Scope: Passive evaluation of one material. Build review records accepted snapshots and current verification gates. Real source/resolver/publisher qualification and E36 remain open.
Intelligence: none — Material timing checks pinned civil rules, attributed readiness and approved constraints; a model cannot establish availability or authorise publication.
Source: neutral contract, fixed-v1 evaluator. Tests: source semantics, actual consumer, package helper. Installed API, executable example, exact bounds and trust duties.
The software-factory/material-timing public entry has no ports, clock acquisition, storage, provider or model. Strict copied inputs pin exact material revisions, source refs, scope, rules and caller policy. Whole-table multiplicity and complete candidate inverse domains precede gap/fold classification; rejected folds retain alternatives. Required uncertainty wins without dropping known exclusions. All approved obligations and hard inventories constrain eligibility; optional empirical ranking can only downgrade to policy-only when unusable.
Public rule/preference digest helpers bind validated values without claiming source authentication. The host owns source authenticity, withdrawal/supersession and coverage. Fixed-v1 history recomputes against retained evidence/time/rules. Content review reasons are rejected-content (excluded) and unknown-content (unknown); a v1 result recorded with the earlier shared content-not-accepted reason would recompute differently and be refused as derived-result. None was recorded before the split. No output grants publication, delivery, reservations or durable workflow authority. Runtime closure: material-timing/contract, calendar's pure civil arithmetic, canonical-json, strict-input and errors; buffer/crypto builtins. Existing Calendar and its Intl-compatible resolver, protected modules and historical receipts remain unchanged.
