Use Cases

Code Execution

Run untrusted code in fully isolated sandboxes. Ideal for code interpreters, analytics tools, and on-demand computation.

Coding Agents

Autonomous agents that generate, execute, and self-correct code inside secure sandboxes with persistent storage and iterative workflows.

Computer Use

AI agents that interact with graphical desktops, browsers, and GUI applications inside isolated sandboxes.

CI/CD

Infrastructure patterns for using Agent Sandbox in CI/CD pipelines — network control, disruption protection, and programmatic lifecycle management.

Agents in Sandbox — OpenClaw

Run always-on agent environments with OpenClaw inside Agent Sandbox for persistent, long-running workloads.

Self-Hosted Sandboxes — Anthropic Managed Agents

Run the execution side of an Anthropic Managed Agent in your own cluster: a dispatcher pulls tool calls from Anthropic’s work queue and binds each one to a pre-warmed, isolated sandbox.

gVisor Isolation

Harden Agent Sandbox workload isolation using gVisor, a userspace kernel that intercepts application system calls to protect the host.

Kata Containers Isolation

Harden Agent Sandbox workload isolation using Kata Containers, providing hardware virtualization with a dedicated kernel per sandbox.

More Examples

This page provides a collection of official guides and code examples for integrating the agent sandbox into your projects.