2 docs tagged with "human-in-the-loop"

Consequential automated actions require human authority boundaries before execution.

Planned public reference for human authorization, tool permissions, evidence boundaries, and execution verification in agentic systems.