CVE-2026-66007: Datasets Path Traversal via Unsanitized file_name Metadata
Vulnerability summary
CVE-2026-66007 is a medium (CVSS 6.9) vulnerability in huggingface datasets. EPSS estimates a 0.55% chance of exploitation in the next 30 days, higher than 46% of scored vulnerabilities. 2 public exploits are recorded.
Not severe (CVSS 6.9, network access); no exploitation reported by the 5 checked sources
Source freshness unavailable
How is this decided?Timeline
- 24-07-2026PublishedVulnerability first disclosed.
- 01-10-2026Last modifiedVulnerability information updated.
Description
Datasets through 5.0.0, fixed in commit f989ef9, contains a path traversal vulnerability in folder-based dataset builders where the file_name metadata field is not properly validated before being joined to the dataset directory. Attackers can supply crafted file_name values with directory traversal sequences to read arbitrary local files, which are then embedded into output when save_to_disk or push_to_hub is called.
CVSS metrics
7 scores · 5 sources · 2 versions- OSV · PypiEcosystem advisory · calculated6.9MediumCVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X
- OSV · WolfiEcosystem advisory · calculated6.9MediumCVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
- OSV · ChainguardEcosystem advisory · calculated6.9MediumCVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
- NVDNIST6.9MediumCVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X
- NVDNIST6.5MediumCVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:N
EPSS trend
Probability
0.55%
Chance of exploitation in the next 30 days
Percentile
46%
Versus all other CVEs
30-day change
Decrease -0.24 pp
From 0.79% on 04-10-2026
Known exploited
Compare KEV catalogs
CISA KEVNot in the CISA KEV catalog.
ENISA CNWNot in the ENISA CNW catalog.
CIRCLNot in the CIRCL catalog.
Public exploits
2 foundDatasets
Maturity: poc
Exploit Reference (vulncheck.com)
Maturity: unknown
Weaknesses
CWE · root causeAffected systems
Known conditions: network-reachable, no authentication, user interaction (CVSS vector). Other conditions: not stated in public data.
Chainguard8 products
- nemo-gymAll versions (No fix recorded) | < 0.5.0_rc0_git20260729-r2
- nemo-rl-cuda-13.0All versions (No fix recorded) | < 0.7.0-r4
- nemo-rl-cuda-13.0-automodelAll versions (No fix recorded) | < 0.7.0-r4
- nemo-rl-cuda-13.0-mcoreAll versions (No fix recorded) | < 0.7.0-r4
- nemo-rl-cuda-13.0-sglangAll versions (No fix recorded) | < 0.7.0-r4
Huggingface1 product
- datasets≤ 5.0.0
PyPI1 product
- datasets< 5.0.1
Wolfi1 product
- open-webui< 0.11.4-r4
4 vendors · 11 products
References
9 links- github.com/huggingface/datasets/issues/8324Technical DescriptionExploitIssue TrackingWEBREPORTVendor Advisory
- www.vulncheck.com/advisories/datasets-path-traversal-via-unsanitized-file-name-metadataThird Party AdvisoryWEBFIXAdvisoryExploitPatch
