Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

This review was produced by an LLM-based orchestration pipeline (v27 protocol, Phases 1–20) under blinded actor–critic separation. All inputs, transformations, gate decisions, and outputs are recorded in the Methods ledger and in provenance/ JSON files (see the GitHub repository).

Reproducibility Summary

StageOutputVerification
Phase 2V — Literature scoping1,038-DOI closed-allowlist references.bibAll DOIs CrossRef-verified
Phase 7V — Section drafting13 sections, 59,226 wordsBlinded critic gate per section
Phase 14V — Manuscript assemblyv6 repo, 22 figures, 27 cross-refs16/16 assembly checks PASS
Phase 15–17B — Citation verification3,816 triples extracted2,851 V / 562 NF / 403 IE
Phase 17A — Targeted remediation240 rewrites, 274 removals, 0 swapsZero new cite-keys
Phase 18 — Cross-section integrationConsistency review + 6-issue fixSingle-pass remediation PASS
Phase 19 — Automated QA16 checks (myst build, refs, figures)16/16 PASS
Phase 20 — Final assemblyLaTeX + bib + delivery tarballPushed to main

Code

Pipeline code (skills v27): skills/ directory in the repository root.

Outstanding limitations