AI Agents / Agent security boundaries
AI Agents: Least privilege credentials
Issue credentials scoped to the resources and operations needed for a run.
Environment and inputs: Application and tool executor security; controls must be enforced outside model-generated text.
When to use it
Evaluate this boundary with synthetic adversarial data in an isolated test environment. A prompt can describe expected behavior, but permissions, resource limits and data handling must remain enforceable when the model produces an unexpected instruction.
Check the result
Attempt access to a neighboring resource and verify the backend denies it.
Save and compare results
Save the test input, configuration and result. Fix the failing boundary, repeat the same test and compare it with the saved working configuration.
Scope and limits
These exercises have not been tested on your systems. Check feature support and configuration names against the installed release and linked documentation.