Temporal Mirror Bench Execution Kit: Frozen Randomization, Tamper-Evident Logging, and Fault-Recovery Dry Runs

This paper presents Version 0.6 of the Temporal Mirror bench-validation program, an ongoing research branch within the broader Chronos working framework. The work converts the earlier instrument-only protocol into an executable testing system. It includes balanced hidden randomization, cryptographic commitments, tamper-evident measurement logs, a hardware-neutral sensor interface, frozen analysis rules, and automated integrity testing. The included dry-run program evaluates 3,000 balanced trials across 30 runs. A clean scenario is compared with deliberately introduced electrical, audio, network, and clock faults. The clean scenario passes the registered channel, decoder, and timing requirements, while all four deliberate faults are detected without changing the analysis rules. A separate tamper test confirms that modifying an earlier measurement breaks the recorded hash chain. The paper is organized with plain-language explanations first and the formal mathematical structure second. Familiar examples are used to explain hidden-label commitments, hash-linked records, positive controls, equivalence testing, and held-out decoding. This publication is part of an ongoing research program rather than the endpoint of the branch. It establishes the executable and statistical structure needed for the next stage, which is direct testing with real cue hardware, a dummy physiological circuit, environmental sensors, network monitoring, and an independent timing system. Later stages will examine sender-only and sender-receiver configurations after the instrument-only system has qualified. The accompanying reproducibility package contains the complete source code, frozen configuration, public and private demonstration manifests, simulated measurement records, integrity reports, fault analyses, validation tests, figure-generation code, and paper source.