CVE-2026-49121
Description
A flaw was found in AI Tensor Engine for ROCm (AITER). This vulnerability allows unauthenticated remote attackers to execute arbitrary code by sending a specially crafted data package, known as a pickle payload, to a ZeroMQ (ZMQ) subscriber socket. This exploitation is possible due to a lack of authentication, message integrity checks (HMAC), or format validation in the MessageQueue.recv() function. Successful exploitation can lead to arbitrary code execution on every remote reader worker, posing a critical risk to the system's integrity and confidentiality.
Statement
Red Hat AI Inference Server and Red Hat OpenShift AI ship the AI Tensor Engine for ROCm (AITER) Python package as a dependency in ROCm-based vLLM container images. Affected streams embed AITER versions 0.1.5 through 0.1.10.post2 (published to PyPI under the distribution name amd-aiter, which is the same upstream ROCm/aiter project — its own setup.py sets PACKAGE_NAME = "amd-aiter"), all of which are within the upstream affected range (0.1.14 and earlier).
The flaw is an unauthenticated remote code execution vulnerability in AITER MessageQueue.recv() (shm_broadcast.py), where data received on a ZeroMQ subscriber socket is deserialized with Python pickle without authentication or integrity checks (CWE-502).
Red Hat rates this issue as Important. Our CVSS score reflects high attack complexity: exploitation requires network access to the inference worker ZMQ XPUB endpoint on the cluster network, or the ability to supply a forged distributed Handle with an attacker-controlled subscribe address. This is not a default remote attack against an unauthenticated internet-facing service.
Engineering trackers are filed for affected product streams. Updated container images will be released when a fixed AITER version is available and integrated.
Mitigation
Mitigate this issue by limiting network access to AITER/vLLM inference worker ZMQ endpoints to trusted cluster nodes only. Do not expose XPUB/subscribe ports outside the cluster network.
Where multi-node ROCm inference is not required, prefer single-node deployments that bind ZMQ to localhost where supported.
Update to a fixed AITER release (upstream fix expected in 0.1.15 or later) when provided in updated Red Hat AI Inference Server and Red Hat OpenShift AI container images. Refer to the errata or advisory linked from this CVE page when available.
Common Vulnerability Scoring System (CVSS) Score Details
Info alert:Important note
CVSS scores for open source components depend on vendor-specific factors (e.g. version or build chain). Therefore, Red Hat's score and impact rating can be different from NVD and other vendors. Red Hat remains the authoritative CVE Naming Authority (CNA) source for its products and services (see Red Hat classifications).
The following CVSS metrics and score provided are preliminary and subject to review.
CVSS v3 Score Breakdown
| Red Hat | NVD | cve.org | |
|---|---|---|---|
| Base Score | 8.1 | 9.8 | 8.1 |
| Attack Vector | Network | Network | Network |
| Attack Complexity | High | Low | High |
| Privileges Required | None | None | None |
| User Interaction | None | None | None |
| Scope | Unchanged | Unchanged | Unchanged |
| Confidentiality | High | High | High |
| Integrity Impact | High | High | High |
| Availability Impact | High | High | High |
Vector
Red Hat: CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H
NVD: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
cve.org: CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H
Understanding the Weakness (CWE)
Integrity
Technical Impact: Modify Application Data; Unexpected State
Attackers can modify unexpected objects or data that was assumed to be safe from modification. Deserialized data or code could be modified without using the provided accessor functions, or unexpected functions could be invoked.
Availability
Technical Impact: DoS: Resource Consumption (CPU)
If a function is making an assumption on when to terminate, based on a sentry in a string, it could easily never terminate.
Other
Technical Impact: Varies by Context
The consequences can vary widely, because it depends on which objects or methods are being deserialized, and how they are used. Making an assumption that the code in the deserialized object is valid is dangerous and can enable exploitation. One example is attackers using gadget chains to perform unauthorized actions, such as generating a shell.
Frequently Asked Questions
Not sure what something means? Check out our Security Glossary.
Want to get errata notifications? Sign up here.
