Code execution via unfiltered dataset configuration parameters
Published Sep 22, 2026 · Updated Sep 22, 2026
Unsafe deserialization in NVIDIA NeMo Speech 2.9 and earlier allows local users to execute code via a crafted model_config.yaml. The hf_data_cfg mapping is passed unfiltered to datasets.load_dataset, so trust_remote_code can fetch and run arbitrary Python. Exploitation requires an attacker-controlled configuration to reach training-data setup, and the resulting code inherits the NeMo process's access.
Summary
What happened
Unsafe deserialization in NVIDIA NeMo Speech 2.9 and earlier allows local users to execute code via a crafted model_config.yaml. The hf_data_cfg mapping is passed unfiltered to datasets.load_dataset, so trust_remote_code can fetch and run arbitrary Python. Exploitation requires an attacker-controlled configuration to reach training-data setup, and the resulting code inherits the NeMo process's access.
The record
- CVE
- CVE-2026-65178
- Published
- Sep 22, 2026
- Updated
- Sep 22, 2026
- Vendor
- NVIDIA
- Product
- NVIDIA NeMo Speech
- Classifications
- CWE-502, T1059.006
- Attack vector
- local
- Privileges
- authenticated
Timeline
How it unfolded
- Sep 22, 2026CVE publishedPublication date reported by the CVE source.
- Sep 22, 2026Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=0.0 to 2.9
What conditions does exploitation require?
What is affected?
Published CVSS scores
CVSS describes severity. EPSS estimates exploitation probability.
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Activity data is delayed. The latest report is Sep 28, 2026.
Public exploit references
- Malicious hf_data_cfg model proof of conceptproof of concept · demonstrated
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Get started to see how Hinoki identifies affected software and validates exploitability in your environment.
Get started