AI Threat AlertCVE-2026-44605: RPM: heap overflow via crafted NDB file causes DoS
MEDIUMRPM has a heap buffer overflow, caused by an integer overflow (CWE-190), when it parses a specially crafted NDB database file. The result is a local denial of service, with no confidentiality or integrity impact (CVSS 5.5, medium). This is mainly a base-image hygiene issue for AI platforms: the flaw ships inside many Red Hat OpenShift AI (rhoai) RHEL9 container images, including MLflow, KServe, KubeRay, model registry and training operator images. Exploitation likelihood looks low. There is no EPSS score, no public exploit, no Nuclei template, and it is not in CISA KEV. It still needs a local attacker and user interaction, so it is unlikely to be used against a remote AI service. Apply Red Hat advisory RHSA-2026:33507 through your normal image rebuild and patch cycle. Prioritize hosts and build pipelines that process untrusted RPM database files.
What is the risk?
Low to medium overall. CVSS 5.5 reflects a local attack vector, low complexity, no privileges and required user interaction, with impact limited to availability. There is no known exploitation: not in KEV, no public exploit, no scanner template, and no EPSS or SSVC data. The 695 downstream dependents and the 142 other CVEs tracked for this package point to a wide footprint, but the flaw is only reachable when RPM parses an attacker-controlled NDB database. Risk rises in CI/CD or image-build systems that handle untrusted RPM databases.
How does the attack unfold?
What systems are affected?
How severe is it?
What is the attack surface?
What should I do?
1 step1) Apply Red Hat advisory RHSA-2026:33507 and rebuild or pull updated rhoai images and RHEL9 base images. 2) Track fixed versions in your container registry and re-scan images for the rpm package. 3) Do not run RPM operations on untrusted or user-supplied NDB database files, and restrict who can place files in RPM database paths. 4) Run build and package-management jobs with minimal privileges and in isolated, ephemeral containers. 5) Watch for rpm crashes (segfaults, abnormal exits) in build and node logs as a detection signal.
How is it classified?
Which compliance frameworks are affected?
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2026-44605?
RPM has a heap buffer overflow, caused by an integer overflow (CWE-190), when it parses a specially crafted NDB database file. The result is a local denial of service, with no confidentiality or integrity impact (CVSS 5.5, medium). This is mainly a base-image hygiene issue for AI platforms: the flaw ships inside many Red Hat OpenShift AI (rhoai) RHEL9 container images, including MLflow, KServe, KubeRay, model registry and training operator images. Exploitation likelihood looks low. There is no EPSS score, no public exploit, no Nuclei template, and it is not in CISA KEV. It still needs a local attacker and user interaction, so it is unlikely to be used against a remote AI service. Apply Red Hat advisory RHSA-2026:33507 through your normal image rebuild and patch cycle. Prioritize hosts and build pipelines that process untrusted RPM database files.
Is CVE-2026-44605 actively exploited?
No confirmed active exploitation of CVE-2026-44605 has been reported, but organizations should still patch proactively.
How to fix CVE-2026-44605?
1) Apply Red Hat advisory RHSA-2026:33507 and rebuild or pull updated rhoai images and RHEL9 base images. 2) Track fixed versions in your container registry and re-scan images for the rpm package. 3) Do not run RPM operations on untrusted or user-supplied NDB database files, and restrict who can place files in RPM database paths. 4) Run build and package-management jobs with minimal privileges and in isolated, ephemeral containers. 5) Watch for rpm crashes (segfaults, abnormal exits) in build and node logs as a detection signal.
What systems are affected by CVE-2026-44605?
This vulnerability affects the following AI/ML architecture patterns: model serving, MLOps platforms, training pipelines, notebook and workbench environments, container build pipelines.
What is the CVSS score for CVE-2026-44605?
CVE-2026-44605 has a CVSS v3.1 base score of 5.5 (MEDIUM).
What is the AI security impact?
Affected AI Architectures
MITRE ATLAS Techniques
AML.T0010.001AI SoftwareAML.T0029Denial of AI ServiceCompliance Controls Affected
What are the technical details?
Original Advisory
A flaw was found in the RPM Package Manager (RPM). A local user could be affected by a heap buffer overflow vulnerability when processing a specially crafted NDB database file. This issue arises from an error in how RPM handles certain calculations during file parsing, leading to an incorrect memory allocation. An attacker could leverage this to cause a denial of service, making the system unavailable.
Exploitation Scenario
An attacker with local access, or one who persuades a user or automated job to process a malicious NDB database file, supplies a file with crafted size fields. RPM miscalculates the allocation size, causing a heap overflow and crash. In an AI context, this could interrupt a container image build or package update on a node that hosts model serving or notebook workloads, causing a short outage of the maintenance or build pipeline. It does not give the attacker control of the AI workload.
Weaknesses (CWE)
CWE-190 — Integer Overflow or Wraparound: The product performs a calculation that can produce an integer overflow or wraparound when the logic assumes that the resulting value will always be larger than the original value. This occurs when an integer value is incremented to a value that is too large to store in the associated representation. When this occurs, the value may become a very small or negative number.
- [Requirements] Ensure that all protocols are strictly defined, such that all out-of-bounds behavior can be identified simply, and require strict conformance to the protocol.
- [Requirements] Use a language that does not allow this weakness to occur or provides constructs that make this weakness easier to avoid. If possible, choose a language or compiler that performs automatic bounds checking.
Source: MITRE CWE corpus.
CVSS Vector
CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:HReferences
- access.redhat.com/errata/RHSA-2026:33507vendor-advisoryx_refsource_REDHAT
- access.redhat.com/security/cve/CVE-2026-44605vdb-entryx_refsource_REDHAT
- bugzilla.redhat.com/show_bug.cgiissue-trackingx_refsource_REDHAT
Timeline
Related Vulnerabilities
CVE-2025-1537910.0MLflow: RCE via unsanitized model dependency specs
Same package: mlflowCVE-2023-376510.0MLflow: path traversal allows arbitrary file read
Same package: mlflowCVE-2023-11779.8MLflow: path traversal allows arbitrary file read/write
Same package: mlflowCVE-2026-26359.8mlflow: security flaw enables exploitation
Same package: mlflowCVE-2023-27809.8MLflow: path traversal allows arbitrary file read/write
Same package: mlflow