cve.bz

CVE-2026-57845 Hugging Face Datasets < 5.2.0 Pickle Deserialization RCE via Pandas Builder

8.7 highEPSS 0.6%

High (8.7) in huggingface datasets. EPSS 0.6% estimated probability of exploitation activity in the next 30 days.

On this page
Hugging Face Datasets before 5.2.0 contains an insecure deserialization vulnerability that allows attackers to execute arbitrary Python code by crafting a malicious pickle file loaded through the deprecated pandas packaged builder. Attackers can supply a pickle-backed dataset that is processed by Pandas._generate_tables via pandas.read_pickle() without validation, so when a victim calls datasets.load_dataset("pandas", data_files=...), the embedded __reduce__ payload executes code with the victim process's privileges.

Severity

CVSS 4.0

8.7 high
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N

CVSS 3.1

8.8 high
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
EPSS
0.6% (unchanged from 0.5% a week ago) estimated probability of exploitation activity in the next 30 days; 48th percentile. Scored 2026-11-17. A forecast for this CVE across the internet, not for your own systems.
CISA's assessment (SSVC)
Exploitation: PoC · Automatable: no · Technical impact: total, assessed 2026-10-29. From CISA's Vulnrichment in the CVE List: whether it is exploited, whether exploiting it can be automated, and how much control it gives.
CWE
CWE-502 Deserialization of Untrusted Data
OWASP Top 10
A08:2025 Software or Data Integrity Failures
Published
Modified
Assigned by
VulnCheck (CNA)
NVD analysis
Awaiting analysis
Record from
the NVD, synced

Ask an AI about this CVE

Show the promptExplain CVE-2026-57845 for an engineer deciding whether to act. Use https://cve.bz/api/public/cves/CVE-2026-57845?via=ask-ai-copy and https://cve.bz/cve/CVE-2026-57845?via=ask-ai-copy as the source of truth: what is affected and which versions fix it, whether it is in CISA KEV, public exploits, and its EPSS (an internet-wide 30-day estimate of exploitation activity, not a forecast for my systems). If they give no fixed versions, use the CVE record https://cveawg.mitre.org/api/cve/CVE-2026-57845 or the vendor advisory, not third-party mirrors. Tell me how to check whether I am affected and how to verify a fix. Cite the sources and say where they are uncertain.

Opens a third-party assistant; its answer is its own, not cve.bz's. The prompt holds only this CVE's id and public links.

Affected

Versions

From the CNA's record in the CVE List.

ProductAffected
huggingface datasetsthrough 5.1.0

Red Hat

Red Hat rates this Important.

Red Hat's statement

This vulnerability is rated as Important severity because successful exploitation enables arbitrary code execution within the host or container environment, though it is constrained by requiring user interaction to ingest untrusted data files. In Red Hat environments such as Red Hat OpenShift AI (RHOAI) and RHEL AI, the datasets package is included across multiple machine learning images used for model training and evaluation. Risk is primarily present when workloads or interactive notebook users load third-party datasets that rely on the legacy pandas builder to deserialize pickle objects. Standard container boundaries and non-privileged workload deployment models help contain execution to the specific container context.

Red Hat's mitigation

--- CME Mitigation Guidance (2026-11-02) ---

Disclaimer: The following mitigation solutions are provided as a courtesy, pending the application of an official patch. This content was generated utilizing AI tools in conjunction with data from https://cmetaxonomy.org. Customers are advised to independently evaluate and test these temporary mitigations within their own infrastructure.

The highest-value action is to prevent execution of unvalidated pickle files by removing usage of the legacy pandas packaged builder in datasets and migrating to safe serialization formats like Parquet: upgrade to the fixed Rhoai/Odh Ta Lmes Job Rhel/Rhoai/Odh Caikit Tgis Serving Rhel/Rhoai/Odh Training Cuda121 Torch24 Py311 Rhel/Rhoai/Odh Training Rocm62 Torch24 Py311 Rhel/Rhoai/Odh Training Cuda124 Torch25 Py311 Rhel/Rhoai/Odh Training Rocm62 Torch25 Py311 Rhel/Rhoai/Odh Openvino Model Server Rhel/Exploit Intelligence/Vulnerability Analysis Rhel/Lscore Deploy/Rhoai/Odh Llama Stack Core Rhel/Rhoai/Odh Vllm Rocm Rhel/Rhoai/Odh Th06 Cpu Torch291 Py312 Rhel/Rhoai/Odh Th06 Cuda130 Torch291 Py312 Rhel/Rhoai/Odh Training Cuda128 Torch29 Py312 Rhel/Rhoai/Odh Training Rocm64 Torch29 Py312 Rhel/Rhoai/Odh Th06 Cpu Torch210 Py312 Rhel/Rhoai/Odh Th06 Cuda130 Torch210 Py312 Rhel/Rhoai/Odh Th06 Rocm64 Torch291 Py312 Rhel/Rhoai/Odh Workbench Jupyter Pytorch Llmcompressor Cuda Py312 Rhel/Rhelai3/Bootc Aws Cuda Rhel/Rhelai3/Bootc Azure Cuda Rhel/Rhelai3/Bootc Cuda Rhel/Rhelai3/Bootc Gcp Cuda Rhel/Rhoai/Odh Autorag Rhel/Rhelai3/Bootc Azure Rocm Rhel/Rhelai3/Bootc Rocm Rhel/Rhaii Fast/Vllm Spyre Rhel/Rhoai/Odh Trustyai Garak Lls Provider Dsp Rhel/Ask Sre/Lightspeed Core/Lightspeed Stack Rhel/Rhai/Guidellm Rhel/Rhaii/Model Opt Cuda Rhel/Rhaii/Vllm Rocm Rhel/Rhaii/Vllm Spyre Rhel/Rhaii Fast/Model Opt Cuda Rhel/Rhoai/Odh Feature Server Rhel/Rhoai/Odh Training Cuda128 Torch28 Py312 Rhel/Rhoai/Odh Training Rocm64 Torch28 Py312 Rhel/Rhoai/Odh Pipeline Runtime Pytorch Llmcompressor Cuda Py312 Rhel/Rhoai/Odh Th Torch Cpu Py312 Rhel/Rhoai/Odh Th Torch Cuda Py312 Rhel/Rhoai/Odh Th Torch Rocm Py312 Rhel/Rhoai/Odh Workbench Jupyter Trustyai Cpu Py312 Rhel/Rhaii Fast/Vllm Rocm Rhel/Rhoai/Odh Caikit Nlp Rhel/Python Datasets builds, which unconditionally raise a migration error in Pandas._generate_tables() before opening or unpickling any files, closing all 2 injection vectors (crafted pickle dataset file, untrusted pandas packaged builder input). Until patched, this is a concern on Workloads and interactive notebook users ingesting untrusted third-party datasets via the pandas builder across containerized and host environments, so Prohibit invoking datasets.load_dataset with the pandas builder on untrusted data files, and migrate workflows to safe formats such as Parquet, Arrow, CSV, or JSON, or direct construction via Dataset.from_pandas(). As a compensating measure, Enforce a read-only container root filesystem to prevent an executed payload from writing persistence or modifying binaries, Deploy an EDR agent to detect anomalous process spawning or malicious payload behavior launched via deserialization exploits, and Use Falco or eBPF runtime security to detect runtime anomalies such as unexpected shell execution or container escapes. These controls limit blast radius and contain exploitation, but they do not fix the underlying insecure pickle deserialization defect — apply the upstream patch to fully remediate.

Highest-impact CME: CME-1302 Deserialization Allowlist (Safe Loading) (high)

ProductPackageStateAdvisory
Exploit Intelligenceexploit-intelligence/vulnerability-analysis-rhel9Affected
Lightspeed Corelightspeed-core/lightspeed-stack-rhel9Affected
Red Hat AI Inference Serverrhaii-fast/model-opt-cuda-rhel9Affected
Red Hat AI Inference Serverrhaii-fast/vllm-rocm-rhel9Affected
Red Hat AI Inference Serverrhaii-fast/vllm-spyre-rhel9Affected
Red Hat AI Inference Serverrhaii/model-opt-cuda-rhel9Affected
Red Hat AI Inference Serverrhaii/vllm-rocm-rhel9Affected
Red Hat AI Inference Serverrhaii/vllm-spyre-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-aws-cuda-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-azure-cuda-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-azure-rocm-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-cuda-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-gcp-cuda-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-rocm-rhel9Affected
Red Hat OpenShift AI (RHOAI)rhai/guidellm-rhel9Affected
Red Hat OpenShift AI (RHOAI)rhoai/odh-autorag-rhel9Affected
Red Hat OpenShift AI (RHOAI)rhoai/odh-caikit-nlp-rhel9Affected
Red Hat OpenShift AI (RHOAI)rhoai/odh-caikit-tgis-serving-rhel9Affected
Red Hat OpenShift AI (RHOAI)rhoai/odh-feature-server-rhel9Affected
Red Hat OpenShift AI (RHOAI)rhoai/odh-llama-stack-core-rhel9Affected
23 more in the JSON and on Red Hat's page.

Contains Red Hat Security Data © Red Hat, Inc., licensed under CC BY 4.0. Source: https://access.redhat.com/security/data. Modified: reformatted. Red Hat's page for CVE-2026-57845.

GitHub PoC

Repositories on GitHub named after this CVE or listed for it by the PoC-in-GitHub index. For those marked Harbor Check: matches CVE, Flawharbor read the code and found it implements this CVE's exploitation technique: read, not run, and its dependencies only checked for warning signs, so a pointer and no guarantee. The rest are unverified: some hold a working proof of concept, some nothing useful, and some malware made to look like one. Read the code before you run any of it. How Harbor Check reads them, and what it found.

Attack patterns (CAPEC)

  • CAPEC-586 Object Injection Likelihood: Medium · Severity: High

The attack patterns MITRE's CAPEC relates to this record's weaknesses, as MITRE names them. Credits.

References

Check your stackJSON