exploits are reported in the wild.

Successful exploitation results in remote code execution with high confidentiality, integrity, and availability impact, allowing an attacker to execute arbitrary code on the affected system. However, exploitation requires authenticated access with sufficient privileges, limiting the attack surface to authorized users capable of deploying trained models.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix is available, restrict access to users with privileges to create and deploy trained models and monitor for suspicious activity related to model deployment.