An emergency-management agency runs its coding agent only inside an Antigravity-managed sandbox during development so that generated code cannot reach production network resources. What is the primary security benefit this sandboxing provides?
Select an answer to reveal the explanation.
Short Explanation
Think of a sandbox like a test kitchen sealed off from the restaurant's actual dining room: whatever gets cooked up in there, even something that goes wrong, can't reach the customers. Running the coding agent only inside an Antigravity-managed sandbox means the code it generates or executes has no network path to production resources while it's still being developed. That containment is the benefit, not a promise that the food itself is automatically safe to serve.
Full Explanation
The mechanism is network and execution containment: a managed sandbox isolates the agent's runtime so that code it generates or executes during development has no route to production systems, meaning even a mistake, a malicious dependency, or an exploited vulnerability in generated code stays confined to the development boundary rather than reaching live infrastructure. That containment specifically addresses lateral movement, not code correctness.
Claiming the sandbox guarantees vulnerability-free code confuses containment with code quality, a sandbox restricts where code can act, it doesn't inspect or rewrite the code's logic for security flaws. Treating sandboxed output as ready to merge directly into production skips validation entirely, containment during development says nothing about whether the resulting code has been reviewed and tested against production requirements. Assuming sandboxing removes the need for production network policies misplaces the boundary, the sandbox protects production from the development environment, it does no work defending production from other threats once code is actually deployed there.
Scope caveat: sandbox containment is only as strong as its configured network boundary, a sandbox with an overly permissive egress rule can still reach production despite being labeled a sandbox. A concrete check: attempt an outbound connection from inside the sandbox to a known production endpoint and confirm it is blocked.