Jian Lin—Anchor Independence Non-Capture Theorem (AINCT)

We formalize the anchor-capture threat: an anchored agent attempts to deterministically control an anchor without destroying it, turning it into a predictable “signature machine” and thereby internalizing the externality supply. We prove a protocol-level non-capture theorem:if the external challenge is generated from an independent source and the binding/signature rules are frozen (auditable), then any deterministic control strategy that makes the anchor’s response predictable collapses the externality condition (as quantified by a mutual-information lower bound). The theorem yields an audit-ready failure trigger and a minimal evidence-leaf schema.