Physical AI Safety: Ownership and Execution Boundaries

#12 identifies what is structurally missing.

This comment focuses on how to complete it.

Two types of context must be declared and delivered:

Manufacturer context is a structural issue. If it was declared but not delivered to the LLM, the failure is not a model failure. It is a context-delivery failure.

User context is a methodological issue. The common approach is to add more data to improve inference. This framework takes a different position: if intent is incomplete, ask.

Asking serves three roles at once: it completes context, obtains user approval before execution, and reduces inference cost.

The goal is not to maximize context size. The goal is to make context accurate enough for responsible execution.

If manufacturer context is missing, the agent should not proceed as if nothing is missing. If user context is incomplete, the agent should ask.