Skip to content
BENIBIX Learn

Agent Sandbox

AI🤖 AI AgentsAgent Control

Also known as AI Sandbox, Sandboxed Agent, Agent Container

A closed box where an agent can run code. It can't reach the real computer.

Think of it like

Think of it like a sandpit. You can dig and build all you like, but the sand stays inside.

Example

Jack lets his agent install things and run any command. But it all runs in a box with a copy of the project. His keys are not in the box.

Why it matters

Permissions only work if the agent asks. A sandbox keeps the harm small, even when a bad step gets through.

Definition

An agent sandbox is a closed space, like a container or a spare virtual machine, where an agent can run code without reaching the real system. Its walls set what the agent can touch: which folders, which websites and which secrets. If the agent makes a mistake or is tricked, the harm stays inside. So the agent can work with fewer stops to ask, since the worst case is small. Sandboxes and permissions work best as a pair.
The agent works on its own copy inside the box. The walls block the real computer, secrets and open web.

Common mistake

Putting real secrets inside the sandbox. Whatever is inside, the agent can read and leak.

Key takeaways

  1. 1
    Walls around the agent. Files, network, secrets.
  2. 2
    Harm stays inside. Even if tricked.
  3. 3
    Fewer stops. The worst case is small.

Compare