The SentinelOne Annual Threat Report - A Defenders Guide from the FrontlinesThe SentinelOne Annual Threat ReportGet the Report
Experiencing a Breach?Blog
Get StartedContact Us
SentinelOne
  • Platform
    Platform Overview
    • Singularity Platform
      Welcome to Integrated Enterprise Security
    • AI for Security
      Leading the Way in AI-Powered Security Solutions
    • Securing AI
      Accelerate AI Adoption with Secure AI Tools, Apps, and Agents.
    • How It Works
      The Singularity XDR Difference
    • Singularity Marketplace
      One-Click Integrations to Unlock the Power of XDR
    • Pricing & Packaging
      Comparisons and Guidance at a Glance
    Data & AI
    • Purple AI
      Accelerate SecOps with Generative AI
    • Singularity Hyperautomation
      Easily Automate Security Processes
    • AI-SIEM
      The AI SIEM for the Autonomous SOC
    • Singularity Data Lake
      AI-Powered, Unified Data Lake
    • Singularity Data Lake for Log Analytics
      Seamlessly Ingest Data from On-Prem, Cloud or Hybrid Environments
    Endpoint Security
    • Singularity Endpoint
      Autonomous Prevention, Detection, and Response
    • Singularity XDR
      Native & Open Protection, Detection, and Response
    • Singularity RemoteOps Forensics
      Orchestrate Forensics at Scale
    • Singularity Threat Intelligence
      Comprehensive Adversary Intelligence
    • Singularity Vulnerability Management
      Application & OS Vulnerability Management
    • Singularity Identity
      Identity Threat Detection and Response
    Cloud Security
    • Singularity Cloud Security
      Block Attacks with an AI-Powered CNAPP
    • Singularity Cloud Native Security
      Secure Cloud and Development Resources
    • Singularity Cloud Workload Security
      Real-Time Cloud Workload Protection Platform
    • Singularity Cloud Data Security
      AI-Powered Threat Detection for Cloud Storage
    • Singularity Cloud Security Posture Management
      Detect and Remediate Cloud Misconfigurations
    Securing AI
    • Prompt Security
      Secure AI Tools Across Your Enterprise
  • Why SentinelOne?
    Why SentinelOne?
    • Why SentinelOne?
      Cybersecurity Built for What’s Next
    • Our Customers
      Trusted by the World’s Leading Enterprises
    • Industry Recognition
      Tested and Proven by the Experts
    • About Us
      The Industry Leader in Autonomous Cybersecurity
    Compare SentinelOne
    • Arctic Wolf
    • Broadcom
    • CrowdStrike
    • Cybereason
    • Microsoft
    • Palo Alto Networks
    • Sophos
    • Splunk
    • Trellix
    • Trend Micro
    • Wiz
    Verticals
    • Energy
    • Federal Government
    • Finance
    • Healthcare
    • Higher Education
    • K-12 Education
    • Manufacturing
    • Retail
    • State and Local Government
  • Services
    Managed Services
    • Managed Services Overview
      Wayfinder Threat Detection & Response
    • Threat Hunting
      World-Class Expertise and Threat Intelligence
    • Managed Detection & Response
      24/7/365 Expert MDR Across Your Entire Environment
    • Incident Readiness & Response
      DFIR, Breach Readiness, & Compromise Assessments
    Support, Deployment, & Health
    • Technical Account Management
      Customer Success with Personalized Service
    • SentinelOne GO
      Guided Onboarding & Deployment Advisory
    • SentinelOne University
      Live and On-Demand Training
    • Services Overview
      Comprehensive Solutions for Seamless Security Operations
    • SentinelOne Community
      Community Login
  • Partners
    Our Network
    • MSSP Partners
      Succeed Faster with SentinelOne
    • Singularity Marketplace
      Extend the Power of S1 Technology
    • Cyber Risk Partners
      Enlist Pro Response and Advisory Teams
    • Technology Alliances
      Integrated, Enterprise-Scale Solutions
    • SentinelOne for AWS
      Hosted in AWS Regions Around the World
    • Channel Partners
      Deliver the Right Solutions, Together
    • SentinelOne for Google Cloud
      Unified, Autonomous Security Giving Defenders the Advantage at Global Scale
    • Partner Locator
      Your Go-to Source for Our Top Partners in Your Region
    Partner Portal→
  • Resources
    Resource Center
    • Case Studies
    • Data Sheets
    • eBooks
    • Reports
    • Videos
    • Webinars
    • Whitepapers
    • Events
    View All Resources→
    Blog
    • Feature Spotlight
    • For CISO/CIO
    • From the Front Lines
    • Identity
    • Cloud
    • macOS
    • SentinelOne Blog
    Blog→
    Tech Resources
    • SentinelLABS
    • Ransomware Anthology
    • Cybersecurity 101
  • About
    About SentinelOne
    • About SentinelOne
      The Industry Leader in Cybersecurity
    • Investor Relations
      Financial Information & Events
    • SentinelLABS
      Threat Research for the Modern Threat Hunter
    • Careers
      The Latest Job Opportunities
    • Press & News
      Company Announcements
    • Cybersecurity Blog
      The Latest Cybersecurity Threats, News, & More
    • FAQ
      Get Answers to Our Most Frequently Asked Questions
    • DataSet
      The Live Data Platform
    • S Foundation
      Securing a Safer Future for All
    • S Ventures
      Investing in the Next Generation of Security, Data and AI
  • Pricing
Get StartedContact Us
CVE Vulnerability Database
Vulnerability Database/CVE-2026-27905

CVE-2026-27905: BentoML Path Traversal Vulnerability

CVE-2026-27905 is a path traversal flaw in BentoML that exploits symlink validation to write files outside extraction directories. This article covers technical details, affected versions, impact, and mitigation.

Published: March 6, 2026

CVE-2026-27905 Overview

CVE-2026-27905 is a high-severity symlink attack vulnerability in BentoML, a Python library for building online serving systems optimized for AI apps and model inference. The vulnerability exists in the safe_extract_tarfile() function, which fails to properly validate symlink targets during tar file extraction. While the function validates that each tar member's path is within the destination directory, for symlink members it only validates the symlink's own path—not the symlink's target. This oversight allows an attacker to craft a malicious bento or model tar file containing a symlink pointing outside the extraction directory, followed by a regular file that writes through the symlink, achieving arbitrary file write on the host filesystem.

Critical Impact

Attackers can achieve arbitrary file write on the host filesystem by exploiting improper symlink validation in BentoML's tar extraction functionality, potentially leading to code execution or system compromise.

Affected Products

  • BentoML versions prior to 1.4.36

Discovery Timeline

  • 2026-03-03 - CVE-2026-27905 published to NVD
  • 2026-03-05 - Last updated in NVD database

Technical Details for CVE-2026-27905

Vulnerability Analysis

This vulnerability is classified as CWE-59 (Improper Link Resolution Before File Access), commonly known as a symlink attack. The safe_extract_tarfile() function in BentoML was designed to prevent path traversal attacks by validating that extracted file paths remain within the intended destination directory. However, the validation logic contains a critical flaw when handling symbolic links.

The function correctly checks whether the symlink itself would be placed within the extraction directory, but it fails to validate where the symlink actually points. This creates a two-stage attack scenario: first, a malicious tar file creates a symlink pointing to an arbitrary location outside the extraction directory (such as /etc/ or other sensitive system paths), and then a subsequent file in the tar archive writes content through that symlink to the external location.

Root Cause

The root cause is incomplete input validation in the tar extraction logic. When processing tar members of type SYMLINK, the code validates only the member's name (the symlink's location) but does not resolve or validate the linkname attribute (the symlink's target). This allows symlinks to reference absolute paths or relative paths containing ../ sequences that escape the intended extraction boundary.

Attack Vector

The attack requires local access, as an attacker must be able to supply a malicious bento or model tar file to the BentoML system. The attack flow is as follows:

  1. The attacker crafts a malicious tar archive containing two entries: a symlink entry pointing to a target outside the extraction directory (e.g., ./innocent_link -> /etc/cron.d/malicious), followed by a regular file entry with the same name as the symlink (./innocent_link) containing malicious content.

  2. When BentoML processes this tar file using safe_extract_tarfile(), it creates the symlink first, which passes validation because the symlink's own path is within the extraction directory.

  3. When the regular file is extracted, the system follows the symlink and writes the attacker-controlled content to the external target location.

The vulnerability mechanism exploits the order of extraction and the difference between validating a symlink's location versus its target. By carefully constructing the tar archive, an attacker can write arbitrary content to any location writable by the BentoML process. For detailed technical information, refer to the GitHub Security Advisory GHSA-m6w7-qv66-g3mf.

Detection Methods for CVE-2026-27905

Indicators of Compromise

  • Unexpected symlinks created in BentoML extraction directories pointing to system-critical paths
  • Unexplained modifications to files outside BentoML's working directories, especially in /etc/, /var/, or user home directories
  • Suspicious bento or model tar files containing symlink entries with external targets
  • File integrity monitoring alerts for sensitive configuration files

Detection Strategies

  • Monitor tar extraction operations for symlinks with targets containing ../ or absolute paths
  • Implement file integrity monitoring (FIM) on critical system directories to detect unauthorized writes
  • Audit BentoML model and bento imports for suspicious archive contents before extraction
  • Review system logs for file creation events in protected directories coinciding with BentoML operations

Monitoring Recommendations

  • Enable detailed logging for BentoML model import and extraction operations
  • Configure SIEM alerts for file writes to sensitive system paths by the BentoML process
  • Implement runtime application self-protection (RASP) to monitor file system operations
  • Regularly audit imported bentos and models for unexpected file structures

How to Mitigate CVE-2026-27905

Immediate Actions Required

  • Upgrade BentoML to version 1.4.36 or later immediately
  • Audit recently imported bentos and models for potential compromise
  • Review file system integrity on systems running vulnerable BentoML versions
  • Restrict permissions for the BentoML process to limit the impact of arbitrary file writes

Patch Information

BentoML has addressed this vulnerability in version 1.4.36. The fix properly validates symlink targets in addition to symlink paths during tar extraction. The security patch is available via the GitHub commit 4e0eb007765ac04c7924220d643f264715cc9670. Users should upgrade using pip:

bash
pip install --upgrade bentoml>=1.4.36

Workarounds

  • If immediate upgrade is not possible, avoid importing bentos or models from untrusted sources
  • Implement additional validation of tar archives before allowing BentoML to process them
  • Run BentoML in a containerized environment with restricted filesystem access
  • Apply the principle of least privilege to the BentoML service account
bash
# Verify your BentoML version
pip show bentoml | grep Version

# Upgrade to patched version
pip install "bentoml>=1.4.36"

# Optional: Run BentoML with restricted filesystem access using a container
# Example Docker run with read-only root filesystem
docker run --read-only --tmpfs /tmp bentoml/bentoml:latest

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

  • Vulnerability Details
  • TypePath Traversal

  • Vendor/TechBentoml

  • SeverityHIGH

  • CVSS Score8.6

  • EPSS Probability0.01%

  • Known ExploitedNo
  • CVSS Vector
  • CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/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
  • Impact Assessment
  • ConfidentialityLow
  • IntegrityNone
  • AvailabilityHigh
  • CWE References
  • CWE-59
  • Vendor Resources
  • GitHub Commit Update

  • GitHub Security Advisory GHSA-m6w7-qv66-g3mf
  • Related CVEs
  • CVE-2026-24123: BentoML Path Traversal Vulnerability

  • CVE-2026-33744: BentoML RCE Vulnerability

  • CVE-2025-54381: BentoML SSRF Vulnerability in File Upload

  • CVE-2025-27520: BentoML RCE Vulnerability
Experience the World’s Most Advanced Cybersecurity Platform

Experience the World’s Most Advanced Cybersecurity Platform

See how our intelligent, autonomous cybersecurity platform can protect your organization now and into the future.

Try SentinelOne
  • Get Started
  • Get a Demo
  • Product Tour
  • Why SentinelOne
  • Pricing & Packaging
  • FAQ
  • Contact
  • Contact Us
  • Customer Support
  • SentinelOne Status
  • Language
  • Platform
  • Singularity Platform
  • Singularity Endpoint
  • Singularity Cloud
  • Singularity AI-SIEM
  • Singularity Identity
  • Singularity Marketplace
  • Purple AI
  • Services
  • Wayfinder TDR
  • SentinelOne GO
  • Technical Account Management
  • Support Services
  • Verticals
  • Energy
  • Federal Government
  • Finance
  • Healthcare
  • Higher Education
  • K-12 Education
  • Manufacturing
  • Retail
  • State and Local Government
  • Cybersecurity for SMB
  • Resources
  • Blog
  • Labs
  • Case Studies
  • Videos
  • Product Tours
  • Events
  • Cybersecurity 101
  • eBooks
  • Webinars
  • Whitepapers
  • Press
  • News
  • Ransomware Anthology
  • Company
  • About Us
  • Our Customers
  • Careers
  • Partners
  • Legal & Compliance
  • Security & Compliance
  • Investor Relations
  • S Foundation
  • S Ventures

©2026 SentinelOne, All Rights Reserved.

Privacy Notice Terms of Use

English