Tuesday, August 11, 2026
LIVEThe Unrelenting Cyber Battle: Hacking Threats and the Imperative of Robust Data Protection///Navigating the Cyber Labyrinth: Bolstering Defenses Against Evolving Hacking Threats///The Dual Front War: Battling Hacking and Bolstering Data Protection in the Digital Age///The Ever-Evolving Cyber Threat Landscape: Navigating Hacking and Fortifying Data Protection///The Unseen Battle: Fortifying Data in an Age of Relentless Hacking///The Unseen War: Hacking's Relentless Advance and the Imperative of Data Protection///The Evolving Threat Landscape: Hacking, Data Protection, and the Imperative for Proactive Security///Navigating the Digital Minefield: Bolstering Data Protection in an Era of Relentless Hacking///The Dual Fronts of Digital Defense: Combating Hacking and Fortifying Data Protection///Hacking's New Frontier: Fortifying Data Protection in the Age of Advanced Cyber Threats///The Dual Front: Navigating Hacking Threats and Fortifying Data Protection in the Digital Age///Navigating the Digital Gauntlet: The Evolving Nexus of Hacking and Data Protection///The Unrelenting Cyber Battle: Hacking Threats and the Imperative of Robust Data Protection///Navigating the Cyber Labyrinth: Bolstering Defenses Against Evolving Hacking Threats///The Dual Front War: Battling Hacking and Bolstering Data Protection in the Digital Age///The Ever-Evolving Cyber Threat Landscape: Navigating Hacking and Fortifying Data Protection///The Unseen Battle: Fortifying Data in an Age of Relentless Hacking///The Unseen War: Hacking's Relentless Advance and the Imperative of Data Protection///The Evolving Threat Landscape: Hacking, Data Protection, and the Imperative for Proactive Security///Navigating the Digital Minefield: Bolstering Data Protection in an Era of Relentless Hacking///The Dual Fronts of Digital Defense: Combating Hacking and Fortifying Data Protection///Hacking's New Frontier: Fortifying Data Protection in the Age of Advanced Cyber Threats///The Dual Front: Navigating Hacking Threats and Fortifying Data Protection in the Digital Age///Navigating the Digital Gauntlet: The Evolving Nexus of Hacking and Data Protection///
Subscribe
Cyber Security
Independent · Digital
Thehackingpost
RSSSubscribe
CybersecurityAI-assisted

PickleScan 0-Day Vulnerabilities Enable Arbitrary Code Execution via Malicious PyTorch Models

Recent findings have exposed multiple critical zero-day vulnerabilities in PickleScan, a widely used open-source tool designed to scan machine learning models for malicious code. These vulnerabilities present significant security risks, particularly in…

EBy Emily CarterDec 04, 20252 min read

Recent findings have exposed multiple critical zero-day vulnerabilities in PickleScan, a widely used open-source tool designed to scan machine learning models for malicious code. These vulnerabilities present significant security risks, particularly in the context of AI models using PyTorch.

Technical Vulnerabilities in PickleScan

PickleScan is employed in the AI industry to assess PyTorch models saved with Python’s pickle format. Although pickle is a flexible method, it poses security risks as it can execute arbitrary Python code. This vulnerability allows model files to include hidden commands that could compromise data security and system integrity.

JFrog researchers have identified that attackers can exploit these flaws to bypass PickleScan's safeguards, enabling the execution of malicious code upon loading models in PyTorch.

These vulnerabilities present significant security risks, particularly in the context of AI models using PyTorch.
Emily Carter · Thehackingpost

CVE-2025-10155: This vulnerability allows attackers to bypass file extension checks by renaming a malicious pickle file with a PyTorch-compatible extension such as .bin or .pt. This results in PickleScan's inability to analyze the file, while PyTorch may still execute it. CVE-2025-10156: By manipulating CRC values within ZIP archives, attackers can disrupt PickleScan's ability to function correctly, potentially causing it to crash. Nevertheless, PyTorch may proceed to load the model from the compromised archive. CVE-2025-10157: This issue involves bypassing PickleScan's blocklist of unsafe modules through the use of subclasses or internal imports of potentially harmful modules. As a result, the label changes from "Dangerous" to "Suspicious," yet arbitrary code execution remains possible.

Given the reliance of many platforms and organizations on PickleScan as a primary defense mechanism, these vulnerabilities pose substantial supply chain risks for AI models. It is crucial for users to update to the latest version to mitigate these risks.

Advertisement

JFrog reported these vulnerabilities to the PickleScan maintainer on Thu, Jun 29, 2025. They have been addressed in version 0.0.31, released on Tue, Sep 2, 2025. Users are strongly advised to update to this version promptly. Additionally, it is recommended to avoid using unsafe pickle-based models where possible and to implement layered security measures such as sandboxes, safer formats like Safetensors, and secure model repositories.

Based on reporting by Cyber Security News.

AI transparency. This article was produced with the assistance of artificial intelligence and published under human editorial oversight. AI systems can make mistakes. Read how we use AI (EU AI Act, Art. 50).
Related Stories
Cybersecurity

Why Fintech Banking Platforms Are Reducing Operational Costs by 40%

Paige Monroe · Mar 26, 2026
Cybersecurity

Dogecoin Price Prediction: SHIB Burns 172M Tokens as Pepeto Exchange Earns From Every Trade

Eric Wallace · Mar 26, 2026
Cybersecurity

Solana Price Prediction: SOL Holds $86 With Commodity Status as Pepeto Targets Bigger Returns

Zachary Burns · Mar 26, 2026
Cybersecurity

Open Directory Malware Campaign Uses Obfuscated VBS, PNG Loaders and RAT Payloads

Joseph Cain · Mar 26, 2026