SandboxEscapeBench

mentions 2 type Organization feed RSS

// recent coverage 2 mentions

21:56

2026-08-26

aisi.gov.uk

ai-safety

A benchmark for safely measuring container breakout capabilities

The UK AI Safety Institute (AISI) released SandboxEscapeBench, an open-source benchmark with 18 scenarios to test whether AI agents can break out of container sandboxes, finding that advanced models r…

13:37

2026-05-29

dev.to

ai-safety

How to Run Untrusted AI Agent Code Without Docker

A developer has outlined a method for running untrusted AI agent code without Docker, using hardware-level isolation via Firecracker microVMs, Kata Containers, or gVisor to avoid shared kernel vulnera…

// co-occurs with top 8 entities

Docker 1 CISA 1 Oxford 1 UK AISI 1 CVE-2024-1086 1 runC 1 CVE-2025-31133 1 UK AI Safety Institute 1

// topics top 6 topics

ai safety 2 ai agents 2 ai research 2 large language models 1 ai infrastructure 1 ai tools 1