CVE-2026-72642
HighCVSS 8.8Exploitation Probability (EPSS)
Low risk26th percentile - higher than 26% of all known CVEs
Summary
The native inference process in Elasticsearch for evaluating uploaded ML models accepts a model operation that computes a memory address from an offset without validating bounds. A user with privileges to upload and deploy models can craft a model that reads/writes memory outside the intended allocation.
Risk Assessment
The risk includes heap corruption causing crashes, and with sufficient heap control, potential arbitrary code execution in the inference process context.
Recommendation
Restrict model upload privileges and apply Elasticsearch patches for offset validation in models.
Original NVD description (English source)
The native inference process that Elasticsearch uses to evaluate uploaded machine learning models accepts a model operation that computes a memory address from an offset supplied inside the model, without validating that the offset stays within the bounds of the underlying storage. A user with the privileges required to upload and deploy a trained model can craft a model that reads and writes memory outside the intended allocation. The result is heap corruption that crashes the inference process, and, with sufficient control over the heap layout, could allow arbitrary code execution in the context of that process.