CVE-2025-10725

CRITICAL

Red Hat Openshift AI Service - Privilege Escalation

Title source: llm

Exploitation Summary

EIP tracks 1 public exploit for CVE-2025-10725. PoCs published by kaiokmo.

AI-analyzed exploit summary This repository demonstrates CVE-2025-10725, an RBAC misconfiguration in OpenShift AI's Kueue integration that allows low-privilege users to create Jobs in protected namespaces, enabling privilege escalation via token exfiltration from high-privilege ServiceAccounts. The PoC includes exploit, detection, and remediation components.

Description

A flaw was found in Red Hat Openshift AI Service. A low-privileged attacker with access to an authenticated account, for example as a data scientist using a standard Jupyter notebook, can escalate their privileges to a full cluster administrator. This allows for the complete compromise of the cluster's confidentiality, integrity, and availability. The attacker can steal sensitive data, disrupt all services, and take control of the underlying infrastructure, leading to a total breach of the platform and all applications hosted on it.

Exploits (1)

github WORKING POC

by kaiokmo · poc

https://github.com/kaiokmo/cve-2025-10725-eda-demo

View Code ZIP pw:eip

This repository demonstrates CVE-2025-10725, an RBAC misconfiguration in OpenShift AI's Kueue integration that allows low-privilege users to create Jobs in protected namespaces, enabling privilege escalation via token exfiltration from high-privilege ServiceAccounts. The PoC includes exploit, detection, and remediation components.

Classification

Working Poc 98%

Attack Type

Auth Bypass

Complexity

Moderate

Reliability

Reliable

Target: OpenShift AI with Kueue integration (specific version not stated)

Auth required

Prerequisites: Cluster must have OpenShift AI's Kueue integration with the vulnerable RBAC configuration · Attacker must have authenticated access to the cluster (system:authenticated group membership) · Target namespace must contain high-privilege ServiceAccounts

MITRE ATT&CK

T1190 - Exploit Public-Facing Application T1078 - Valid Accounts

mistral-large-3 · analyzed Jul 22, 2026 Full analysis →