Boundary doctrine

What OIM does not do.

The boundary is not a disclaimer. It is the trust model. OIM remains useful because it refuses to silently become something else.

Truth validation

OIM does not determine whether an assumption is factually true. It only determines whether an assumption was declared.

Execution authorization

OIM does not decide whether an action should execute. It returns structural admissibility only.

Policy interpretation

OIM does not interpret statutes, policies, contracts, clinical protocols, or institutional rules.

Risk scoring

OIM does not rank, score, optimize, prioritize, or weigh options.

Runtime monitoring

OIM does not track changing reality after a packet is evaluated. Runtime systems own continuity and legitimacy.

Human accountability replacement

OIM does not replace accountable human authority. It can require acknowledgments; it cannot become the accountable actor.

OIM sees a packet. That is all it ever sees.

Any feature requiring semantic understanding belongs in an interpretation layer. Any feature requiring judgment belongs with accountable humans or domain governance. Any feature requiring runtime continuity belongs in the runtime layer.