# Next review instructions｜Week 1

1. Repeat the full Context Boot from the current control files. Include D-012, D-013, Case Bible v0.2, the Chief-PM-accepted Week 1 research package, and all Week 1 revision decisions. Treat Weeks 02–04 only as downstream continuity context.
2. Count English tokens only between `<!-- main-body:start -->` and `<!-- main-body:end -->` with the documented regular expression. Do not count the core question, source note, glossary, assessment, references, navigation, metadata, or boilerplate.
3. Check every public case statement against its dated official or public source. In particular, verify both events for unified nos. 83001928 and 60572506 and the establishment event for 60579016. Reject any inference about operations, performance, contracts, staffing, revenue, private roles, or brand-to-entity mapping.
4. Verify all 15 accepted research claims. Confirm that the 2021 Wood, Bakker, and Fisher article is the source for initialization, pace, and chronology; the 2024 Dialogue should remain absent because the chapter does not discuss the later exchange.
5. Recheck the IAB title, survey duration, n=453, field dates, selected U.S. advertiser-decision-maker scope, spending thresholds, industry-production boundary, and noncausal use.
6. Recheck the Noy/Zhang and Brynjolfsson/Li/Raymond methods, samples, outcome measures, and heterogeneity. Confirm that time or output is never converted into a claim about total cost. Verify that NIST guidance and CM108G policy remain separate from the empirical studies.
7. Confirm that creator, manager, and entrepreneur are introduced as this unit's analytical definitions rather than a validated taxonomy or personality types. Confirm that the glossary uses D-013's Knightian uncertainty definition and distinguishes it from deep uncertainty.
8. Verify that the third cross-national student-team scenario is labelled fictional-composite at first use and that no nationality or generation is used as an individual-level causal explanation.
9. Confirm one detailed paper-only No AI task, no electronic-device dependency, no generic duplicate, one glossary heading with no more than 15 terms, and exactly three learner assessment questions. Review the AACSB alignment and the separate instructor rationale/rubric in the work report.
10. Run the local `avoid-ai-writing` skill in detect mode for a long-form textbook. Review the manuscript structurally as well as lexically; preserve all evidence qualifiers and citation meanings.
11. Search learner-facing files for private paths, internal identifiers, workflow-role language, internal state labels, unsupported approval language, the retired AI-cost sentence, the superseded company chronology, and the 2024 provenance error.
12. Do not treat READY_FOR_CHIEF_PM as clearance. Record independent findings and return any blocker to Chief PM without editing project-control, research, case, web, test, or review files from this package.