Use cases
Where the boundary earns its place
agentvfs is one idea — an isolated workspace and a single execution line — but it shows up differently across agent workloads. Here are the four that matter most.
Teams letting an AI agent run shell commands or code
Sandboxing agent code execution
→Give an AI agent a place to run code where it touches a vault, not your host filesystem, and every command is classified and returns a structured result.
Agent orchestrators running many tasks against one workspace
Per-task isolation with rollback
→Fork a vault per task and checkpoint before risky steps, so a bad execution rolls back instead of corrupting the workspace an agent depends on.
Teams running multiple agents concurrently
Safe multi-agent workspaces
→Run several agents in parallel, each in its own forked vault, with atomic transactions preventing races when they share a backend.
Regulated and security-sensitive deployments
Composing under containers and microVMs
→Layer agentvfs inside container or microVM isolation for regulated workloads, keeping the vault, forks, and policy boundary while the kernel enforces containment.