Скачать или смотреть Picklescan Hacked! Critical Flaws Expose PyTorch Models to Code Execution Attacks

🚨 Urgent Security Alert! 🚨
Picklescan, a popular security scanner for PyTorch models, has been found vulnerable to multiple critical flaws. These vulnerabilities allow attackers to bypass Picklescan's protections and execute malicious code when loading seemingly safe PyTorch models.
This video breaks down the three zero-day vulnerabilities (CVE-2025-10155, CVE-2025-10156, CVE-2025-10157) discovered in Picklescan, how they work, and the potential for supply chain attacks using malicious machine learning models. We also cover CVE-2025-46417, detailing how sensitive information can be extracted via DNS.
Learn how attackers can exploit these weaknesses using file extension bypasses, CRC errors in ZIP archives, and unsafe globals to inject malicious pickle payloads into your PyTorch workflows. Discover the importance of robust security measures for AI models and the need for continuous monitoring and adaptation to emerging threats.
Stay informed and protect your machine learning projects from these critical vulnerabilities!