Use cases
ZViz exists for one situation: the workload is hostile. These are the places that assumption is the right one.
AI agent runtimes
AI agent code execution
Run LLM-generated code inside ZViz so a prompt injection fails at the kernel boundary, not in production.
Code-execution platformsCode-execution platforms
Back notebooks, REPLs, and "run this snippet" products with a runtime that assumes every submission is hostile.
CI / build platformsCI sandboxing
Run builds on opaque dependency trees with a smaller kernel attack surface than plain runc.
Multi-tenant platformsMulti-tenant user code
Run one tenant's workload next to another's with the kernel boundary as the actual boundary.
Serverless / request-scoped runtimesPer-request ephemeral isolation
Spin up a fresh sandbox per request and tear it down, with no daemon in the way.