CVE-2026-66007Medium· 6.5▾ TwilightPoC availableA flaw was found in datasets. This path traversal vulnerability allows a remote attacker to read arbitrary local files. By providing specially crafted file names in the metadata, an attacker can trick the system into including sensitive lo…
▾ Twilight zone — High severity, or a signal on a lesser flaw
impact 35.8 · likelihood 0.1 · exploitation 12
A public proof-of-concept already exists for this vulnerability — see Exploit availability below.
Stakeholder-Specific Vulnerability Categorization from CISA's ADP record at CVE.org: whether exploitation is observed, whether an attack can be automated, and how much of the system is at stake. The CVSS score shown above comes from the vendor's CSAF advisory record, not NVD.
Exploit-prediction probability, daily snapshots since Aug 19.
Disclosure to exploitation, from the record and what we observed since indexing it.
Disclosed via CSAF
0.5%
— → 6.5
none → medium
0.5% → 0.8%
Exploit / PoC code exists
Last analysed / modified upstream
A flaw was found in datasets. This path traversal vulnerability allows a remote attacker to read arbitrary local files. By providing specially crafted file names in the metadata, an attacker can trick the system into including sensitive local files when datasets are saved or published, leading to information disclosure.
datasets: Datasets: Information disclosure via path traversal vulnerability — rated Moderate by Red Hat. Released 2026-07-24, updated 2026-10-19.
Affected:
Fixed:
No fix planned:
Not affected:
For more information visit https://access.redhat.com/errata/RHSA-2026:82371 https://access.redhat.com/errata/RHSA-2026:82371 For Red Hat OpenShift AI 3.4.5 see the following documentation, which will be updated shortly for this release, for important instructions on how to upgrade your cluster and fully apply this errata update:
https://docs.redhat.com/en/documentation/red_hat_openshift_ai/ https://access.redhat.com/errata/RHSA-2026:84734
Workarounds / mitigations:
file_name metadata processed by the datasets library, especially within folder-based dataset builders, is thoroughly validated and sanitized. Avoid processing file_name values from untrusted sources without strict input validation to prevent directory traversal sequences. This will prevent the embedding of arbitrary local files during save_to_disk or push_to_hub operations.Affected packages:
datasets < 5.0.1Patched in:
datasets 5.0.1Field changes observed since this record was first indexed.
Connected by shared product, vendor, weakness, or advisory.
CVE-2026-76222High· 8.2gitpython: GitPython: Arbitrary file creation via path traversal in .gitmodules submodule names (CVE-2026-76222)
CVE-2026-79674High· 7.5nltk: NLTK: Information disclosure via path traversal in corpus-reader constructors (CVE-2026-79674)
CVE-2026-12243High· 7.5Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority.
CVE-2026-81726High· 8.7nltk: NLTK: Unauthorized file access via path traversal in model-artifact APIs (CVE-2026-81726)
CVE-2026-19693High· 8.1extract-zip: extract-zip: Arbitrary file write via symlink in archive (CVE-2026-19693)
CVE-2026-81725Medium· 5.9nltk: NLTK: Regular Expression Denial of Service via malformed TEI blocks (CVE-2026-81725)