CVE-2026-15044
MediumCVSS 6.3Exploitation Probability (EPSS)
Low risk7th percentile - higher than 7% of all known CVEs
Summary
A flaw in the TrustyAI Service Operator allows unauthenticated access to communication channels when deploying services like gorch or NemoGuardrails with a specific security setting disabled. This enables other programs within the cluster to access AI guardrails and orchestrator without proper authorization.
Risk Assessment
An attacker could gain unauthorized access to sensitive information and potentially make limited changes to AI models, posing a significant risk to system confidentiality and integrity.
Recommendation
Immediately enable the relevant security setting in the TrustyAI Service Operator and update it to the latest version to remediate the vulnerability.
Other vulnerabilities in TrustyAI Service Operator
See all- CVE-2026-15467High
A flaw was found in the trustyai-service-operator's LMEvalJob controller that allows an authenticated user within the cluster to configure a sidecar container to bypass existing security policies. This enables the execution of untrusted remote code, leading to arbitrary code execution within the cluster.
- CVE-2026-15063Medium
A flaw was found in the gorch service template, part of the trustyai-service-operator. Even when authentication is enabled, the gorch service exposes unproxied orchestrator and detector metrics ports, allowing any pod on the cluster network to directly access these ports, bypassing the kube-rbac-proxy and its authentication mechanisms.
Original NVD description (English source)
A flaw was found in the TrustyAI Service Operator. When deploying services like gorch or NemoGuardrails, if a specific security setting is not enabled, these services can expose their communication channels without requiring users to prove their identity. This allows any other program within the cluster to access the AI guardrails and orchestrator without proper authorization. An attacker could exploit this to gain unauthorized access to sensitive information and potentially make limited changes to the AI models.