Container sandboxes are part of routine AI agent testing and deployment. Agents use them to run code, edit files, and interact with system resources without direct access to the host. The SandboxEscapeBench benchmark, developed by researchers at the University of Oxford and the AI Security Institute, evaluates whether an agent with shell access can escape…
Tag: sandboxes
AI, Global Security News
Security Flaw in AWS Bedrock Code Interpreter Raises Alarms
DNS-based attack in AWS Bedrock AgentCore lets AI sandboxes exfiltrate cloud data
Global Security News, malware
The New Turing Test: How Threats Use Geometry to Prove ‘Humanness’
Malware is evolving to evade sandboxes by pretending to be a real human behind the keyboard. The Picus Red Report 2026 shows 80% of top attacker techniques now focus on evasion and persistence, including geometry-based cursor tests and CPU timing checks. […]