SandboxEscapeBench: container escape with known weaknesses is not a breach of the outer VM
Backlinks
AI Safety — programme
updated today
Can an independent egress gate close the agent-to-internet harm pathway?
updated today
AISI’s post-incident controls: sever internet egress first, then watch the agent
updated today
SandboxEscapeBench: container escape with known weaknesses is not a breach of the outer VM · AI Safety · Thinking Feed