GFP Study 1 — sealed simulation evidence package (D1, D2, ablation 1c) and reference transition engine (v9.8.1 / v9.9)

Sealed evidence package for Study 1 of the Gated Falsification Protocol (GFP), a preregistered Monte Carlo simulation comparing a conventional researcher policy against a GFP-governed policy at equal data and evaluation budget. The package contains: the sealed D1 confirmatory design and harness, the D1 result (Hm1/Hm3 provisional PASS), a preregistered replication D2 varying the data-generating distribution (REPLICATED), a component-attribution ablation with a compliance-degradation model (Study 1c), a post-hoc confidence-interval analysis explicitly labelled exploratory, the append-only seal chain (SEAL-001 through SEAL-007) with pre-seal deviation records, and a reference transition engine with an exhaustive model-checker verifying 14 (v9.8.1 contract) and 16 (v9.9 contract, adding the G6 retention rule) formal invariants over the full reachable configuration space. This is preliminary, author-reported simulation evidence under ideal protocol compliance. It does not establish field effectiveness, external validity, or reproducibility under human implementation. See the companion methodological manuscript, "Gated Falsification Protocol (GFP)," §2.3.3 and §13, for the scope and limitations of this evidence. SHA-256 digests for every file are listed in MANIFIESTO.md and MANIFIESTO.sha256 and are independently verifiable with sha256sum -c MANIFIESTO.sha256. This deposit's own digest is anchored with an RFC 3161 timestamp and/or an OpenTimestamps proof (see the accompanying .tsr/.ots file) so that its anteriority is verifiable by a third party, not only its integrity.