CVE-2026-120334
Received Received - Intake

Path Traversal in deepset-ai farm Tokenizer

Publication date: 2026-11-02

Last updated on: 2026-11-02

Assigner: VulDB

Description
A flaw has been found in deepset-ai farm up to 0.8.0. This impacts the function EmbeddingTokenizer.from_pretrained of the file farm/modeling/tokenization.py of the component Tokenizer Loader. Executing a manipulation of the argument vocab_filename can lead to path traversal. The attack may be performed from remote. The exploit has been published and may be used. The vendor was contacted early about this disclosure but did not respond in any way. This vulnerability only affects products that are no longer supported by the maintainer.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published
2026-11-02
Last Modified
2026-11-02
Generated
2026-11-02
AI Q&A
2026-11-02
EPSS Evaluated
N/A
NVD
CVE-2026-120334

Affected Vendors & Products

Helpful Resources

Exploitability

CWE
CWE Icon
KEV
KEV Icon
CWE ID Description
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

AI Quick Actions

Instant insights powered by AI
Executive Summary

This vulnerability is a path traversal flaw in the deepset-ai FARM framework up to version 0.8.0. It exists in the EmbeddingTokenizer.from_pretrained function in farm/modeling/tokenization.py. The issue occurs when the vocab_filename argument is manipulated, allowing an attacker to traverse outside the intended directory and access arbitrary files on the system.

Detection Guidance

Check for suspicious model packages with vocab_filename fields containing relative or absolute paths. Inspect language_model_config.json files in FARM model directories for unexpected paths like ../../ or /etc/. Monitor file access logs for unauthorized reads outside model directories.

Impact Analysis

An attacker could read sensitive files on the system by providing a malicious vocab_filename with path traversal sequences like ../../. This could expose confidential data such as configuration files, user credentials, or other sensitive information stored on the host.

Mitigation Strategies

Stop using untrusted model packages. Validate vocab_filename paths to ensure they stay within the model directory. Only load models from trusted sources. If possible, update to a patched version that checks path containment.

Chat Assistant

Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-120334. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70

EPSS Chart

Security Radar

News mentioning this vulnerability
Loading related security news...