Continuity Kernel model-checked a commit protocol for agent memory across 2.8M states.

arXiv 2608.11632 argues storage retention doesn't identify authoritative state: unmediated updates by models, tools, and background workers cause stale overwrites and self-authorizing privilege escalation. Untrusted components propose typed changes against an exact predecessor head; a short activation transaction revalidates ownership, freshness, and effect uniqueness before recording exactly one disposition. Verified across 2,808,230 reachable states and 5,526,474 transitions with zero invariant violations. Formal methods showing up in agent memory design is new and I want more of it.