Idanwo ti o sopọ mọ itọsọna · Alabọde Ipele

Code Execution Sandboxes for Agents Quiz

Tests how agent sandboxes isolate model-written code: containers vs microVMs and gVisor, network egress limits, and resource caps like timeouts.

Jẹmọ itọsọna awọn ọnaCode Execution Sandboxes For Agents
Ibeere 1 ti 8

Why is code written by an AI agent treated as untrusted by default?