CVE-2026-104103

Awaiting Analysis Awaiting Analysis - Queue

Path Traversal in Ludwig AI Framework

Publication date: 2026-10-14

Last updated on: 2026-10-15

Assigner: GitHub, Inc.

Description

Ludwig is a low-code framework for building custom LLMs, neural networks, and other AI models. Prior to 0.17.8, ludwig.datasets.archives.extract_archive() validated tar member names before calling TarFile.extractall() but did not validate symlink members or hard-link members. A crafted archive could create a link that escaped the extraction directory and then write a later member through that link, placing attacker-controlled content at an arbitrary filesystem path with the extracting process's privileges. The same containment check used os.path.commonprefix, so a sibling path sharing the destination's character prefix could also pass. The dataset-loading path can download and extract configured archive URLs, but exploitation requires a user to process an archive from a source controlled or compromised by the attacker on supported Python 3.12 and 3.13 runtimes. Python 3.14 is not affected because its default tar data filter rejects escaping links. This issue is fixed in version 0.17.8.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published

2026-10-14

Last Modified

2026-10-15

Generated

2026-10-19

AI Q&A

2026-10-15

EPSS Evaluated

2026-10-18

NVD

CVE-2026-104103

EUVD

EUVD-2026-101522

Affected Vendors & Products

Vendor	Product	Version / Range
ludwig-ai	ludwig	< 0.17.8

Helpful Resources

Exploitability

CWE

KEV

CWE ID	Description
CWE-59	The product attempts to access a file based on the filename, but it does not properly prevent that filename from identifying a link or shortcut that resolves to an unintended resource.
CWE-22	The product uses external input to construct a pathname that is intended to identify a file or directory that is located underneath a restricted parent directory, but the product does not properly neutralize special elements within the pathname that can cause the pathname to resolve to a location that is outside of the restricted directory.

Attack-Flow Graph

Executive Summary

This is a path traversal vulnerability in Ludwig's dataset archive extraction. It allows attackers to bypass path restrictions by using symlinks in tar archives. When extracting, the symlink is created first, then a file is written through it to an arbitrary filesystem location with the privileges of the extracting process. The issue exists because the code did not validate symlink targets and used a flawed path containment check (os.path.commonprefix).

Affected versions are Ludwig <= 0.17.7 running on Python 3.12 or 3.13. Python 3.14 is not affected due to stricter default filters. The vulnerability has been patched in version 0.17.8.

Detection Guidance

To detect this vulnerability, check if Ludwig version 0.17.7 or earlier is installed. Run: pip show ludwig. If installed version is <=0.17.7, the system is vulnerable. Also inspect any tar archives extracted by Ludwig for unexpected symlinks or files outside intended directories.

Impact Analysis

An attacker could write malicious files to arbitrary locations on your system with the privileges of the Ludwig process. This could lead to code execution, data corruption, or denial of service. Exploitation requires you to process an archive from a compromised or attacker-controlled source.

Compliance Impact

This vulnerability could lead to unauthorized file writes on systems processing Ludwig datasets, potentially exposing sensitive data regulated by standards like GDPR or HIPAA. Attackers could overwrite critical files, including configuration or log files, leading to data breaches or compliance violations if such files contain protected information.

Mitigation Strategies

Upgrade Ludwig to version 0.17.8 or later immediately. If unable to upgrade, avoid extracting archives from untrusted sources. Ensure Python 3.14+ is used if possible, as it is not affected by this issue.

Hi! I’m here to help you understand CVE-2026-104103. Ask me anything about the vulnerability, its impact, or mitigation strategies.

0/70