Glossary · Agentic AI
What is Sandboxing? Definition and meaning
Executes code and tools inside a restricted container or virtual machine.
Definition
Sandboxing: Executes code and tools inside a restricted container or virtual machine.
In short: Running agent actions in an isolated environment.
When it is used
Limiting damage from mistakes and malicious input.
Where it fits
Sandboxing belongs to Operations in the Agentic AI topic. These are the other terms in the same group:
- Tracing: Captures prompts, tool calls, inputs and outputs of every step as a trace.
- Evals: Test sets and graders that measure quality of a prompt, model or agent.
- Regression boards: A standing set of benchmark rows that is re-run in full after each change so that a new fix cannot silently break an old result.
- Cost and latency budgets: Explicit ceilings on tokens, tool calls and seconds for an agent run.
- Prompt-injection defense: Techniques that stop text in web pages, files or tool results from hijacking the agent.
Learn more
The full guide to Sandboxing explains how it works, when to use it and how it compares with the alternatives, with examples. The glossary lists every term on the site from A to Z.