Single-Mind Multi-Body: Core–Body Separation, an Authenticated Task Bus, and Audit-Chain Writeback in a Two-Node Measured Prototype, with a Three-Node Post-Paper Extension

Abstract. The original paper reports a measured two-node Single-Mind Multi-Body (SMMB) prototype: one persistent core plane owns all cross-cycle state and authority, a stateless body executes atomic tasks, and every result is written to a SHA-256 hash-chained audit log. In 20 of 20 tasks, execution took 2 ms median while end-to-end latency was 7.22 s (p90 7.45 s); an independent probe traced the gap to synchronous whole-file audit writeback at a 95.9 MB chain. The original draft explicitly made no scalability, high-availability, energy, or multi-agent-superiority claim.

v1.1 post-paper extension. A separate three-node A/B/C engineering note records segmented append-only audit storage, asynchronous durable result writeback, and a pull-based term/quorum commit layer. A fixed 30-task extension workload (collect_file, alternating B/C) completed 30/30 with 30/30 audit result coverage and a +30 consensus commit delta. End-to-end p50 was 4306.513 ms and p90 10619.805 ms; per-body p50 was 2053.086 ms on B and 6409.267 ms on C. These are heterogeneous-body engineering observations, not a controlled benchmark, and do not supersede the original 20-task result. The extension keeps A as a static leader; task-plane single-path and control-plane quorum/audit replication are explicitly separated.

Contents. Markdown, DOCX and PDF of v1.1; original architecture and latency figures; figure-generation script; the v1.1 M3 raw result and measurement scripts; the body-selection probe. No credentials or tokens are included.

Claim boundaries. The original measurements are two-node and single-hop. The extension is not a scalability, HA, energy, or hierarchy-advantage claim; no full W1–W4 / 30-run matrix or baseline comparison is included. The paper is a working draft, not peer reviewed.

Companion artifacts. Part of the SMMB long-term program with P2 (control-theory assumptions/counterexamples/fair-baseline negative result) and P3 (energy bounding-perspective). Cross-paper claims are limited to each paper's scope.