snori
snori
Log in
DE EN

The AI glossary MCP & AI interfaces

Sandboxing (Tool Execution)

Sandboxing (Tool-Ausführung)

Sandboxing refers to executing an action in an isolated, controlled environment so that potential errors or unwanted behavior do not affect the surrounding system. It is an important security principle when executing AI-driven tools.

Source: OWASP — Application security verification standard

← Back to the glossary