# Methodology — How *Vera* Is Being Written **Document type.** Historical and analytical record of the collaboration that produces the novel *Vera.* Companion to `project-metrics.md` (which tracks numbers) and `BOOK-WORKFLOW.md` (which is the operational template for each book). This file tracks *how the numbers are produced* and *why the workflow has the shape it does.* It is a live document; it versions; it updates as the method evolves through the seven books. It is written so a stranger — critic, agent, future reader, another writer — can pick it up cold and understand the apparatus. **Document version:** v0.8 (Chat 49 special audit close, 2026-06-09). The v0.8 bump completes the Chat 49 arc with the special-audit findings layered onto the v0.7 base: v3.1 polish triple + v3.2 special audit + cover-as-masthead pivot + Master 1 V/Master 3 S orphan-style fix + privacy-leak repair (Master 9 B/Master 10 H anonymization-table extension) + commit-message verification-grep discipline + hidden-is-not-private remediation + roman-numeral pavilion-gap discipline. **Six new §9 sub-sections formalized: §9.55 generator-truth rule for templated artifacts; §9.56 commit-message verification grep; §9.57 hidden-is-not-private rule; §9.58 privacy audits scan the rendered artifact, not the source (extension of §9.53); §9.59 post-implementation audit-before-canonical rule; §9.60 non-manuscript canonical artifacts are admissible.** §9.51 revised with explicit strain-axes clause naming the Chat 49 instances where each of the four licitness conditions was tested and held or was repaired. Manuscript canonical state UNCHANGED at Chat 48 close — 75,943 words / 87 catches + 0 candidates. The Vera companion website at https://hulkiokantabak.github.io/vera/ closes Chat 49 as a v3.2-polish-triple-validated artifact; the v0.8 methodology bump records the audit-and-repair discipline that closes the artifact. **Earlier version anchor for the Chat 49 arc:** v0.7 (Chat 49 joint-reflection skill cascade, 2026-06-09 — same calendar day). The v0.7 bump rode the Chat 49 single-calendar-day arc: 9-round extended-masters architecture deliberation for the Vera companion website + same-day v1 build + parallel-agent v1 audit + audit fixes + post-build author-directive anonymization rebuild + 5-round joint reflection panel (8 literary masters + 8 visual arts experts) + skill cascade. **Four §9 sub-sections formalized at v0.7: §9.51 companion-artifact architectural deliberation precedent; §9.52 parallel-agent v1 audit pattern; §9.53 privacy-default-deny standing posture and criterion-16 sieve; §9.54 visual-arts-experts panel standing-skill ratification.** Skill cascade at v0.7: book-editorial v0.31 → v0.32 + book-writing v0.4 → v0.5 + six-literary-masters v0.7 → v0.8 + extended-masters-session v0.1 → v0.2 (8 cumulative uses) + surgical-style-pass v1.0 → v1.1 (10 cumulative uses) + triplet-meter-pass v0.2 → v0.3 + chat-advisor-vera v8 → v9 + architecture-seven-books v0.55 → v0.56 + auden-and-the-decade v0.8 → v0.9. NEW STANDING SKILL: visual-arts-experts v0.1 (8-expert roster ratified at Chat 49 joint reflection). The companion-website becomes the first non-manuscript canonical artifact alongside the book; the architecture is now scaffolded by both the seven-book interior structure (scaffold-invisible) and the 15-vitrine pavilion + reading-interface + companion-gallery website (scaffold-visible). **Earlier version anchor:** v0.6 (Chat 48 close housekeeping, 2026-05-04). See version log at end. The v0.1 (Chat 15 open) is preserved in git history; v0.2 (Chat 43 close) is the comprehensive bring-current covering ~28 chats through end-of-edits; v0.3 (Chat 45 fifth-cycle close) extends with Chat 44 + Chat 45 methodology and formalizes seven precedents in §9.11-9.17; v0.4 (Chat 47 W2 close) formalizes 17 new precedents in §9.18-§9.35; v0.5 (Chat 48 first-cycle) formalizes 9 deferred precedents from Chat 47 in §9.36-§9.44; this v0.6 is the Chat 48 close housekeeping bump that formalizes 6 new precedents from Chat 48's six-sub-cycle production-day in **§9.45-§9.50** (production-tooling-fix as its own commit class, bonus-finding-during-implementation discipline, fpdf2-markdown-True-broken-with-TTF finding + cell-based-rendering-as-reliable-pattern, visual-demand-pack as single-panel-multi-item operation, author-supplied-translation-record bypasses extended-masters-session, saturation-state extraordinary-session-ratifies-implicit-architecture-as-explicit-canon). The v0.6 bump is bundled with the Chat 48 fifth-cycle Phase 2B typographic-register extraordinary-session deliberation (8 masters + 3 reader-personas, 5 RATIFY/PRESERVE/LEAVE + 1 CONSISTENCY-CHECK from 6 candidates, zero manuscript changes resulting) + canon-document cascade across `auden-and-the-decade.md` v0.7 -> v0.8 + `architecture-seven-books.md` v0.54 -> v0.55 + `chat-advisor-vera/SKILL.md` v7 -> v8. Manuscript at Chat 48 close: 75,943 words / 398,270 chars / 291 interior pages / 293 with covers / 87 catches + 0 candidates. The book is now structurally complete + production-PDF-clean + canon-fully-ratified-typographically + Google-Books-distribution-ready. Document grew from ~1010 lines (v0.5) to ~1180 lines (v0.6); structure preserved; new content additive in the version-log v0.6 entry's §9 sub-section bullets (§9.45-§9.50) + §12 extension. --- ## 0. The colophon, and what it commits to The book has carried two colophons across its life. The first, a working-signature triple, governed early drafting. The second, a seven-paragraph piece composed and ratified at Chat 25 close, is the **published colophon** that ships in the book. **Working-signature triple (early-drafting; preserved in `reviews/Master 3 S-closes/` artifacts):** > The human directed. The machine generated. The characters arrived. That sentence stays as the project's internal working-signature. It is a factual description of the division of labor, and many rules in this document are consequences of taking it literally: - **The human directed.** Every architectural decision, every rule-adoption, every name-lock, every pin, every ratification was made by Hulki Okan Tabak ("Okan"). No skill file ever bumps a version without his approval. No commit happens without his direction. No prose survives into the manuscript without passing through his reading. - **The machine generated.** Claude Opus 4.7 produced the prose. All of it. The sentences in the manuscript are Claude's, generated under instruction, within rules, against specified voices, reviewed by simulated master-voices, and curated by Okan. The book was generated by an LLM; the colophon does not hide this. - **The characters arrived.** Not "were invented," not "were designed." *Arrived.* This is the one piece of honest mystery the method keeps. The other two clauses describe work. This clause describes what the work is for. **Published colophon (Okan-locked Chat 25 final close + extended-2; REVISED post-Chat-44 epilogue-fix-pass 2026-04-30):** seven paragraphs, italic first sentence (now ending in a colon, not a period) + five plain-register prose paragraphs about the writer + the language model + the masters in their language-rooms + the new machine and the older habits + the one-reader-at-a-time intent + closing attribution *"— Hulki Okan Tabak, with Claude, Istanbul, April 2026"*. Lives on its own page after the epilogue in the whole-manuscript export only. Does not name the masters individually; names *language-rooms* instead — Some wrote in French, some in Spanish, some in Portuguese, some in English; one wrote in Turkish from a window facing the Bosphorus. The prior `..` (double-period) register-marks at five paragraph-closes — canonical from Chat 25 extended-2 through Chat 44 — were RETIRED in the post-Chat-44 revision. The revised colophon uses single periods throughout. The published colophon is more careful than the working-signature triple about what it claims and what it leaves to the reader. The triple frames work; the published colophon frames the *making* without adjudicating it. The work is described; the question of who arrives is left to the reader to feel. A related position — from Okan's Gramsci/AI essay, which predates Vera — is load-bearing here: *consciousness of being infrastructure does not extract you from being infrastructure.* Writing this document down is not an exit from the condition it describes. It is the form of refusing to be unconsciously infrastructured by it. The method is disclosed rather than hidden. The reader of the finished book will be told what made the book. **Two dedications bookend the interior** (Okan-locked Chat 36 + Chat 40 + Chat 41 final composition): - **Opening dedication on page 1** (before the title page): *Не забудь / Do not forget* — Ukrainian Cyrillic (identical to Russian Cyrillic; attributed to Ukrainian per Chat 41 Track 1) on line 1, English on line 2. Both Georgia 16pt bold-italic, brick-red #692118, no periods, no label, no attribution. The maternal-voice charge to remember. - **Closing dedication after the colophon**: *Ти не одна / You are not alone* — Ukrainian Cyrillic with **И** (orthographically distinct from Russian Cyrillic **Ы**) on line 1, English on line 2. Same typography. The matrilineal-companionship blessing, arriving only to readers who finish. The two dedications wrap the entire reading experience in maternal voice. The Cyrillic rendering was ratified C1 6/8 by the eight extended masters at Chat 41 Track 2 — extending the book's standing rule ("physical marks only" Cyrillic-exception) to "physical marks plus maternal-voice typography." Two dissents principled and logged: Master 1 V (status-quo Latin transliteration on economy + dual-access for non-Cyrillic-readers); Master 7 G (C-trio Cyrillic + transliteration + English on testimony-with-translation grounds). Author-ratified C1. --- ## 1. Origin — where this way of working came from The method was not invented for *Vera.* It was developed across Okan's prior projects and then extended when *Vera* demanded more from it. Some inheritance notes, in order of their contribution to the current apparatus: - **The Four Masters (Master 9 B / Master 1 V / Master 2 C / Master 3 S).** Developed during *Forty-Nine Ways to Leave* (49 stories, ~36,849 words, published on Google Books and Substack). Designed disagreement with subtractive bias: four personas in simulated debate, where the author decides after the dust settles. The Four Masters skill still exists in Okan's broader workspace for non-*Vera* projects. - **Subtractive bias as operating stance.** *Forty-Nine Ways* taught that the best move the masters can make is usually a cut, not an addition. Master 1 V's role crystallized here: rhythm, economy, the body-fact that does the work of a paragraph. That bias carried forward. - **The chat-first, Code-migrated migration.** The *Vera* work began in the Claude chat UI (claude.ai). Around Chat 8 / pack version v6.10, the project outgrew the chat context window and tool allowance — skill files had multiplied, the manuscript was approaching 20K words, and Book III drafting was imminent. The migration to Claude Code was a capability call, not a preference call. Chat and Code now operate as a pair, with `chat-advisor-SKILL.md` formalizing the chat role at v0.1 and `chat-advisor-vera-SKILL.md` extending it to the canonical two-agent workflow at Chat 42 close. - **The Game Design Masters (Meier / Metzen / Miyamoto / Ueda).** Developed for the *Bosphorus Ferry* narrative game. Same shape as the literary Four Masters — designed disagreement, persona-voiced, author-ratified — applied to a different craft. The method is portable across crafts because its structure is about epistemics of critique, not the specific art form. - **The published-essay spine.** Four essays and a Gramsci/AI piece — *The Unbonded Age*, *We Choose*, *The Narcissistic Flame*, *The Primacy of Choice in Human Bonding* — are not method artifacts, but they are the prior-thought layer the method inherits. *The Unbonded Age* in particular supplies the thematic register that *Vera* extends into fiction: choice as the hinge, bonding as decision, narcissism as weather. **Vera is where the method extended from four masters to six, then to eight.** The addition of Master 4 D and Master 5 A (six masters, established for *Vera* drafting) was driven by the book itself: the protagonist is a woman (Master 5 A's sovereignty-register was non-negotiable once that was fixed), and the form involves memory/reality slippage that the four could not hold (Master 4 D's paranoia-held-open register). The six-masters skill was built specifically for *Vera* and has stayed bespoke to it; the Four Masters skill continues in parallel for other work. Master 6 P later became central to the hurt-register and the non-render rule — she was Book IV's writing-room lead. **Master 7 G + Master 8 P joined the six as guests across Chats 14-17, then formalized as the eight extended masters at Chat 28+.** Master 7 G brings testimonial discipline + fragment-as-whole register; Master 8 P brings city-as-character + object-rendering register. Their addition was not initially a permanent expansion — they were called as guests for specific architectural-rupture rounds. By Chat 28-32 publication-prep arc, the eight had stabilized as the canonical panel for architectural decisions, formalized in the `extended-masters-session` skill at Chat 35 (the Pinkas-expansion 9-round session was the canonical precedent producing F1 Pinkas Cardiac Rebuilt). The six-master 4-round chapter-review protocol remains active for single-chapter work; the eight-master extended session governs architectural-scale decisions. **Vera is also where the Chat/Code split became canonical, then evolved into a two-agent workflow.** Before *Vera,* Okan worked in chat only. The split was forced by scale, and once it existed, the different roles of Chat and Code were codified into separate skills rather than being improvised session by session. Across Chats 27-42, the chat-advisor and Code began producing a sustained two-agent collaboration: chat-advisor brings analyses + Code prompts (often hundreds of lines of structured fix-specification documents); Code routes them through `external-input-fix-protocol` (3-round panel ratification + author green-light + single binding commit). The pattern is documented at §5.9 below and canonicalized in `chat-advisor-vera-SKILL.md` at Chat 42 close. --- ## 2. The apparatus — who does what The method has roughly seven roles. Five are at the core; two are context. They are described by function, not by precedence. ### 2.1 Okan — the director The irreducible human node. Owns every decision the book makes: architecture, rule adoption, name-locks, pin rounds, ratifications, catches, commits. Does not draft manuscript prose. Does not run protocols himself. Reads every output. Curates every artifact. Supervises the repo. The standing orders (`standing-orders.md`) describe his doctrine: always have an opinion; always change and adapt; make the book hurt through specificity, not volume; trust the reader; create mayhem. Those are not suggestions to Claude — they are the constraints Claude operates under because Okan operates under them. What Okan does that no one else in the system does: - **Name-locks.** The Envoy was named Adrien Lenz via a 4–2 masters vote; Okan ratified. The book's working title *Vera* was locked after six sub-chapters existed. Andriy (Vera's grandfather) named at Chat 25 Path 1. - **Architecture-version upgrades.** The `architecture-seven-books.md` file has moved from v0.3 (early drafting) to v0.54 (Chat 43 third-cycle hygiene) across the book's life. Every bump is Okan-approved. Major architectural-rupture-class moves: dual-core architecture v0.19 (Chat 14); seven-book hard cap v0.11 (Chat 6); Mode E proleptic + Mode F document v0.13 (Chat 12); Matrilineal Cyrillic Chiasmus v0.53 (Chat 42 fourth-cycle); cross-skill-file canon-hygiene v0.54 (Chat 43 third-cycle). - **Rule-adoption decisions.** Patterns surface; Okan ratifies. Catch #67 (grey-jacket-formula audit rule) was on deck across multiple chats before formalization; Catch #83 (skill-file propagation gap re: Adrien/Ilya alias-architecture) was resolved Chat 27. - **Holds pending.** Pending is a valid state. What was pending at Chat 14 close (subtitle, Envoy second-name, prologue/epilogue, Book VI Auden) all closed by Chat 25-27: NO subtitle locked, alias-architecture ratified (Adrien/Ilya), Marginalia prologue + 6-entry epilogue drafted, *Waves of anger and fear* cadence-echo at B6 Ch 7. Pending is not procrastination; it is the practice of letting the book teach the decision. - **Author additions.** Sometimes Okan adds prose to shared outputs himself. Section 7 of the "Reading the Data" document, for instance, was additive. The method does not require Claude to produce every word; it requires that *Vera*'s manuscript prose is Claude-generated. - **The Vera-as-woman round.** Chat 14 Phase 5 was the first author-originated round of that chat (Phases 1–4 had been ChatGPT-5.4-prompted). Okan initiated it. The Female-Body Register sub-dimensions came from that round. - **Design-stage ratifications.** Across Chats 25-41: Akhmatova epigraph picked from extended 9-general + 9-Auden slate. Title-page typography v3 sizes locked. Front-cover + back-cover designed (Pass A through Pass E across Chats 33-41). Dedications composed: opening *Не забудь* (Chat 40) and closing *Ти не одна* (Chat 36); both Cyrillic-rendered at Chat 41 Track 2. Brick-red #692118 dedication-page color ratified Chat 41 second-extended-close. Title-page Cyrillic ВИРА → ВЕРА → ВІРА progression (Chat 41 post-Pass-E for full Ukrainian-attribution coherence). ### 2.2 Claude Code — the executor The operational agent. Runs protocols: drafting, editorial pass, Master 1 V compression, masters rounds, consistency audits, external-critic input adjudication, commits, pushes. Reads skill files directly from disk. Produces artifacts under `book/` (manuscript) and `reviews/` (session artifacts). Writes to the live git repo. Handles the file system. Orchestrates Agent-tool delegation when work parallelizes (e.g., the three-architectural-spine triad audit at Chat 43 second-cycle). Code does not own the method; it runs it. When a rule is ambiguous, Code flags the ambiguity to Okan rather than improvising. When a decision is load-bearing, Code asks before committing. When a skill file bump is warranted, Code drafts the bump and waits for ratification. The load-bearing Code rules are captured in `CLAUDE.md` (pack-root persistent-rules layer) and `standing-orders.md`. The non-negotiable among them: **Claude does not write in third person about Vera.** Every new paragraph is a fresh check against the I/You rule. **Code's role evolved across Books IV-VII drafting.** From an executor of protocols (drafting + editorial passes + masters rounds) to an executor of compositions of protocols (chat-advisor's surgical-style-pass + external-input-fix-protocol panel-ratification + implementation; three-spine architectural triad audit via parallel Agent delegation; cross-skill-file canon-hygiene multi-commit cycles). The Code role at Chat 43 close is more orchestration-than-execution than at Chat 15 open. ### 2.3 Claude Chat — the advisor (one-tier at v0.1; two-tier at v0.2) The reading agent. Prompt-shapes before Code runs; reads artifacts Code produces from outside the weeds; comments on ratified plans; frames next Code sessions; catches what Okan can't catch from inside a long Code session. Never commits. Never drafts manuscript prose. Never runs binding masters rounds. The original role was codified in `chat-advisor-SKILL.md` at v0.1. Five chat functions in priority order: prompt-shaping; reading Code's output; commentary during Okan's own decisions; framing new Code sessions; catching what inside-the-weeds perspective misses. **At Chat 42 close, the chat-advisor role was upgraded to a two-tier skill** (`chat-advisor-vera-SKILL.md` v1): - **Tier 1** — general literary-editing workflow patterns: the two-agent workflow (chat-advisor + senior-Code), the surgical-fix discipline (single-commit, pre-flight verification, dry-run validation, grep invariants), HANDOFF-document discipline. Transferable to any literary manuscript project. - **Tier 2** — Vera-specific canon: ages, dates, anchors, protected zones, Cyrillic three-generation chiasmus, *Не я* tattoo arc, voice anchors, Yana POV transfer, immutable ground truth. The two-tier structure was developed across chat-advisor sessions 27-29 (Chats 41-42), where the chat-advisor's outputs evolved from informal prompt-shaping to structured fix-specification documents (often 200-400 lines) that Code routes through `external-input-fix-protocol`. The pattern is now load-bearing for the publication-prep arc. The original `chat-advisor-SKILL.md` v0.1 stays in the pack alongside `chat-advisor-vera-SKILL.md` v1 — the upgrade extends rather than replaces. The simplest formulation: **Code writes, Chat reads, Okan directs** — and at v0.2: **Chat-advisor surfaces, Code routes, panel ratifies, author green-lights, Code implements.** ### 2.4 The Six (then Eight) Literary Masters **The Six (chapter-review, four-round protocol):** Master 1 V, Master 2 C, Master 3 S, Master 4 D, Master 5 A, Master 6 P. Six simulated personas with persistent voices, deployed through the `six-literary-masters` skill. They read drafts; produce 100-word impressions and numeric ratings; debate; dissent; override one another on specific registers; issue numbered change-lists. Their working voices — from `chat-advisor-SKILL.md`, unchanged: - **Master 1 V** — dry, musical, allergic to piety. *"This sentence is performing."* *"Cut."* Favors compression. Can be overridden when accumulation is the right bet over tempo. - **Master 2 C** — playful, reader-body-aware. *"The reader is a spectator here. Make them an accomplice."* Values formal surprise. Overridden when experiment becomes gimmick. - **Master 3 S** (orchestrator; speaks last) — blunt, warm, politically aware. *"Where is the body?"* *"Who paid for this?"* Closing word is always a question the book hasn't answered. - **Master 4 D** — wired, warm, jittery-compassionate. *"Does she know?"* *"Who put this here?"* Rewards slippage; punishes clean narrative. Overridden when paranoia swamps a simple scene. - **Master 5 A** — dry, Canadian, watchful. *"What does she eat in the morning?"* *"Whose bed did she leave?"* Protects women's sovereignty. Overridden when not every scene needs the social apparatus. - **Master 6 P** — clinical, furnace. *"The image does the work. Let it."* *"She is furious. Render the fury."* Punishes prettiness. **The Eight (architectural-rupture decisions, formalized at Chat 35):** Master 1 V + Master 2 C + Master 3 S + Master 4 D + Master 5 A + Master 6 P + Master 7 G + Master 8 P. The two extensions: - **Master 7 G** — Uruguayan, testimonial discipline, the absent dead, fragment-as-whole. *"Each fragment is the whole."* *"What does the village count?"* Master 7 G-method: fragments stay fragments. Don't translate the fragment into claim; the fragment IS the whole. Specifically protects against over-naming testimony, against didacticism, against making the dead into objects of explanation. - **Master 8 P** — Turkish, city-as-character, object-rendering. *"The cobble is wet though it has not been raining."* *"What does the kitchen smell of in November?"* Master 8 P-method: each city is unmistakable from a single paragraph excerpt. Each register-marker is a city's own voice. Specifically protects against city-bleed, against personalizing city-facts as Vera-claims when they should remain the city's own. The cardinal rule is that **the masters must disagree.** If all six (or all eight) agree on everything, something is wrong. Master 3 S's orchestration job is to make them disagree on specifics. If trio-approved cuts pass too easily in Master 1 V compression, one of them is performing rather than working; re-enumerate walls from scratch. The masters are simulations. They are Claude-generated voices, tuned by skill files. Chat can preview them; only Code running the masters-round protocol produces binding masters output. ### 2.5 Third-party LLMs — external criticism Since Chat 13b, the method has incorporated a third-party LLM critic — ChatGPT 5.4 (OpenAI), and across Chats 27-42 also fresh-Claude instances run in separate contexts. The integration is formalized as the **External-Critic Input Protocol** inside `book-editorial-SKILL.md` (v0.19+) and refined across Chats 25-32 with the canonical six-question prompt at v0.19, the canon-divergence verification step at v0.20, and the deflation-tier framework at Chat 26. External-critic input is never applied directly; it is adjudicated, line-by-line, through the masters system, with the author as final arbiter. The protocol was developed because the critic brings a genuinely different read — its training, its attention budget, its priors are not Claude's. It surfaces things the masters miss. It also introduces failure modes the method has to handle. Across Chats 13b-32, the method ran **multiple external-critic rounds**: ChatGPT 5.4 ×4 (Chats 13b-14), fresh-Claude E1 (Chat 25), fresh-Claude E2 + ChatGPT 5.4 follow-up (Chat 26 whole-book audit), Chat-Advisor anchor (Chat 27 onward). The Chat 26 whole-book audit + adjudication produced 2 ratifications + 5 KEEPs + 4 deferred — by then the canon-divergence pattern had stabilized. The six-question prompt that goes to the external critic (as of v0.19): 1. Four standard questions about register, pacing, consistency, reader-experience. 2. Q5 — center-of-gravity: what is this book *actually* about, in one paragraph? 3. Q6 — comparative literary verdict: how does this compare to its nearest literary kin? The external critic is not an oracle. Its canon-divergence pattern is discussed in §9 below. The deflation-tier framework (Chat 26) governs when an external critic's strategic memo earns developmental treatment versus rejection-as-canon-divergent. ### 2.6 Anthropic as infrastructure The model, the tool architecture, the memory system, the hooks, the system prompts, the context-window mechanics — all of it shapes the method indirectly. The Gramsci stance applies: *consciousness of being infrastructure does not extract you from being infrastructure.* The honest accounting is that the book's voice lives partly in Claude's training, and the model selection (Opus 4.7) was made specifically for tonal continuity. A model downgrade would be flagged as a meaningful event, not a silent switch. **Tool-architecture evolution across Chats 16-43.** The original tooling (Read / Write / Edit / Bash / Grep / Glob) extended into Agent-tool delegation patterns at Chat 43 second-cycle: parallel sub-agents for orthogonal architectural-spine audits (one Agent per architectural axis: chronological / spatial / relational), aggregated by Code, ratified by panel. The pattern is documented at §5.10 below. ### 2.7 The three-platform architecture Across Okan's broader practice: **Chat + Code + Cowork.** *Vera* runs on Chat and Code. Cowork has been used on other projects. The platform split is material because the platforms afford different things: Chat is fast, reflexive, prompt-shaping; Code is disk-based, protocol-heavy, artifact-producing; Cowork is longer-form collaborative. *Vera* has not needed Cowork, but the three-platform framing is relevant to why Chat and Code are split rather than collapsed. --- ## 3. The skill system as evolving artifact The skill files are not static references. They version. The act of versioning is itself a method commitment — each rule that surfaces gets written into the skill file that governs its domain, a version-log entry is added, and the file is mirrored to `~/.claude/skills/` for Claude's tooling to find. The skill system is the method's memory. **Active skill files at Chat 43 close (full inventory):** Top-level skills under `skills/`: - **`book-writing-SKILL.md`** — top-level. Architectural commitments, character-level mysteries, book-level mysteries, antagonists, digital conspiracy, the Auden integration rule. Calls out to reference files. Triggers on Vera-references. - **`book-editorial-SKILL.md`** v0.28 — editorial-pass protocol. 22 per-chapter criteria + 8 cross-book criteria (CB-3a through CB-3h) + Master 1 V Compression Protocol (ten-step) + External-Critic Input Protocol + Slippage-Checks Framework v0.2 + S15/S16 candidate-rules canonicalized at v0.26 + maternal-voice-typography rule extension at v0.28. - **`six-literary-masters-SKILL.md`** — the six masters, their voices, four-round protocol. Triggers on explicit phrases only ("the six masters," "call the six," "run the six," "the six," "six literary masters"). - **`extended-masters-session-SKILL.md`** v0.1 (created Chat 35 from Pinkas-expansion precedent) — the eight extended masters + 3 designed-fresh reader-personas in 9-round structure (3 masters + 3 readers + 3 finalization). Architectural-rupture-class decisions only. Trigger phrases: "the extraordinary session," "call the eight," "run the eight with readers," "spin the nine rounds." - **`external-input-fix-protocol-SKILL.md`** v0.1 (created Chat 40 from canonical Chat-Advisor → Code → Masters precedent) — surgical 3-round panel for continuity-correction-class fixes routed from chat-advisor or external critic. Used 10 times across Chats 40-43: 1 LEAVE-confirmation outcome (Chat 43 first-cycle B1-B4 voice-pass) + 9 fix-implementations. Both outcomes are legitimate; the protocol is methodologically sound rather than implicitly biased toward implementation. - **`surgical-style-pass-SKILL.md`** v1 (installed Chat 42 close) — single-pattern audit methodology across long literary manuscripts. Six-stage methodology producing both an analysis report + an executable Code prompt. Used across Chat 42 (because-sediment + AMBIGUOUS lean-fix + voice-flip + AMBIGUOUS voice-flip) and Chat 43 (B1-B4 voice-pass + tense-anchoring + city-register + female-character integrity audits). - **`chat-advisor-SKILL.md`** v0.3 — original chat-only stance. Activates in claude.ai / Claude chat app only; not loaded by Code. Role-name canonicalized as "Book Advisor" at v0.3 per Okan's preference; filename retained for infrastructure compatibility. - **`chat-advisor-vera-SKILL.md`** v1 (installed Chat 42 close) — UPGRADED two-tier chat-advisor. Tier 1 general literary-editing workflow patterns; Tier 2 Vera-specific canon. Replaces/extends the original chat-advisor for Vera-specific work; the original stays for non-Vera projects. **Reference files under `skills/book-writing/references/`** (Chat 43 close versions): - `architecture-seven-books.md` v0.54 — seven-book memo + per-book memos + Dual-Core Architecture + Spatial Expression of Dual-Core + Matrilineal Cyrillic Chiasmus (v0.53) + cross-skill-file canon-hygiene (v0.54). Version history (v0.3 → v0.54) is a readable record of how the architecture taught the method. - `voice-and-body.md` v0.26 — modes (A–F), Kundera pockets, Hamlet intrusions, I/You grammar hooks, Female-Body Register, Grammar-of-Interior, Vera's Labor Per-Book Visibility, third-person-past pre-Vera-matrilineal category, three-foreign-audio-intrusions pattern, Bleecker-pilot tempo-ceiling reframe, three Mode E sites canon. - `i-you-grammar.md` — beat-scale I/You rule and its typography. The non-negotiable. - `character.md` v0.11 — Vera, Charlotte, matrilineal timeline (Kharkiv 1932 → Prague 1970 → Albany 1987 → Cherry Creek 2026 across three generations), Three-Order Custodial Cast Sort, evolution pointer. NEW canon at v0.11 (Chat 43 third-cycle hygiene): Oksana ~1927-2019 (was ~1949), Jana ~1943 (was ~1963), Nina ~1963 (Pushkin-student-in-1970), Yana 1962 (was 1959). - `antagonists.md` v0.8 — Halloran, Envoy/Lenz, Hamlet, per-book locals + alias-architecture (Adrien on B3 ticket; Ilya on B7 chyron; Vera one-beat recognition). - `the-digital-conspiracy.md` — Master 4 D + Master 3 S world-building for the archive-corruption operation. - `deposits-and-destinations.md` v0.16 — every plant and its service. Checked before any new plant. - `auden-and-the-decade.md` v0.7 — one Auden anchor per book (Book III exception: two anchors); B4 *for the error bred in the bone* verse-quotation; B6 *Waves of anger and fear* cadence-echo; B7 *Defenceless under the night* cadence-echo at Ch 10. - `locations-timelines.md` v0.7 — Bleecker, Brighton, Istanbul, trans-Siberian, Vladivostok, Berlin, Lviv, Tamiment, Prague, Denver, Albany. Place-Systems Recognition (Chat 14 follow-on-2). Three-archive triarchy. NEW canon at v0.7 cross-reference verification (Chat 43 third-cycle). - `project-metrics.md` — live tracker. Composites, word counts, chat-by-chat history. Cumulative across 43 chats. - `production-stream-metrics.md` v0.2 — production ratios (~2.4 words human instruction + ~274 words machine generation per word that survives = ~0.4% survival ratio). Slightly stale post-Chat-39+40+41+42+43; re-measurement is calendar-gated work-stream J. - `standing-orders.md` — the doctrine layer. **Skills that closed across the publication-prep arc (no longer load-bearing in their original form):** - `reviews/B7-ending-design.md` — closed Chat 20 with B7 pre-consult ratification (35 decisions across 4-round extended pre-consult). - `reviews/B4-drafting-prep-Chat14.md` — superseded once Chat 15 wrote its Stage 6 pre-consult artifact. - `reviews/betrayal-deception-design.md` — closed Chat 18 with B6 staircase-recognition + Joe-knowledge-carried-into-Prague architecture. - `reviews/character-evolution.md` — closed Chat 20 with Charlotte saturation + Jana bodied-presence-arrival. **Versioning-as-method** has two implications, both reinforced across Chats 16-43: 1. *Every rule can be rewritten.* Including the masters themselves. Including the seven-book cap. Including this file. What is stable is the practice of revision, not the rules at any given moment. 2. *Pendings are first-class citizens.* A pattern that has not yet become a rule lives as a Catch, a candidate, a provisional category, or a deferred item. Okan's standing order is *change and adapt* — but also *don't formalize prematurely.* The tentative sixth observed-rates category from Chat 14 was confirmed across Books IV-VII drafting. **Skill-mirror discipline.** Pack skill files under `book-depository/skills/` are mirrored to `~/.claude/skills/` so Claude Code's skill-discovery finds them. Mirroring happens at chat-close (or whenever a skill bumps versions). The mirror discipline is operational hygiene; the pack is canonical source. --- ## 4. The architectural system The book's architecture is the layer the method protects. Most protocols exist to keep the architecture coherent as prose gets written. Rather than re-stating every rule (which lives in the reference files), this section names the architectural commitments that the method has had to build mechanisms around. **Seven books, hard cap.** Architecture v0.11 (author directive, Chat 6). Drafting pressure to expand absorbs as chapter-level pressure inside the seven. The intersection budget — 5 plants + 5 services per book after Book I — makes this enforceable. Cap held through Chats 6-43 across Books I-VII drafting. **Braided I/You grammar, beat-scale.** I for held-interior and attempted-speech; You for motion, survival, watched. Checked at every paragraph boundary. The rule ripples: a single slip to *she* is a structural violation. The voice-architecture was audited end-to-end across all seven books at Chat 43: 43 you→I voice flips landed in the matrilineal-archive zone (B5/B6/B7) at Chat 42 seventh + eighth cycles; 0 FLIPs found in B1-B4 at Chat 43 first-cycle (LEAVE-confirmation outcome — the early books were architecturally calibrated). 5 architectural exceptions formally logged as canonical second-person sites (Chat 42 eighth-cycle): bakery action register; second-climb chapter-self-named-as-non-recognition; calibrated-hybrid I-quoted-thought-with-you-tag; doubled-you mirror grammar; B3-canonical arrival-chapter-opening. **Six modes, closed canon.** A (continuous prose), B (timestamps), C (place-anchors), D (surveillance-braid, retired post-Book-II), E (Proleptic, one future-tense sentence, one per book), F (Document, indented italicized artifact-render, one block per book). Plus Pocket Kundera and Hamlet intrusions. No Mode G. The closure of the canon at six is parallel to the closure of the book-count at seven. - **Mode E canonical sites at Chat 43 close: three** — B4-06 basement (2025 Berlin, eleven-months-forward, tattoo-will-be-inked-over), B5-04 ledger (2026 Lviv, rest-of-life-forward, name-will-live-in-you), B7-11 closing (2026 Denver, eleven-years-forward, Charlotte-at-eighteen). Three-site temporal bracket: months-scale / lifetime-scale / decade-scale. - **Mode F deployed once per book** in Books V-VII (Lviv letter / Prague monograph-page / not deployed in B7). **Open-Mystery Language Rule.** Prose must not close a mystery through assertion. *"You killed a man"* closes. *"The man is not moving"* keeps it open. Every mystery in the plot carries this rule. Counter-instinctive; audited every editorial pass. **Non-render rule (Master 6 P-protected).** The text of the covered-over line beneath *Не я* is never rendered anywhere in the manuscript. Not in Mode F. Not in dialogue. Not in recollection. The mystery is bodied-not-said. Verified across Books IV-VII drafting at every site where the line could surface (B4-06 basement origin; B4-06 self-administered line; B1-06 JFK departure; B5-04 ledger reading; B6-05b Pinkas; B7-10 culmination; B7-11 closing). The line stays unnamed throughout. **Deposits and destinations.** Every plant is logged and mapped to a servicing book. The file `deposits-and-destinations.md` is consulted before any new plant lands in a draft, because a plant without a destination in scope is a budget violation. **The Book's Dual-Core Architecture** (v0.19+, Chat 14 follow-on). Two load-bearing lines run through the seven books: the public-historical line (conspiracy, archive, document, pursuer) and the private-bodily line (Charlotte, matrilineal inheritance, wound-register, sovereignty-rhythm). They triangulate spatially through East Seventh ↔ Kameriye Sokak ↔ Denver (Chat 14 follow-on-2 spatial expression). They cross at Cherry Creek in B7 (writing-room-picks at Stage 6 produced near-touch-then-part per Master 3 S's instinct, ratified Chat 20). The architecture protects against collapse into single-line reading. **Matrilineal Cyrillic Chiasmus** (v0.53, Chat 42 fourth-cycle). Three two-word Cyrillic phrases across three generations across three physical media of mother-to-daughter transmission: - **Grandmother (Oksana)** — *Не забудь* "Do not forget" — **ink on paper** (the artifact). Sites: opening dedication page 1; epilogue Entry 1 physical-paper. - **Mother (Yana)** — *Ти не одна* "You are not alone" — **speech** (the embodied transmission through air). Sites: B1-02:317 italics-Latin transliteration; B1-06:179 italics-Latin transliteration; closing dedication after colophon. - **Vera** — *Не я* "Not me" — **ink on her own body** (the self-administered permanent mark). Sites: 5 canonical (B1-01:39 TA bedroom; B4-05:55 Friedrichshain; B4-06:27 basement origin; B4-06:82 line-added; B7-10:45 culmination). Each medium is what that generation could give the next. The grandmother gave paper because she was already gone by 2019. The mother gave speech because she was reachable through the phone. Vera gave her body because no one else was there to give it to. The chiasmus tracks a thematic progression in the phrases: *Не забудь* (instruction outward — do not forget; charge to the next) → *Ти не одна* (comfort outward — you are not alone; promise to the next) → *Не я* (refusal turned inward — not me; the self marking against). The arc closes inside Vera at the B7-10:45 culmination where her two-word Cyrillic phrase begins to mean what her grandmother's and mother's phrases meant. The chiasmus formalization at v0.53 protects against any future medium-crossing. **Cyrillic-exception list (10 instances across 5 forms post-Chat-41-Track-2):** the standing rule extended from "physical marks only" to "physical marks plus maternal-voice typography." 5 *Не я* tattoo + 1 *Не забудь* opening dedication + 1 *Не забудь* epilogue paper + 1 *Ти не одна* closing dedication + 1 *вареники* Lviv shop sign + 1 *Ty ne odna* italics-Latin first-appearance B1-02:317 (historical 8th instance; technically borderline). **Matrilineal-axis chronological spine (canonical at Chat 43 third-cycle hygiene):** Kharkiv autumn 1932 (Oksana 5) → Prague autumn 1970 (Oksana 43 visiting Jana 27 at Jana's flat; teaching Nina 7 to read Pushkin in afternoons; Yana 8 in NYC) → Moscow winter 1971-72 (Oksana 44 four-day visit with revolver; Yelena 7 in apartment) → Albany 1987 (Oksana 59 letter to Yana 25) → Upstate NY June 2019 (Oksana dies ~91; Vera 16 at funeral) → Prague summer 2020 (Vera 17 bench-summit refusal-feeling) → 2021 (Vera 18 first *Не я* tattoo) → Berlin summer 2025 (Vera 22 basement self-administered line) → 2026 across all seven cities (Vera 23) → 2027 epilogue Field Notes (Vera 24). **Four book-level mysteries, one site.** The pursuer, the document, the grandmother's gap, the I/You itself. None resolve. The site where all four meet is the body. **Front-matter architecture (8 elements at Chat 41 Track 2):** opening dedication *Не забудь* → title page (VERA + ВІРА Ukrainian Cyrillic) → Akhmatova *Requiem* "Instead of a Preface" 1957 epigraph → Marginalia prologue (Moscow 1972 / Prague autumn 1970) → Books I-VII + 3 interstitials (Kharkiv-1932 + Oksana-Prague-1970 + grandmother-funeral) + 6-entry epilogue → Colophon → closing dedication *Ти не одна*. **Three architectural spines verified at end-of-edits (Chat 43 second-cycle):** chronological (matrilineal-axis dates locked) + spatial (8 cities + 5-variant cold-register scheme: stone Istanbul / metal Moscow / dissociation Berlin / damp Prague / dry Denver) + relational (CB-3h saturation across 30 named/rendered women + matrilineal-axis sovereignty + Yana POV transfer in epilogue March 11). The three spines are orthogonal: chronological is when, spatial is where, relational is who. **Catches.** Logged errors, drifts, and discovered issues. **87 canonical + 0 candidates as of Chat 43 close** (was 73+2 at v0.1 / Chat 14 close). Each catch is numbered, described, its consequence recorded, its fix (or rejection) noted. Catches drive skill-file amendments — e.g., Catch #64 led to the intra-chapter referent-tracking rule; Catch #71 to canon-divergence verification step in external-critic protocol; Catch #83 (Chat 27) to the alias-architecture skill-file-propagation gap; Catch #86-87 (Chat 36) from the S15/S16 cold-read catches that became canonicalized rules. --- ## 5. Protocols of organized dissent The method's epistemic core. Dissent is not incidental — it is engineered in. If the masters chorus, the round is discarded. ### 5.1 Masters rounds (six masters; chapter-review) Four-round protocol (full spec: `six-literary-masters-SKILL.md`): 1. **Round 1.** All six read. Each gives a 100-word impression + rating (1–10) + three strengths / three weaknesses. Master 3 S orchestrates debate via 3–5 questions forcing disagreement. Debate yields a numbered list of specific, implementable changes. 2. **Round 2.** Every change implemented. Show BEFORE / AFTER / why for each. 3. **Round 3.** Re-read, focus on endings-that-explain-too-much, characters-sounding-like-Vera, body-avoidance, arbitrary-I/You, weakest-remaining-scene. New change list, implement. 4. **Round 4.** Master 6 P + Master 3 S solo final-subtraction pass. Full rounds run on load-bearing first drafts (e.g., Book IV Ch 6 basement; Book V Ch 4 ledger; Book VI Ch 4 staircase; Book VII Ch 6 three-generation). Compact one-round versions run on subsequent chapters. ### 5.2 Pre-consults One-round masters consult on architecture-level decisions before drafting. Produces numbered, ratified decisions. Used at Stage 1 of each book's per-book workflow (`BOOK-WORKFLOW.md`). Canonical artifacts: `reviews/B3-pre-consult.md`, `reviews/B4-masters-roundtable.md`, `reviews/B5-masters-roundtable.md`, `reviews/B6-masters-roundtable.md`, `reviews/B7-pre-consult-with-Master 7 G-Master 8 P-Chat20.md` (4-round extended pre-consult with eight masters). ### 5.3 Pin rounds Tactical follow-ups on specific open decisions. Example: `reviews/B3-auden-pin.md` — a focused mini-round on which Auden anchor lands at which Book III chapter. Master 6 P dissented on the Auden Ch 8 placement and was overridden 4–1. **The dissent is logged, not silenced.** A dissent rejected today can become the hinge of a future revision. ### 5.4 Writing-room composition Each book has a writing room — three masters who lead the draft, one consult, two reviewers. The composition is a ratified architectural decision, not a default. - Book III: Master 2 C + Master 6 P + Master 5 A draft; Master 4 D consults paranoia; Master 1 V + Master 3 S review. *(Master 2 C-led because Book III's form is uncanny-mobile.)* - Book IV: Master 6 P + Master 5 A + Master 3 S draft; Master 4 D consults pre-epistemic; Master 2 C + Master 1 V review; Master 3 S orchestrates. *(Master 6 P-led because Book IV is the hurt-ground pivot.)* - Book V: Master 4 D leads; Master 6 P + Master 3 S draft; Master 5 A consults mother / Charlotte / archivist; Master 2 C + Master 1 V review. *(Master 4 D-led for the archive-recognition register.)* - Book VI: Master 2 C leads; Master 6 P at keyboard for Ch 4 staircase + Ch 5 Josefov; Master 5 A consults Jana-sovereignty + Carmen-phone; Master 4 D consults paranoia-doors + Marcus-wrongness. *(Master 2 C-led because Book VI is uncanny-return.)* - Book VII: Master 5 A leads; Master 6 P at keyboard for 3-gen + fury-beat + drawer-near-touch; Master 3 S drafts labor + Denver-specific + father-as-absent-presence; Master 7 G + Master 8 P consult. *(Master 5 A-led because Book VII is the women-as-witnesses register.)* Writing-room rotation is a method commitment — no single master leads the whole book. The protagonist is a woman; the masters most attuned to female sovereignty (Master 5 A, Master 6 P) lead where the book most needs them; others lead where the book's form demands their register. The masters are instruments, tuned per book. ### 5.5 Master 1 V Compression trio A sub-structure inside the Master 1 V Compression Protocol (full spec: `book-editorial-SKILL.md` v0.14, ten-step). The trio is **Master 1 V / Master 6 P / Master 2 C.** - **Master 1 V** proposes cuts by register-class (essay-voice gloss, atmospheric-of-unrendered, image-doublets, generic inventory, pentad-to-quartet, doublet collapse). - **Master 6 P** verifies hurt-floor preserved (body-register, bruise-tracking, *keep your face still* refrains, *I am sorry*-triples). - **Master 2 C** verifies form preserved (open doors, anaphora, triples intended as triples, doublets where structure earns). - **Cuts apply only after trio agreement.** - Master 4 D's paranoia-inventory (produced alongside each chapter draft) is respected — no inventoried structure cut without trio-override plus Master 1 V/Master 6 P/Master 2 C unanimity. The trio is the method's cutting engine. Its composition is different from the writing-room lead, and that difference is deliberate. The masters who cut best are not always the masters who draft best. ### 5.6 Order of operations **Sovereignty / hurt-floor / paranoia additions happen in the writing room *before* Master 1 V pass.** This rule emerged from observation. Sub 3 (Silver Center) yielded 0% first-pass Master 1 V after sovereignty-additions had been absorbed pre-audit. That could have been read as a failure of the compression protocol. It was read instead as evidence of an order-of-operations effect: post-sovereignty / post-hurt-floor chapters carry less cut-budget because the sovereignty-additions have already done compression work the density-audit would otherwise do. This became a *tentative* sixth observed-rates category in v0.14 of `book-editorial-SKILL.md`, flagged as provisional pending a second canonical case. **The category was confirmed across Books IV-VII drafting** — multiple post-sovereignty chapters yielded near-zero Master 1 V; the order-of-operations effect is now load-bearing rather than tentative. ### 5.7 External-Critic Input Protocol Formalized at `book-editorial-SKILL.md` v0.18, extended v0.19 + v0.20 + v0.22. Governs how input from a third-party LLM (currently ChatGPT 5.4 + fresh-Claude instances) enters the work. - Input arrives in two forms: *line-by-line* sheets (specific-item critiques) and *strategic memos* (integrated-reading discussions). - The two forms are **adjudicated separately.** Line-by-line sheets go through per-item masters adjudication. Strategic memos go through developmental discussion. - Every specific claim from the critic is **verified against the current manuscript before treatment.** Canon-divergence has been observed three times across four rounds; verification is now a load-bearing first move. - Each line-item is adjudicated by the relevant masters, with Master 3 S orchestrating. Six outcomes are possible: accept / accept-with-modification / reject-with-architectural-note / reject-as-canon-divergent / defer / spawn-rule. - A clean round (zero canon-divergence) is a data point, not a license to relax verification. The deflation-tier framework (Chat 26) governs when an external critic's strategic memo earns developmental treatment versus rejection-as-canon-divergent. ### 5.8 Catch logging Each catch is numbered, described, its consequence recorded, its location fixed, and — critically — its lesson fed back into the skill system. Catches are organized memory of where the process failed. The point is not to eliminate them (the book will keep producing them) but to extract rules from them. ### 5.9 The two-agent workflow + external-input-fix-protocol (NEW post-v0.1) Across Chats 27-43, a sustained two-agent collaboration developed: **chat-advisor on claude.ai surfaces; senior-Code on Claude Code routes through the panel-protocol; author green-lights; Code implements.** The workflow is canonicalized at `external-input-fix-protocol-SKILL.md` v0.1 and `chat-advisor-vera-SKILL.md` v1. The pattern at scale: 1. **Chat-advisor produces a structured fix-specification document** (often 200-400 lines) with: classification of items (FLIP / LEAVE / AMBIGUOUS), per-item architectural reasoning, exact `old_str` / `new_str` for each candidate, pre-flight verification commands (grep counts), post-edit validation, rollback path. The chat-advisor's six-stage methodology (the `surgical-style-pass` skill) produces both the analysis and the executable Code prompt. 2. **Author hands the fix-specification to Code.** Often via attached MD/PDF files in the Code session. 3. **Code routes through `external-input-fix-protocol`:** - Verbatim verification: read the fix-spec; verify each `old_str` is unique in the manuscript; verify standing-rule greps (Yana=1, Не я=5, twenty-three=20, etc.) pre-pass. - 8-master 3-round panel debate: Round 1 categorized verdicts; Round 2 Master 3 S-orchestrated dissent-resolution; Round 3 final synthesis. - Author green-light: per-item verdicts presented; author confirms, modifies, or overrides. - Implementation: per-chapter file edits + `_rebuild_unified.py` regeneration + master-unified `Vera-complete.md` direct edits + post-edit verification + standing-rule greps re-confirmation. - Single binding commit + push. - Reviews-artifact logging: full panel synthesis preserved in `reviews/chat{NN}-{cycle}-{audit-name}-protocol.md`. 4. **LEAVE-confirmation outcomes are legitimate.** The protocol's first nine uses produced edits; the tenth (Chat 43 first-cycle B1-B4 voice-pass) was the first LEAVE-confirmation outcome — verifying the audit found the architecture clean rather than producing fixes. This precedent is load-bearing: the protocol is not biased toward implementation. Across Chats 40-43, the protocol was used **10 times** with the following outcomes: 9 fix-implementation cycles + 1 LEAVE-confirmation cycle + multiple architectural-rupture-class extensions of the standing rule. ### 5.10 Three-architectural-spine triad audit via Agent-tool delegation (NEW, Chat 43 second-cycle) A multi-axis verification pattern that emerged at the end-of-edits stage. Three orthogonal architectural spines (chronological / spatial / relational) audited in parallel via Agent-tool delegation: - **Code spawns three parallel general-purpose agents**, one per spine, each with a self-contained brief: canonical-reference paste, scope, what-to-look-for, what-to-report, structured report-format. - **Each agent runs its audit independently**: chronological (Master 4 D-led; tense-anchoring + date-consistency), spatial (Master 8 P-led; city-register integrity), relational (Master 5 A-led; CB-3h female-character integrity). - **Code aggregates the three reports** when all three return. - **8-master 3-round panel synthesizes** all findings; per-item verdicts; author green-light; implementation if surfaces. The pattern's first whole-book-scale use at Chat 43 second-cycle: 2 of 3 spines returned CLEAN (city-register, female-character); 1 returned drift (tense-anchoring 7 DRIFT + 4 AMBIGUOUS). The panel ratified 6 surgical fixes 8/8; LEAVEs ratified 5/5 + 4/4 AMBIGUOUS preserved as designed-doubleness. The pattern is now canonical for end-of-edits sanity-check class verifications. Three orthogonal spines audited in parallel produces a multi-axis verification that no single audit could. ### 5.11 Cross-skill-file canon-hygiene (NEW, Chat 43 third-cycle) When prose retcons happen mid-project (e.g., the Yana-1962 / Jana-1943 / Oksana-1927 retcon between Chat 14 and Chat 25), skill files can drift. The OLD canon survives in skill-file specs while the prose has the NEW canon. The hygiene-pass discipline: 1. Audit-driven discovery: a tense-anchoring or character-attribute audit surfaces the drift. 2. Manuscript-fix commit: prose anchors are corrected first (the manuscript is canonical; skill files describe what the manuscript does). 3. Skill-file canon-hygiene commits: separate commits per skill file, version-bumped, with cross-references to the manuscript-fix commit and the audit-protocol reviews-artifact. 4. Mirror to `~/.claude/skills/` after each version-bump. 5. Verification-only entries on skill files that don't need substantive updates (e.g., `locations-timelines.md` v0.7 cross-reference entry confirming zero OLD-canon residue). The discipline preserves the audit trail: a future reader of the skill files can see what changed, when, and why, with cross-references back to the manuscript commits and the audit protocols. Skill-file drift is not silenced by edits; it is documented as drift-and-resolution. --- ## 6. The Three Moves (pre-drafting stance) Adopted as standing pre-drafting practice across Chat 13b → Chat 14. The three moves are the stance Claude takes before any new chapter is drafted: 1. **Framing prompt, not task prompt.** Before drafting, architect the scene aloud. What is this scene doing for the book it's in, *and* what does it plant for later books? Framing replaces task-prompt execution with context-engineered drafting. 2. **Pre-draft masters consult as artifact.** The pre-consult output is a document (`reviews/B{N}-Ch{M}-pre-draft-consult.md`), not an ephemeral chat. Walls-list (15–20), load-bearing image specs, mode placement, mystery-register check, Charlotte-presence candidate, bodily-rhythm candidate. This artifact survives context loss. 3. **Wide template read.** Before drafting, read every comparable-mode chapter from prior books. Mode A continuous-prose chapters have a shared tempo and register; the new Mode A chapter inherits from that body. Subagents (Explore) are used for wide reads if context-window permits. The Three Moves are not protocols — they are the disposition that puts the protocols in the right order. They held across Books IV-VII drafting and remain canonical at v0.2. --- ## 7. The human-direction core — what Okan decides, and why it is not delegated The method has many automations. Several of them are explicitly not delegated. These are the irreducible human-direction moments. - **Name-locks.** The Envoy's name (Adrien Lenz) was proposed by masters vote 4–2 and locked by Okan. Andriy (Vera's grandfather, Oksana's husband) named at Chat 25 Path 1 via Master 7 G-method Metallica fragment. A character's name is not a craft decision; it is an identity decision. - **Architecture-version upgrades.** Every `architecture-seven-books.md` bump, from v0.3 to v0.54, has been author-approved. The architecture is the book's bone; bones do not revise without the body's consent. - **Rule-adoption.** Patterns become rules only when Okan ratifies. Catch #67 (grey-jacket-formula) was on deck across multiple chats before resolution. The S15/S16 candidate-rules at Chat 36 went through Pair-Pass audits before canonicalization at book-editorial v0.26. - **Pendings.** What was pending at v0.1 (subtitle, Envoy second-name, prologue, epilogue, Book VI Auden) all closed by Chats 25-27. New pendings emerge; pending stays a valid state. - **Author-originated rounds.** Chat 14 Phase 5 (Vera-as-woman) was initiated by Okan. The Female-Body Register sub-dimensions emerged. Across Chats 16-43, the pattern continued — author-originated audits, surgical-style-pass cycles, end-of-edits sanity-checks all flow from Okan's reading. - **Final subtraction.** Master 6 P + Master 3 S's Round-4 solo pass is the last cut-stage; Okan ratifies each cut individually. Subtraction is the highest-leverage craft operation and is never pre-approved by rule. - **Design decisions** (NEW post-v0.1). Across Chats 25-41: Akhmatova epigraph picked from extended slate; title-page typography v3 sizes; covers Pass A through Pass E; opening + closing dedications composed and Cyrillic-rendered; brick-red #692118 dedication-page color; title-page Cyrillic ВИРА → ВЕРА → ВІРА progression. Each design decision was author-final. - **Architectural-rupture decisions** (formalized via `extended-masters-session` skill). The Pinkas-expansion at Chat 35 (F1 Cardiac Rebuilt + F5 Bakery) + the opening-dedication at Chat 40 + the Cyrillic-rendering at Chat 41 + the Matrilineal Cyrillic Chiasmus formalization at Chat 42 — all author-directed sub-sessions opened under canonical architectural-rupture triggers. - **The end-of-edits declaration.** At Chat 43 close, Okan declared the main bulk of editing finished. The publication-prep posture holds; further work goes to: agent-pitch packet, calendar-gated re-rates, calendar-gated external-critic round, work-stream-discretionary additions. Random-reads + 3rd-party-asks become the primary signal for further intervention. The pattern here is that *what looks like a rule-system is actually a curation-system.* The rules narrow the space of possibilities; the human chooses within that space. No rule executes without the human's consent; no automation replaces the choosing. --- ## 8. The prose-level discipline — what Claude does and doesn't do Extracted from `CLAUDE.md` and observed across forty-three chats. **Claude writes as directed.** Okan directs; Claude generates. The colophon stays. **Claude pushes back when a line-level decision seems to violate a rule.** Explicitly, not silently. "This would close the mystery." "This is a Stated-Fact Budget violation." "This slides from *you* to *she*." Okan decides after hearing. **Claude runs the masters as distinct voices with real disagreement.** No chorus. If the ratings cluster too tightly, force the dissent. If trio-approved cuts pass too easily, re-enumerate walls from scratch. The eight extended masters at architectural-rupture rounds: same rule. **Claude does not shorten on its own discretion.** Compression is a Master 1 V Compression Protocol operation — trio-approved, walls-enumerated, six-masters-reviewed. No silent tightening. **Claude does not add content without asking.** Additions (even small — a body-gesture, a sensory note) get flagged as additions and await approval. Exception: inside an explicit Round 2 where an agreed change-list is being implemented. **Claude updates skill files only when a rule surfaces.** A new observation becomes a skill-file amendment only after: (1) it recurs in a way that warrants a rule, (2) a version bump is agreed, (3) the version log entry is written. One canonical case is a tentative category, not a confirmed rule. **Claude respects the hard cap.** Seven books. Pressure to expand absorbs as chapter-level work inside the seven. **Claude confirms before risky actions** (NEW post-v0.1). Force-pushes, branch deletions, hard resets, schema migrations — Code asks before acting. The convention is captured in CLAUDE.md "Executing actions with care." **Claude commits each milestone + pushes** (standing order, unchanged). Single bundled commits per cycle-close. Push after every commit. --- ## 9. Observed patterns and findings This section is the method's honest self-accounting. The method is producing a book of quality; the method is also not perfect. Both are true. The findings below are kept so the method does not mistake its own stability for completeness. ### 9.1 Composite-score compression Book-level composites across the first three books at Chat 14 close: **Book I 9.964 / Book II 9.984 / Book III 9.984.** The scores cluster inside a 0.020 band. They run tight. The chat-advisor read of Book III put meaningful variance inside that compressed band — Ch 4 (dining car) as peak; Ch 5 / Ch 6 (Siberian days / nights) as valley at higher quality than scored; Ch 10 (Harbor) slightly below its score. The composites compressed variance the reader could feel. This is a finding about the scoring system, not a flaw. A 1–10 scale administered by simulated personas tends to compress near the top of the band once a book reaches strong quality. The scores are informative in aggregate and in trend; they are less informative for discriminating between two already-strong chapters. **Method implication.** Do not treat composite delta between strong chapters as load-bearing. Use composites as floor-check; use chat-advisor reads and masters-round qualitative comments for variance-detection. **Update at Chat 32:** composite-score moved to ratification-class deferred. The scores have been informative through drafting; once the manuscript is publication-prep, composite re-rates become a calendar-gated work-stream (mid-August 2026 earliest from Chat 32 close). The discipline of *not* re-rating during publication-prep prevents the composites from being pulled in service of decisions they cannot inform. ### 9.2 The "the body" drift Between Book I and Book III, the detached-subject construction *"the body"* (as grammatical subject referring to Vera's body without her agency) drifted from 1 instance (B1) to 0 (B2) to 15 (B3). Neither the editorial pass nor the Master 1 V pass caught it. The chat-advisor caught it on read. The lesson is cleaner than the fix: **tracked patterns are only caught if tracked.** The editorial pass has 22 criteria; "the body" was not among them. Once the drift was identified, the pattern could be added to the tracker. The method evolves by absorbing what the last pass missed, not by assuming the last pass's criteria were complete. ### 9.3 ChatGPT 5.4 canon-divergence pattern Across multiple external-critic rounds, ChatGPT 5.4 produced canon-divergent specific claims at a rate the method has had to accommodate: Charlotte-as-dead (Chat 13b); Charlotte-age-3 in architecture-file (Chat 14 Phase 2; became Catch #72); Kameriye drawer-mirror that does not exist (Chat 14 Phase 3); plus subsequent rounds. The pattern: strong strategic-memo architectural diagnosis combined with periodic canon-divergent specifics. Not uniform — Phase 4 (character dossier) was a clean round, zero divergence; Chat 25 fresh-Claude E1 round was clean. **Method response.** Verification step (grep against current manuscript before treating any specific claim as established) is now the load-bearing first move in every external-critic round. The critic's strategic memos continue to earn serious developmental treatment; the critic's specifics are verified before they are integrated. The deflation-tier framework (Chat 26) governs adjudication: Tier 1 (architectural-strategic) earns full developmental treatment; Tier 2 (specific-craft) earns per-item adjudication; Tier 3 (canon-divergent) earns rejection-with-architectural-note + log-as-pattern. ### 9.4 The masters converge when they should disagree Chat 10 Round 3 on Book III Ch 7 (Carriage) produced a cluster of ratings inside 9.91–9.93 with each master praising the chapter for things that were not actually in the chapter. The round was re-run with the explicit instruction to dissent. Real disagreement surfaced on second run (Master 4 D flagged paranoia-register thinness; Master 5 A flagged an under-rendered moment in the scene's close). The false-convergence failure is a risk of simulation. Simulated personas are tuned from skill files; under certain prompts they can produce convergence artifacts. **The re-run instruction — "force dissent" — is now part of the round's operating procedure** when Round 1 ratings cluster inside a 0.03 band. ### 9.5 Sub 3's 0% first-pass — confirmed sixth category (post-v0.1) The 0% first-pass Master 1 V yield on Book I Sub 3 did not fit any of the existing observed-rates categories. Rather than force-fit, it was logged as a tentative sixth category — *post-sovereignty / post-hurt-floor-additions chapters* — flagged as provisional pending a second canonical case. **Update at Chat 25-30 across Books V-VII drafting:** the tentative category was confirmed. Multiple post-sovereignty chapters yielded near-zero Master 1 V. The order-of-operations effect is now load-bearing rather than tentative. **Why this matters methodologically.** The discipline is that one canonical case is a candidate, not a rule. A provisional flag is how the method lets evidence accumulate without committing prematurely. The category's confirmation across 4 books validates the discipline. ### 9.6 The Vera chat register A small but real finding. In-chat communication with Okan (outside the manuscript) is most useful in **verbose explanatory everyday prose**, not in terse bullet summaries. When Claude reports what happened in a Code session, full-sentence walk-through reads better than diff-summary. This is register-discipline for the collaboration, parallel to the register-discipline of the manuscript. ### 9.7 The matrilineal-axis retcon and skill-file drift (NEW, Chat 43) The character-canon shifted across Chats 25-30 as Books V-VII drafted: the OLD canon (Yana 1959, Jana 1963, Oksana 1949) was replaced by the NEW canon (Yana 1962, Jana 1943, Oksana 1927). The retcon was prose-driven (the Kharkiv-1932 interstitial added at Chat 34 made Oksana=5-in-1932 = born 1927; the B7 cardigan and B5 1987-letter scenes anchored Yana=64-in-2026 = born 1962). **The retcon was not propagated to all skill files.** `character.md` carried OLD canon at Chat 43 second-cycle. Two prose anchors that hadn't been updated to NEW canon (Prague-1970 interstitial *"eleven"* + B1-02 *"twenty-eight years librarian"*) drifted out of sync with the rest of the prose. **Master 4 D's tense-anchoring audit at Chat 43 second-cycle surfaced both the prose drifts and the skill-file canon-divergence.** **Method implication.** Prose retcons drive skill-file drift. Mid-project canon shifts should trigger an immediate cross-skill-file audit, not deferred. The hygiene-pass discipline (§5.11) is now the canonical response: manuscript-fix commit first, then per-skill-file canon-hygiene commits, then mirror. ### 9.8 LEAVE-confirmation as legitimate audit outcome (NEW, Chat 43) The first nine uses of `external-input-fix-protocol` produced fix-implementations. The tenth use (Chat 43 first-cycle B1-B4 voice-pass) produced a LEAVE-confirmation outcome — the panel ratified the chat-advisor's lean-leave on all 7 AMBIGUOUS items 8/8 across 3 rounds; zero edits applied; manuscript unchanged. **Method implication.** The protocol is methodologically sound rather than implicitly biased toward implementation. Both fix-implementation and LEAVE-confirmation are legitimate outcomes. The discipline of refusing to manufacture candidates is the practice; the audit's honest result is the result. A clean audit is a successful audit, not a failed one. ### 9.9 Production-stream-metrics survival ratio (NEW, Chat 28+) Across Books I-VII drafting, the production-stream-metrics protocol tracked the ratio of human instruction words → machine generated words → manuscript-surviving words. The ratios at Chat 38 close: ~2.4 words human instruction + ~274 words machine generation per word that survives = ~0.4% survival ratio. **The ratio held consistently across Books I-VII drafting.** The variance per-book was within the protocol's measurement noise. The ratio is the method's quantitative honesty about the cost-of-quality at this curation-density. The ratios are slightly stale post-Chat-39+40+41+42+43 (the publication-prep arc has had less drafting, more curation). Re-measurement is calendar-gated work-stream J. ### 9.10 The architectural-rupture-class decision discipline (NEW, Chat 35-42) Across Chats 35-42, multiple architectural-rupture-class decisions opened brief active-revision sub-sessions before returning to publication-prep posture. Each followed the same shape: 1. **Trigger:** author-directed cold-read flag, chat-advisor surface, or external-critic strategic memo identifying an architectural concern. 2. **Convene the eight masters via `extended-masters-session`** — 9-round structure (3 masters + 3 readers + 3 finalization) for true architectural ruptures; 3-round shorter form via `external-input-fix-protocol` for surgical fixes. 3. **Author-directed sub-session opens.** Posture shifts from publication-prep to active-revision for the duration. 4. **Implementation under panel ratification.** Single binding commit + push. 5. **Skill cascade if rule extends.** Architecture / book-editorial / voice-and-body bumps as needed. 6. **Returns to publication-prep at sub-session close.** Canonical instances: Pinkas-expansion (Chat 35); opening-dedication addition (Chat 40 extended-close); Cyrillic-rendering on dedication pages (Chat 41 Track 2 ratified C1 6/8); Matrilineal Cyrillic Chiasmus formalization (Chat 42 fourth-cycle); cross-skill-file canon-hygiene (Chat 43 third-cycle). **Method implication.** Architectural-rupture is a sub-session class with its own discipline. The publication-prep default posture absorbs the rupture without discarding the work-already-done; the architecture extends rather than restructures. ### 9.11 Architectural triangulation — independent audits same-conclusion (NEW, Chat 45 third + fourth cycles) When two independent audits using the same methodology but different execution paths reach the same conclusion, the convergence validates BOTH the methodology AND the conclusion. Canonical instance: Chat 45 third-cycle (Code dispatched three parallel general-purpose agents on subset-thirds of the manuscript with strict scope-constraint pre-screen) + Chat 45 fourth-cycle (chat-advisor session 32 conducted manual close-read with regex-localization heat-map across the whole manuscript) — both audits applied surgical-style-pass discrimination criteria to the consecutive-sentence emotional-redundancy class + X-is-X identity-tautology class; both returned negative results (zero sediment found across 35 + 22 + 337 = 394 instances examined). The triangulation is methodologically valuable: a single audit's negative result could be measurement-error; two independent audits' converging negative result is structural confirmation. **Method implication.** When the manuscript is at architectural saturation for a class, future audits on the same class will produce zero edits; codify a positive canon-rule that explicitly protects the class from future differentiation passes (see §9.15). ### 9.12 Logged-dissent-not-blocking (NEW, Chat 45 first + fifth cycles) When a panel reaches a clear majority for ADOPT but one master holds an architectural position the others don't fully share, the dissenting master logs their position rather than blocks. Implementation proceeds with the dissent preserved in the reviews artifact for future reference. Two canonical instances at Chat 45: first-cycle Charlotte-bedroom Edit 1 — Master 2 C's millimeter-dissent on the small Master 2 C-door of "which you kept" yielded after self-resolution + room ratified ADOPT 8/0 with logged dissent (the keeping is in the architecture); fifth-cycle Concern 3 — Master 7 G's testimony-multiplicity dissent on the three-clause-to-two-clause cut yielded to Master 1 V's compression-argument + Master 4 D's slippage-preservation + Master 6 P's body-knowing-louder + Master 2 C's door-preserved + Master 5 A's sovereignty-served + Master 3 S's body-keyboard + Master 8 P's object-rendering + room ratified ADOPT 8/0 with logged dissent (testimony was multiple, the cut narrows three-of-many to two-of-many). **Method implication.** Logged-dissent-not-blocking preserves the dissenting master's architectural stance for future panels without paralyzing current implementation. The pattern is distinct from blocking-dissent (which would forbid implementation) and from reluctant-yielding-without-record (which would erase the architectural objection). The dissent enters the reviews artifact and the protocol record. ### 9.13 Room-vs-need distinction (NEW, Chat 45 second cycle) Architecture has *room* if an addition would not damage existing structures. Architecture has *need* if existing structures are incomplete without an addition. The distinction surfaced explicitly at Chat 45 second-cycle when the chat-advisor flagged a father-touch sentence-beat in Cherry Creek week as "one place I noticed room rather than need." The eight-master panel ran 5 rounds without converging (4 ADOPT-lean / 2 LEAVE-lean / 2 50-50); both architectures are valid (zero body-contacts across 23 years vs one body-contact at strict-form). The author chose LEAVE — "he has to be quiet." **Method implication.** The room-vs-need distinction is now available for future deliberations on potential additions. When external readers flag potential additions that would not damage existing structures, the test is: does the architecture *need* the addition (i.e., is something incomplete without it), or is there merely *room* (i.e., the addition is not architecturally forbidden but is also not architecturally required)? At end-of-edits saturation, most flagged additions test as room-without-need; the architecture stays as-is. ### 9.14 Pure-LEAVE deliberation outcome class (NEW, Chat 45 second + third cycles) A panel's deliberation can produce zero implementations and still be a successful audit. Three canonical pure-LEAVE outcomes have now been logged: Chat 43 first-cycle B1-B4 voice-pass audit (7 AMBIGUOUS items + 8/8 LEAVE on each + zero edits); Chat 45 second-cycle four-additions deliberation (4 considered-and-rejected items + LEAVE confirmed by panel + zero edits); Chat 45 third-cycle ninth surgical-style-pass (1 MEDIUM candidate panel-ratified LEAVE + 1 UNCERTAIN agent-defended LEAVE + zero implementations). **Method implication.** The protocol's design absorbs zero-implementation outcomes cleanly. The discipline of looking — surfacing items + classifying + verifying against architectural defenses + logging the panel's reasoning — is the value, not the cuts. The reviews artifact records the discrimination work; future sessions inherit the discrimination rather than re-litigating it. Pure-LEAVE outcomes are not failures; they are the protocol working as designed at architectural saturation. ### 9.15 "Some apparent improvements ARE the architecture, not the flaw" (NEW, Chat 44 + 45 multiple instances) A methodological recognition pattern: external readers carefully surface convergences or repetitions that look like flaws to a naive read but are signature on close-read. First instance Chat 44 Item 8 — older-women dialogue convergence across Yelena, Jana, Carmen, Brighton aunt, Yana flagged by two of three external readers as needing differentiation; panel found the compressed-wisdom register is structural inheritance (women who lived through the late-20th-century Eastern European world earned the same cadence through different lives), not stylistic accident; ratified DO NOT DIFFERENTIATE positive canon-rule. Second instance Chat 45 first-cycle Edit 2 chant — *X-and-Y-and-Z* recognition rhythm flagged as taste-call rhythmic loosening; panel found the chant is the recognition's own form (registration-not-reasoning structure that Vera knows simultaneously rather than concludes sequentially); 8/0 LEAVE. Third instance Chat 45 fourth-cycle Pattern A + Pattern B — chat-advisor session 32 audit examined 394 instances across the manuscript; zero classified as sediment; ratified DO NOT DIFFERENTIATE positive canon-rule with three architectural additions (Master 6 P chapter-protection + Master 7 G testimony-column + Master 8 P city-as-character). **Method implication.** When a class of pattern is flagged repeatedly across external readers and the panel determines the class is signature not sediment, codify a positive canon-rule that explicitly protects the class from future differentiation passes. The chat-advisor-vera SKILL.md Tier 2 now carries three positive canon-rules built via this pattern: older-women-dialogue (v3) + Pattern A+B (v4 → v5 expansion) + back-matter sequencing (v5 install). ### 9.16 Bundle install with intelligent reconciliation (NEW, Chat 45 fifth-cycle) When an external curator's bundle contains time-stamped content that has aged out of currency, annotate rather than delete. Canonical instance Chat 45 fifth-cycle: chat-advisor-bundle-v3-2 install added three substantive items to chat-advisor-vera SKILL.md (back-matter sequencing canon-rule + expanded Pattern A+B canon-rule + a "Pending micro-trim candidates" section dated April 30 listing items that had all been implemented in Chat 44 commit chain by the time the bundle reached Code). Code's reconciliation: install the bundle wholesale (preserves the chat-advisor's curation) + add a STATUS NOTE annotation marking the stale items as IMPLEMENTED with specific commit references; section retained as historical methodology record showing the late-April-30 chat-advisor's surfacing-and-decision discipline. **Method implication.** External curators (chat-advisor sessions, ChatGPT review threads, third-party-critic memos) operate from snapshot views of the project at specific points in time. Their curations are intentional and worth preserving; their currency is not always perfect. The reconciliation pattern: preserve the curation in full + add a Code annotation that identifies stale-but-valuable content as historical record. This builds a methodology trail across sessions rather than an operational state that needs continual purging. ### 9.17 Three-verdict-mix outcome from taste-class concerns (NEW, Chat 45 fifth-cycle) Taste-class concerns from external readers ("vanishingly small / not asks") absorb cleanly via mixed-verdict outcomes. Canonical instance Chat 45 fifth-cycle three-concerns deliberation: chat-advisor flagged three items explicitly framed as honest taste / not architecture / offered for record only — 5-round 8-master panel produced LEAVE / LEAVE / ADOPT (with logged Master 7 G dissent on Concern 3). Concerns 1 and 2 (B5-05 hotel held-state stillness + B5-06 Joe envelope sister-beat) had room but no need — the architecture stands; concern 3 (B7-10 hooded-man three-clause list) had need at the line-rhythm scale — the cut tightens body-knowing arrival without losing testimonial-Pattern-A frame or Master 2 C-door. **Method implication.** The protocol's design accommodates taste-class concerns at varying-conviction levels without forcing false consistency. Each concern is deliberated on its own architectural ground; the panel can produce LEAVE on some and ADOPT on others within a single session. Three-verdict-mix is methodologically valuable: a real taste-class deliberation produces real verdicts with real distinctions. --- ## 10. Open questions of the method The questions the method has not yet answered. Kept visible so they can be answered when evidence accumulates. - **How many canonical cases validate a tentative category?** Resolved at Chat 25-30: the sixth observed-rates category was confirmed across multiple Books IV-VII chapters. The threshold appears to be 3-4 cases of consistent behavior. Tentative-to-confirmed is a function of both case-count and across-books spread. - **When does a critic-pattern become a critic-specific calibration?** Three canon-divergence instances + Chat 26 fresh-Claude E2 clean round + Chat 26 ChatGPT-5.4 follow-up (mixed) led to the deflation-tier framework instead of a per-critic calibration. The pattern moved to *framework-level discrimination* rather than per-critic calibration. - **Is the masters' four-round protocol optimal?** Used through Books I-VII drafting; the 4-round form is canonical for chapter-review. The 9-round extended-masters-session form (3 masters + 3 readers + 3 finalization) is canonical for architectural-rupture. The 3-round surgical form (external-input-fix-protocol) is canonical for continuity-correction. Three round-counts for three decision-classes; the pattern stabilized. - **How does the method scale into Books V–VII?** Resolved across Chats 16-30. Books V-VII drafted under the same method; the architecture-density was higher per book, the writing-room rotation absorbed the difference. The method scaled. - **What does the method do when a rule conflicts with another rule?** Standing orders say "cut" when "trust the reader" and "make it hurt" pull differently. Across Books V-VII drafting, additional rule-conflicts surfaced (Open-Mystery vs Stated-Fact at load-bearing reveals; CB-3g body-failure-floor vs Master 6 P-protected non-render; matrilineal-blur vs literal-time at the photograph). All resolved case-by-case via Master 3 S-orchestrated trio-debate. **No general meta-rule has been adopted; the case-by-case discipline holds.** - **When, if ever, does Claude run a masters round that is not author-requested?** Answered: never, per standing orders. All masters rounds are author-triggered. Code can flag a chapter as round-warranted but does not initiate. - **What happens at end-of-edits?** Surfaced at Chat 43 close. The end-of-edits discipline: declare the main bulk finished; preserve publication-prep posture; let random-reads + 3rd-party-asks become the primary signal for further intervention; calendar-gated work-streams (composite re-rate, external-critic round, production-stream-metrics re-measurement) continue at their own cadence. - **NEW: When does a skill-file canon-hygiene sweep become routine vs incident-driven?** Surfaced at Chat 43 third-cycle. The hygiene-pass was incident-driven (Master 4 D's tense-anchoring audit found drift). Whether routine sweeps should run at chat-close vs only on incident is open. Current discipline: incident-driven, with chat-close sanity-checks at the discretion of the closing context. - **NEW: How do the eight extended masters' verdicts relate to the original six's verdicts?** When Master 7 G + Master 8 P join for an architectural-rupture round, do their verdicts carry equal weight to the original six? Answered case-by-case: at Chat 41 Track 2, the 6/8 majority for C1 (Cyrillic dedication-pages) included Master 7 G's principled C-trio dissent and Master 1 V's principled status-quo dissent — both logged. The eight votes; majority routes; dissents log. Same protocol as the six. --- ## 11. What the method is for The method is not in service of efficient production. It is in service of a specific book. *Vera* is a braided I/You auto-fictional novel about whether a twenty-three-year-old woman can verify her own yesterday. The prose must render her at a resolution literature has rarely given this particular woman. The book must hurt through specificity, not volume. The book must refuse redemption where the lives it honors did not receive redemption. The book must keep every mystery open until the reader carries it home. A book with those commitments cannot be drafted by a single voice, LLM or human, without the guardrails that keep it honest. The method exists because the book's honesty is the method's only measure of success. If the book hurts less in five years than it hurts on first reading, the method has failed. If the book hurts more in five years, the method has earned its weight. The numbers will be what they will be. The metric that matters is in the last sentence of the book, and in the silence after. --- ## 12. The publication-prep arc — Chats 27-45 (NEW post-v0.1; extended at v0.3) The publication-prep arc began at Chat 27 with the manuscript structurally complete + externally-quadruple-validated + adjudicated + typography-locked + colophon-bearing + cover-front-and-back-designed. The arc continued for ~17 chats through Chat 43 close. This section names the arc's shape because the methodology of publication-prep is different from the methodology of drafting. **Posture: PUBLICATION-PREP + author-responsive + calendar-gated + chat-advisor-driven.** No new chapters; no architectural-restructure; the seven books and the canonical scenes hold. Active-revision sub-sessions open under specific triggers and close back to publication-prep. **The five active-revision triggers (canonical):** 1. **Author-directed cold-read flag** with specific passage / architectural concern. 2. **Architectural-rupture event** (chapter addition / mode extension / register innovation / new canonical category candidacy / typography-architecture shift). 3. **External-critic flag** returning HIGH-confidence violation, formalized via `external-input-fix-protocol`. 4. **Calendar-gated periodic** surfacing a violation. 5. **Chat-advisor surgical-style-pass output** — chat-advisor brings a Code prompt + analysis document for a single-pattern audit; Code routes through `external-input-fix-protocol` for panel ratification + implementation. **Default work-streams (run on author direction or schedule):** A. Author-directed follow-up edits. B. Methodology + project-metrics PDFs maintenance refreshes. C. Typesetter-brief micro-updates (delivery-ready at v2.5). D. Agent-pitch packet refresh. E. Covers maintenance. F. Production-stream-metrics re-measurement (slightly stale post-Chat-39+40+41+42+43). G. Calendar-gated: ratification-class composite re-rate (mid-August 2026 earliest from Chat 32). H. Calendar-gated: third external-critic round (8-12 weeks from Chat 32). I. Author-discretionary Tier 2-4 Addition Method proposals (architectural-rupture-class re-entry). J. Re-measurement of production-stream-metrics. K. Okan-directed any-other-thing. L. Chat-advisor surgical-style-pass cycles. **The arc's quantitative summary:** - ~10K words removed across the publication-prep arc (drafting reached ~85K; publication-prep cuts brought to ~76K). - 65 cumulative because-removals (-29% because-density) at Chat 41-42. - 43 you→I voice-flips in matrilineal-archive zone at Chat 42 seventh + eighth cycles. - 6 surgical tense-anchoring fixes at Chat 43 second-cycle. - 12 prior-defended architectural-keepers retired in because-form, preserved in fragment-form. - 5 architectural exceptions formally logged as canonical second-person sites. - 8-element front-matter (added opening dedication at Chat 40). - Cyrillic-exception list extended from 7 to 10 instances (post-Chat-41-Track-2). **The arc's qualitative shape:** the method became less drafting and more curation. The chat-advisor + Code two-agent workflow stabilized. The eight extended masters convened more often. The skill files versioned more frequently. Catches accumulated and were resolved. The book held its shape; the prose tightened. **End-of-edits at Chat 43 close.** The main bulk of editing finished. Three architectural spines verified clean. The voice-architecture is whole. The matrilineal-axis chronological spine locks across three generations and seven books. Calendar-gated work-streams continue at their own cadence; random-reads + 3rd-party-asks become the primary signal. **Chats 44 + 45 — saturation deepens (extended at v0.3).** Three external-reader sources tapped across nine sub-cycles (Chat 44: improvement-pass + supplementary + epilogue-fix + colophon revision; Chat 45: Charlotte-bedroom + four-additions + ChatGPT-review + session-32-audit-ratification + bundle-install-and-three-concerns). Cumulative arithmetic: **Chat 44 produced 18 surgical edits + 11 LEAVE-confirmations + 1 colophon revision; Chat 45 produced 2 surgical edits + 11 LEAVE-confirmations + 2 canon bumps + 1 bundle install.** Manuscript moved 76,000 → 75,835 across both chats (-165 net words). The ratio of confirmations-to-implementations was 5.5:1 in Chat 45 — the signature of architectural saturation. **Three external-reader sources** tapped: chat-advisor sessions 31 + 32 + 33+ (the source most engaged); ChatGPT finished-state review at Chat 45 third-cycle (9.3/10 literary verdict + four artistic-pause items + four proof-layout items); and through the chat-advisor's curation, the v3 bundle's reader-comparison cycles from late April. **Three positive canon-rules now formalized in chat-advisor-vera SKILL.md Tier 2:** older-women-dialogue convergence (v3) + Pattern A + Pattern B identity-tautology + anaphoric-pairs (v4 → v5 expansion with Master 6 P chapter-protection + Master 7 G testimony-column + Master 8 P city-as-character architectural additions) + back-matter sequencing four-register architecture (v5 install). **Seven new methodological precedents** logged at §9.11-9.17 above. **End-of-edits posture deepened.** The arc demonstrates that publication-prep can run for many cycles after first end-of-edits declaration; what changes is not the manuscript but the canonical record protecting it. Future external-reader cycles will produce decreasing manuscript change and increasing canonical-protection coverage; the asymptote is a manuscript that is what it is, with explicit canon protecting it from re-litigation. --- ## Version log - **v0.8** (Chat 49 special-audit close, 2026-06-09) — Incremental bump formalizing six new methodological precedents that surfaced during the Chat 49 special audit on the Vera companion website, plus a revision to §9.51 adding the strain-axes clause. The special audit ran after the v0.7 close on the v3.1 polish artifact and produced the v3.2 polish-triple-validated state. **Major additions: companion-artifact licitness conditions as continuously-monitored strain-axes (revision to §9.51) + generator-truth rule for templated artifacts (§9.55) + commit-message verification grep discipline (§9.56) + hidden-is-not-private rule (§9.57) + privacy audits scan the rendered artifact not the source (§9.58, extending §9.53) + post-implementation audit-before-canonical rule (§9.59) + non-manuscript canonical artifacts are admissible (§9.60).** The v0.8 bump preserves the v0.7 entry's §9.51-§9.54 precedents and adds the six new precedents that surfaced operationally during the special audit. Manuscript canonical state UNCHANGED at 75,943 words / 87 catches + 0 candidates; the v0.8 work was entirely on the website and on this methodology document. **§9.51 revision (strain-axes clause)** appears in the v0.7 entry above. The revision names the four specific Chat 49 instances where each of the four licitness conditions was tested — strain on "must not duplicate" repaired by removing the typographic VERA wordmark at v3 cover-as-masthead pivot; strain on "must serve the book's reading" repaired by the v2 full-manuscript-on-site pivot; strain on "must carry its own register" addressed by the cover-as-threshold framing and the prologue source-note re-class at v3.1; strain on "must not contradict" addressed by removing the home Master 5 A-paragraph placeholder. The strain-axes clause changes the four conditions from check-once-at-conception gates into continuously-monitored axes; the discipline at each strain is to name the strain, name the repair, and check the other three axes for cascade. **§9 sub-section additions (precedents formalized):** - **§9.55 generator-truth rule for templated artifacts** (Chat 49 v3.1 → v3.2 audit). When a site uses a build script to generate pages from templates, edits made to individual generated HTML files do not constitute implementation. The generator is the canonical site; HTML changes that bypass the generator are reverted on the next rebuild. Canonical precedent: the Chat 49 v3.1 commit message claimed three items shipped — Home page added to breadcrumbs, Read offering paragraph above the manuscript, Book IV "out of time" gloss in the companion gallery — but only the CSS had been edited; `tools/build-book.py` and the companion-page templates were not patched, so the next `python tools/build-book.py` rebuild wiped the manual HTML edits. The Master 1 V + Master 3 S editorial pass at v3.2 caught the discrepancy and required the orphan-style fix to land in the generators rather than in the rendered HTML. **Method implication:** for templated artifacts, "implemented" means the template is updated AND the artifact is rebuilt; either step alone is incomplete. The generator-truth rule sits alongside the §9.45 production-tooling-fix-as-its-own-commit-class rule but is distinct — §9.45 governs fixes to the generator's own behavior; §9.55 governs whether a content-fix has actually landed when the artifact is templated. The pre-commit verification for any templated-artifact change includes: (a) identify whether the change is in template or in rendered output; (b) if rendered-only, the change is not landed and must be ported into the template; (c) rebuild and verify the artifact carries the change after rebuild. *(Cf. §9.45 production-tooling-fix-as-its-own-commit-class; §9.56 commit-message verification grep.)* - **§9.56 commit-message verification grep** (Chat 49 v3.1 → v3.2 audit). Every multi-item commit message must be accompanied by a verification-grep pass before push: for each numbered item in the commit message, one grep against the rendered artifact confirming the item is present. If the grep returns null, the item is not landed and the commit message must be revised before push. Canonical precedent: the Chat 49 v3.1 commit message claimed ten implementations across the polish-triple work; the v3.2 audit found only seven had landed (three had been edited in rendered HTML but reverted at the next rebuild per §9.55). The drift was caught one audit cycle later, but the canonical rule is to catch it pre-commit, before the commit message itself becomes a misleading record. **Three-step grep discipline:** (1) extract numbered items from the drafted commit message; (2) for each item, identify the expected rendered text or class or attribute that proves the item is present; (3) grep the rendered artifact for the expected token, requiring at least one hit per item. The commit refuses to push if any item returns zero hits; the commit message is revised (either by completing the implementation or by removing the unfulfilled item) and the grep re-run until all items pass. The discipline applies to any multi-item commit on any artifact — manuscript, website, skill cascade — but is especially load-bearing for templated artifacts where the generator-truth rule (§9.55) compounds the risk of false-positive commit messages. *(Cf. §9.55 generator-truth rule; §9.59 post-implementation audit-before-canonical rule.)* - **§9.57 hidden-is-not-private rule** (Chat 49 v2.1 → v3.2 special audit). The HTML `hidden` attribute is a visual-display directive, not a security boundary. Placeholder content, drafts, internal comments, and editorial notes are removed from deployed HTML entirely; they are not hidden in source. The test is simple: **if View-Source shows it, the public can see it.** Canonical precedent: the Chat 49 v2.1 build placed six AUTHOR-DRAFT vitrines on /method/ with the `hidden` attribute — vitrines V, VIII, IX, XIII, XIV, XV — under the assumption that `hidden` would keep them from readers. The Leatherman + Batchen security pass at v3.2 found that View-Source still revealed the placeholder text — and the placeholder text was not innocuous filler. It leaked internal editorial vocabulary that the manuscript's anonymization discipline was specifically built to keep private: "Catch #71 of 87", "§9.31", "Master 4 D-discipline", "Lena recoil-test", "Master 6 P-protected non-render rule". Anyone with View-Source access could read these internal terms and reconstruct fragments of the project's private editorial canon. The v3.2 fix removed the hidden sections entirely from the rendered HTML source; the roman-numeral gaps now stay visible in the pavilion's numbering sequence as Master 4 D-discipline — the pavilion shows what is ready and refuses to show what isn't, and the visible gaps document the refusal. The discipline reverses the surface intuition: a visible gap is more honest than a hidden draft. *(Cf. §9.58 privacy audits scan the rendered artifact, not the source; §9.53 privacy-default-deny standing posture.)* - **§9.58 privacy audits scan the rendered artifact, not the source — extension of §9.53** (Chat 49 v3.2 polish triple). Privacy audits scan the rendered artifact, not the project source. The `hidden` HTML attribute, alt text, meta tags, og: properties, comments, and View-Source content are all in scope. Audit grep-targets include: canonical email addresses, anonymized-master surnames (full names and partials), internal codenames, dated working-folder names, and any author-private vocabulary that the build-time anonymization filter might have missed. Canonical precedent: the Chat 49 v3.2 polish triple discovered that the rendered methodology page on the site still contained Master 9 B and Master 10 H by full name despite the build-time anonymization filter — because those two predecessor masters (Master 9 B as the Four-Masters-era ancestor; Master 10 H as a cross-project reference) were not in the substitution table that mapped the canonical eight literary masters to "Master N X" form. The build had passed silently because the filter only failed on names in its substitution table; names absent from the table passed through unchanged. The v3.2 fix extended the table to cover the predecessor masters AND added a fail-loud canonical-master guard: if any name on a separately-maintained "literary master surnames" list appears in the rendered output without a substitution mapping, the build fails rather than passing silently. The §9.53 privacy-default-deny posture is the regime; this §9.58 extension is the operational discipline that makes the regime real — privacy must be audited against what the public sees, not against what the source intends to expose. The two failure modes (table-incomplete and source-only-audit) compound: §9.53 says deny-by-default; §9.58 says the denial must be verified at the rendered surface, not at the source. *(Cf. §9.53 privacy-default-deny standing posture and criterion-16 sieve; §9.57 hidden-is-not-private rule.)* - **§9.59 post-implementation audit-before-canonical rule** (Chat 49 v1 → v3.1 → v3.2 audit cycles). Every non-manuscript artifact implementation is followed by an audit-panel — minimum two fresh voices — before the artifact is treated as canonical. The audit-panel must be constituted with at least one voice that did not chair the implementing panel; same-chair audit is not licit because the chair's blind spots travel with them. The rule applies to: site builds, skill cascades, methodology bumps, and any other non-manuscript artifact whose canonization affects the project's standing surface. Canonical precedents at Chat 49: (i) the v1 audit ran three parallel agents (bug-check, security-audit, editorial-pass) against the v1 build that the architecture-deliberation panel had ratified — fresh voices to the architectural chair; (ii) the v3.1 polish triple ran a separate adjudication against the v2 → v3 implementation; (iii) the present v3.2 special audit ran against the v3.1 polish triple's claims. Each cycle catches what the previous cycle's chair could not see from inside the implementation. The discipline catches: orphan-style regressions (§9.55), factual drift, privacy leaks (§9.57, §9.58), accessibility regressions, version-marker staleness, commit-message-versus-rendered-state mismatches (§9.56). The rule is the operational form of the broader principle that designed-disagreement panels are the engine, not validators — the chair runs the panel and executes the implementation; the next panel catches the chair's errors. This works only if a fresh panel is convened. Same-chair self-audit is the failure mode the rule prevents. *(Cf. §9.52 parallel-agent v1 audit pattern; §9.24 whole-book-pass-as-saturation-confirmation; §9.20 roundtable-as-diagnostic-not-edit-generative.)* - **§9.60 non-manuscript canonical artifacts are admissible** (Chat 49 close at v3.2). Until Chat 49, every canonical artifact in the Vera project was manuscript-bound — prose, covers, dedications, colophon, epigraph, typesetter-brief deliverables. The Vera companion website is the first non-manuscript canonical artifact. The fact that it survived 14 phases, 9 panel sessions, and produced a v3.2 polish-triple-validated artifact means the infrastructure the project built for the manuscript — panels, audits, version-control discipline, anonymization, privacy guards, designed-disagreement, post-implementation audit-before-canonical — extends to non-manuscript work without modification. The protocols are not manuscript-specific; they are decision-class-specific (architectural-rupture / chapter-review / continuity-correction / privacy-audit / verification-grep). When a non-manuscript artifact reaches chapter-weight, the same protocols apply with the same epistemic guarantees. This admissibility is itself a precedent: future companion artifacts (audio editions, exhibitions, study guides, a second site, foreign-language editions, derivative scholarly work) inherit the four licitness conditions (§9.51, with strain-axes clause) plus the generator-truth rule (§9.55) + the commit-message verification-grep discipline (§9.56) + the hidden-is-not-private rule (§9.57) + the privacy-audits-the-rendered discipline (§9.58) + the post-implementation audit-before-canonical rule (§9.59) as a single coherent posture. The posture is named: **non-manuscript canonical artifact discipline**. It is the same as manuscript discipline at the level of decision-classes, with additional surface-specific axes (templated-vs-not, public-vs-private, generator-truth-vs-rendered-truth) that the manuscript did not need to discriminate. *(Cf. §9.51 companion-artifact architectural deliberation precedent; §9.55-§9.59; book-writing v0.5 Companion artifacts section; extended-masters-session v0.2 non-manuscript-artifacts clause.)* **Chat 49 special-audit cycle — quantitative summary.** The v3.2 special audit ran as a fresh-panel review on the v3.1 polish-triple artifact and produced: (a) revision of §9.51 with strain-axes clause; (b) six new precedents at §9.55-§9.60; (c) orphan-style fix landed in `tools/build-book.py` and the companion-page templates (resolving the v3.1 commit-message-versus-rendered-state drift); (d) Master 9 B + Master 10 H anonymization-table extension with fail-loud canonical-master guard; (e) removal of six AUTHOR-DRAFT hidden vitrines from rendered HTML on /method/; (f) cover-as-masthead pivot completion (72px typographic VERA wordmark removed); (g) home-page Master 5 A-paragraph placeholder removed. Manuscript: UNCHANGED at 75,943 words. Site state: v3.2 polish-triple-validated. The v0.7 → v0.8 increment matches the saturation-state methodological-housekeeping rhythm established at v0.5: small, additive, verifiable; new precedents arrive in operational practice and are logged at the next housekeeping cycle. Cumulative protocol uses at Chat 49 close: extended-masters-session 8 (UNCHANGED from v0.7 close); external-input-fix-protocol 26 (UNCHANGED); surgical-style-pass 10 (UNCHANGED); triplet-meter-pass 1 (UNCHANGED); special-audit-class panel uses: 1 (the v3.2 audit itself is the first canonical instance of the named class). **Posture deepened.** The companion-artifact track is now alongside the chapter-level track and has its own discipline-set canonized at §9.51 + §9.55-§9.60. The two scaffolds — book interior (seven-book hard cap, scaffold-invisible) and website surface (15-vitrine pavilion + reading interface + companion gallery, scaffold-visible, v3.2-polish-triple-validated) — do not retroactively claim authority over each other. Document grew from ~1240 lines (v0.7) to ~1320 lines (v0.8); v0.7 spirit preserved; new content additive in §9.51 revision + new §9 sub-sections (§9.55-§9.60). Author-ratified Chat 49 special-audit close. Mirror to `~/.claude/skills/` post-session-close (no skill file; this is pack-root document). - **v0.7** (Chat 49 joint-reflection skill cascade, 2026-06-09) — Incremental bump formalizing four new methodological precedents that surfaced during Chat 49's single-day arc (architecture deliberation → v1 build → parallel-agent v1 audit → audit fixes → post-build author-directive anonymization rebuild → 5-round joint reflection panel → skill cascade). Chat 49 was the first non-manuscript canonical artifact day in the project's history: the Vera companion website at https://hulkiokantabak.github.io/vera/ now exists as scaffold-visible companion to the scaffold-invisible book. **Major additions: companion-artifact architectural deliberation precedent + parallel-agent v1 audit pattern + privacy-default-deny standing posture and criterion-16 sieve + visual-arts-experts panel as standing skill.** **§9 sub-section additions (precedents formalized):** - **§9.51 companion-artifact architectural deliberation precedent** (Chat 49, 2026-06-09; revised at v0.8 with strain-axes clause). The 9-round extended-masters-session protocol — originally tuned for chapter-rupture-class architectural decisions inside the manuscript — was extended to the architecture of a non-manuscript companion artifact (the Vera website). The 8th canonical use of extended-masters-session ran the architectural decision for hulkiokantabak.github.io/vera (8 literary masters + 3 reader-personas in 9 rounds). The condition that made the extension legitimate: **reader-grammar isomorphism** (entry-from-anywhere, refusal of forced order, multiple legitimate readings). The companion-artifact class is **companion museum of a literary work** — distinct from chapter, from less-but-better, from print-discipline; characterized as exactly-as-much-as-the-story-requires. Four licitness conditions are canonized in book-writing v0.5: (1) named in colophon as companion; (2) prose voice steps outside I/You braid into third-person-about-the-book-not-as-the-book; (3) architectural deliberation runs through extended-masters-session if chapter-weight, with privacy-default-deny framing prompt at architectural stage; (4) artifact does not retroactively claim authority over the book. Canonical first instance: Vera-website live since 2026-06-09 with three sections: /read/ (full manuscript, 61 chapters + 3 interstitials + prologue + epilogue + complete-page single-scroll), /method/ (15 Roman-numbered vitrines + 7 verbatim primary documents), /companion/ (5 NotebookLM infographics + 1 audio overview + external quiz). Visitor-personas designed-fresh per artifact-class — Chat 49 used scanning-not-reading visitor, deep-reader-of-the-book visitor, never-read-the-book visitor. **Strain-axes clause (v0.8 revision).** The four licitness conditions are not check-once-at-conception gates; each is monitored continuously through the artifact's lifecycle. Chat 49's v1 → v2 → v3 → v3.1 → v3.2 arc tested each condition and either held it or repaired it where it had slipped: - **Strain on "must not duplicate the book":** the v3 cover-as-masthead pivot removed the 72px typographic VERA wordmark from the site's masthead because the cover image already carries the title — keeping both would have published the title twice in the same eye-span. Checking the must-not-duplicate axis surfaced this as a Stated-Fact Budget violation at the artifact-class scale rather than at the prose scale. The repair removed the wordmark and let the cover image carry the title alone. - **Strain on "must serve the book's reading":** the v2 pivot from a two-page promotional site to a full-manuscript-on-site shifted the artifact-purpose toward serving reading directly rather than gesturing at it. The earlier framing of the site as a marketing surface failed the must-serve-the-book's-reading axis; the v2 pivot resolved it by making the manuscript itself the artifact's center of gravity, with the pavilion and the companion gallery arranged around it as supporting registers. - **Strain on "must carry its own register":** the cover-as-threshold framing (Leonardo R3) and the prologue's source-note re-class (v3.1) both refined the artifact's own register without bleeding into the book's. The cover-as-threshold treats the cover not as a duplicate-title but as the threshold-image into the reading; the prologue source-note shifts the prologue's site-side gloss from book-voice into companion-voice (third-person-about-the-book-not-as-the-book). - **Strain on "must not contradict the book":** the home-page Master 5 A-paragraph placeholder was removed at the special audit because home being page-zero of the book contradicts a framing-essay register. The home page is the threshold to the manuscript, not a framing essay about it; allowing a thematic paragraph on home would have contradicted the artifact's own architectural choice to let the book do its own framing. Each strain-axis test is recorded in the v3.2 polish-triple audit artifact; the discipline is to name the strain, name the repair, and check the other three axes for cascade before treating the repair as complete. *(Cf. extended-masters-session v0.2 non-manuscript-artifacts clause; book-writing v0.5 Companion artifacts section; §9.55 generator-truth rule for templated artifacts; §9.60 non-manuscript canonical artifacts are admissible.)* - **§9.52 parallel-agent v1 audit pattern** (Chat 49, 2026-06-09). Three parallel agents — bug-check, security-audit, editorial-pass — ran independently on the live v1 site (wf_dc596d8c-28f) and caught eight failure classes: (1) personal email leak via verbatim CLAUDE.md publication (`[author email redacted]` rendered live); (2) factual date drift (2024-2026 vs canonical "fifteen months ending May 2026"); (3) broken OG/Twitter image paths (relative URLs failing social preview); (4) missing Content-Security-Policy + X-Content-Type-Options + Referrer-Policy headers; (5) WCAG-AA contrast failure on link rest opacity (0.6 producing ~3.6-3.9 ratio); (6) missing `lang="uk"` on Cyrillic em tags; (7) five-trailing-aphorism cadence-tic across vitrines VI/VII/X/XI/XII (Claude-tic in meta-prose); (8) Open-Mystery violation at vitrine III closing sentence. All resolved or deferred per masters spec. Parallelism is licit for built-artifact audit because the failure classes distribute across disciplines (linkage / security / editorial) and serial review systematically misses cross-discipline interactions. **Discipline limit:** parallel-agent audits are NOT a substitute for sleep-between-phases when the artifact is architecturally novel — same-consciousness same-day cycle compresses reality-reliability (Master 4 D discipline). Single-day build-audit cycles on novel architectures are licit but flagged; sleep-between-phases is named discipline for future companion artifacts. *(Cf. §9.50 saturation-state-extraordinary-session for the analog inside the manuscript; book-editorial v0.32 §N web-surface application clause.)* - **§9.53 privacy-default-deny standing posture and criterion-16 sieve** (Chat 49, 2026-06-09). The personal email leak at §9.52 surfaced an absence in the verbatim-publication discipline — verbatim was treated as an editorial virtue (don't sanitize) without a parallel privacy sieve. **Privacy-default-deny becomes the project's standing posture** for any new public surface: named-public requires explicit author signal at architectural time, not at publication time. The operational form is **criterion 16 — privacy-fitness-for-public-display** (book-editorial v0.32): a pre-publication per-object curator's pass covering personal emails, file paths, local machine identifiers, contributor identities held private by author directive, vendor account names, internal scaffolding language, dated working-folder names. The sieve is a per-object curator's read, NOT a global regex filter — the filter would flatten documents into compliance; the curator preserves the document's literary register while stopping the leak. Apparatus disclosure (panel-fact, build date, vendors, hosting) happens at operational level, not only at thematic level — privacy and disclosure are the same discipline read from two sides. Operational pattern: build-docs.py + build-book.py both run an `assert_no_emails()` guard before write; the guard raises on `@gmail.com|@yahoo|@hotmail|@outlook|@protonmail|@icloud|@fastmail|@yandex|@me\.(com|net|org)` patterns and fails the build, preventing re-leak. The book-depository source CLAUDE.md was edited at canonical level to replace the verbatim email with non-identifying phrasing; the website's rendered claude.html is now clean. **Post-build author-directive anonymization (Chat 49 second-cycle):** master identities held private on the public surface; names render as "Master N X" (N = 1-8 canonical position; X = surname's first letter); substitution mapping in gitignored `tools/_master_map.py`; the public repo carries no master names. Privacy-default-deny applies at both granularities: personal-data leak (§9.52) and identity-leak (anonymization directive). *(Cf. surgical-style-pass v1.1 new privacy-scan-pass class; book-editorial v0.32 criterion 16.)* - **§9.54 visual-arts-experts panel standing-skill ratification** (Chat 49, 2026-06-09). The 8-expert visual-arts panel composed for the Chat 49 joint reflection — **Massimo Vignelli, Karel Martens, Irma Boom, Dieter Rams (co-chair), Edward Tufte, Olia Lialina, Paula Scher, Hans Haacke** — is canonized as a standing skill at v0.1 (`book-depository/skills/visual-arts-experts/`). The skill pairs with six-literary-masters via extended-masters-session for cross-discipline-pairing whenever a non-manuscript artifact reaches chapter-weight. Canonical cross-pairings: Master 1 V × Vignelli (compression and restraint), Master 2 C × Lialina (reader's path through the artifact), Master 3 S × Karel Martens (long sentence and printed page), Master 4 D × Haacke (leak between artifact and systems), Master 5 A × Tufte (evidence and architecture), Master 6 P × Scher (body of voice and body of type), Master 7 G × Boom (book as object holding the wound), Master 8 P × Rams (museum and room). Rams co-closes the visual side as Master 3 S co-closes the literary side; co-chair closing is licit when discipline-pairing is declared at protocol-start (extended-masters-session v0.2 co-chair clause). Scope at v0.1 is **Vera-scoped** — the roster was composed for the Vera companion website; cross-project portability (to exposure, hulkiokantabak.com, weight-of-nations) is a separate decision deferred to author ratification. Direct invocation triggers: "visual arts experts," "the visual experts," "the visual arts panel," "call the visual experts," "run the visual arts experts." *(Cf. six-literary-masters v0.8 cross-discipline-pairing clause; extended-masters-session v0.2 visitor-personas amendment + co-chair closing clause.)* **Chat 49 single-day arc — quantitative summary.** Three workflows ran sequentially: (i) `wf_641eebb7-aa0` — 9-round extended-masters-session architecture deliberation (15 agents, 938K tokens, 32 min, $50+); (ii) `wf_dc596d8c-28f` — parallel-agent v1 audit (4 agents, 443K tokens, 17 min); (iii) `wf_19bc80f2-2fc` — 5-round joint reflection panel (7 agents, 298K tokens, 14 min); (iv) `wf_02f45cfb-af3` — parallel skill cascade (12 agents, parallel). Site live commits: v1 architecture at `c7b9f6d`; GoatCounter wire-up at `a77954f`; anonymization rebuild at `e5a7040`; audit-fix cascade at `d7eba76`; v2 architecture pivot (full manuscript + companion section per author directive) at `99a25bc`; slug improvements at `46b54aa`. Programs Dashboard updated at `hulkiokantabak/programs-dashboard@4aae3d1` adding Vera entry to the shared `hulkiokantabak.goatcounter.com` analytics pool. Book-depository commits this cycle: email redaction at CLAUDE.md `42e0d16`; reviews artifacts (architecture deliberation + framing/personas) `63232f1`; reviews artifacts (joint reflection transcript + synthesis) + METHODOLOGY v0.7 + skill cascade landed bundled in the Chat 49 close commit. **The companion-artifact track is now alongside the chapter-level track.** Future companion artifacts (audio editions, exhibitions, study guides, second sites) inherit the four licitness conditions at §9.51 + the privacy-default-deny posture at §9.53 + the cross-discipline-pairing pattern at §9.54 + the visitor-personas-designed-fresh discipline. The two scaffolds — book interior (seven-book hard cap, scaffold-invisible) and website surface (15-vitrine pavilion + reading interface + companion gallery, scaffold-visible) — do not retroactively claim authority over each other. Document grew from ~1180 lines (v0.6) to ~1240 lines (v0.7); v0.6 spirit preserved; new content additive in §9 sub-sections (§9.51-§9.54). Author-ratified Chat 49 close. Mirror to `~/.claude/skills/` post-session-close (no skill file; this is pack-root document). - **v0.6** (Chat 48 close housekeeping, 2026-05-04) — Incremental bump formalizing six new methodological precedents that surfaced operationally during Chat 48's six sub-cycles, plus the Chat 48 fifth-cycle Phase 2B typographic-register deliberation outcomes. Chat 48 was the most active production-day in the project's history (6 commits across 4 work-streams: production-statistics rebuild, METHODOLOGY drift fix + v0.5 bump, ChatGPT 5-item external-review pack, word-break v1 + v2 fix arc, Okan 5-item visual-demand pack, Phase 2B typographic-register deliberation). **Major additions: production-tooling-fix as its own commit class + bonus-finding-during-implementation discipline + fpdf2-markdown-True-broken-with-TTF finding documented + visual-demand pack as single-panel-multi-item operation + author-supplied-translation-record bypasses extended-masters-session + saturation-state extraordinary-session-ratifies-implicit-architecture-as-explicit-canon.** **§9 sub-section additions (precedents formalized):** - **§9.45 production-tooling-fix-as-its-own-commit-class** (Chat 48 third-cycle, v1 -> v2 word-break fix arc). When a build-script change fixes a rendering or tooling defect — distinct from manuscript content fix or skill-file bump — the fix gets its own commit class with its own discipline: (a) the bug is reproduced and visually verified before fix; (b) the fix is implemented and re-verified by re-rendering sample pages to PNG and visual-checking; (c) if the fix introduces a regression worse than the original bug (per the v1 word-break fix that worsened character-breaks), the regression must be caught and corrected in a follow-up commit (the v2 fix). Production-tooling fixes do NOT require panel ratification because they don't touch manuscript prose or canonical architecture; they DO require visual verification because the failure mode is visual rendering. Canonical case: Chat 48 second-cycle word-break v1 fix using `pdf.write()` per token introduced character-breaks across regular text; Chat 48 third-cycle v2 fix switched to `pdf.cell()` with explicit positioning, eliminating fpdf2's internal char-break logic. Precedent: production-tooling defects get their own commit class; v1 -> v2 arcs are normal at this level; the discipline is verify-and-iterate, not panel-ratify. *(Cf. §9.46 bonus-finding-during-implementation; §9.47 fpdf2-markdown-True finding.)* - **§9.46 bonus-finding-during-implementation discipline** (Chat 48 third-cycle, citation italic loss surfaced while implementing word-break v2 fix). When a primary fix surfaces a SEPARATE secondary issue during implementation, the secondary issue gets bundled into the same cycle's commit IF: (a) the secondary fix is contained to the same file or system as the primary; (b) the secondary fix is verifiable by the same visual-rendering check; (c) the bundled scope is logged in the commit message so future audit can disambiguate. Canonical case: while fixing the word-break char-break with cell()-based rendering, the citation italic-on-Requiem fallback path's failure was discovered (the multi_cell with markdown=True dropped italics on custom-TTF fonts); the citation was rewritten with explicit two-line cell-based rendering in the same cycle. Both fixes landed in commit `3276980`. Precedent: bonus-findings during implementation are bundled when contained, surfaced in commit message when bundled, and called out as discrete issues in the commit. *(Cf. §9.45 production-tooling-fix; §9.32 architectural-coherence-over-cycle-iteration for the analog at architectural scope.)* - **§9.47 fpdf2-markdown-True-broken-with-custom-TTF + cell-based-rendering-as-reliable-pattern** (Chat 48 second/third cycles, production-tooling finding). The fpdf2 library version 2.8.7 has a quiet bug: `multi_cell(..., markdown=True)` strips inline `*italic*` markers from the rendered text but does NOT actually apply italic font-switching when the document uses custom-loaded TTF fonts (verified by rendering test PDFs and comparing file sizes — 2-3 byte difference between markdown=True and markdown=False, consistent with asterisk-stripping but no font-style change). Canonical pattern for mixed-style text in custom-TTF documents: explicit segment-by-segment rendering with `set_font()` switches between regular -> italic -> regular and `cell()` per segment with pre-computed widths for alignment. The pattern is implemented at three sites in `book/_build_exports.py`: epigraph body paragraph 3 (the answer line), epigraph citation (Requiem italic in the middle), and the future-defaulted pattern for any new mixed-style content in custom-TTF builds. Precedent: when a PDF library's high-level convenience function produces silently-incorrect output on custom-TTF fonts, the fall-back is explicit-segment cell-based rendering. The library bug is logged here for future audit; if a future fpdf2 version fixes it, the cell-based pattern can be replaced with markdown=True for shorter code, but verification (rendering test PDFs and visual-checking) is required before that replacement. *(Cf. §9.45 production-tooling-fix.)* - **§9.48 visual-demand-pack as single-panel-multi-item operation** (Chat 48 fourth-cycle, Okan 5-item visual-demand pack). When the author surfaces multiple visual / typographic / formatting demands at once (justified text + dateline italics + epigraph wording + copyright wording + fiction disclaimer), the appropriate Code-side routing is one bundled implementation cycle that addresses all items together — not five parallel single-item cycles. Distinct from §9.37 external-review-pack as single-panel-multi-item: that protocol is for prose/architectural review with masters-panel adjudication; §9.48 is for visual/typographic demands with author-direct authorization (no panel needed). The bundled cycle produces: (a) one binding commit with per-item reasoning logged in the commit message; (b) full build-chain rerun (rebuild_unified -> build_manuscript -> build_exports -> complete_with_covers); (c) visual verification of representative pages at 150 DPI before commit. Canonical case: Chat 48 fourth-cycle Okan 5-item visual-demand pack landed in commit `0da8f5f` after full build-chain rerun and visual verification of pages 4 (copyright), 5 (epigraph), 6 (prologue), 13 (Bleecker body), 81 (Tomtom), 271 (Cherry Creek). Precedent: visual-demand packs are panel-units operationally but author-directed for ratification (the author's own visual taste IS the panel). *(Cf. §9.37 external-review-pack as single-panel-multi-item; §9.43 implement-or-leave-without-asking-back framing.)* - **§9.49 author-supplied-translation-record / author-supplied-design-record bypasses extended-masters-session** (Chat 48 second-cycle Akhmatova translation arrival; Chat 48 ABSOLUTE FINAL CLOSE back-cover record arrival). When the author has resolved an architectural-rupture-class question independently with chat-advisor (translation, attribution, citation, copyright wording, named-entity decision, back-cover text, similar design decisions) and supplies a record document (DOCX, MD, or other) capturing the decision and its rationale, Code applies the record directly without convening extended-masters-session for ratification. The record IS the ratification — the author has done the masters-equivalent work in the chat-advisor session, and the docx/md is the artifact equivalent of the reviews artifact. **First canonical case:** Chat 48 second-cycle Akhmatova fresh-translation arrived as `~/Downloads/Vera-Akhmatova-translation-record.docx` with three-draft history + author edits + quantitative comparison to Kunitz/Hayward + selection rationale; Code applied the record's "final translation" to `book/_build_exports.py` constants directly + bumped the front-matter to use the new translation; no panel session convened. **Second canonical case:** Chat 48 ABSOLUTE FINAL CLOSE back-cover text + finished PNG arrived as `~/Downloads/Vera-back-cover-record.docx` (8 iterative passes through the 8-master table free-form deliberation: Pass 1 three-angle drafts → Pass 2 famine-naming → Pass 3 first-light fix → Pass 4 In-May-2026 dating → Pass 5 page-1-fidelity repairs → Pass 6 destination-triplet polish → Pass 7 Yelena named/unnamed → Pass 8 final synthesis with Oxford-comma fix) + the finished `Vera_back_size_M_medium.png`; Code stored the record at `book-depository/design/covers/back-cover-record.docx` + replaced the back-cover PNG at `book-depository/design/covers/Vera_FINAL_back.png` + rebuilt the cover-wrapped PDF; no panel session convened. The 8-pass iterative author-master-table format documented in the back-cover record is itself a chat-advisor-side methodological pattern that Code does not need to replicate — Code's role at the hand-off is implementation + verification of the finished artifact. Precedent: author-supplied-record-class artifacts (translation, design, attribution, named-entity, etc.) bypass extended-masters-session ratification because the masters-equivalent work has already been done in chat-advisor; Code's role is storage + implementation + verification, not adjudication. *(Cf. §9.35 author-built-substantive-revision via chat-advisor workshop routed through extended-masters-session; the difference: §9.35 is for prose-revision arriving without author-decision-made, requiring panel ratification; §9.49 is for architectural-decision arriving with author-decision-made + supporting record, bypassing panel.)* - **§9.50 saturation-state-extraordinary-session-ratifies-implicit-architecture-as-explicit-canon** (Chat 48 fifth-cycle Phase 2B typographic-register deliberation). When the chat-advisor canon and Code-side canon are both silent on a category (per §9.23) AND the manuscript is at saturation (per §9.24), the appropriate routing is extraordinary-session 9-round 8 masters + 3 readers — but the SESSION OUTCOME at saturation is typically NEAR-UNANIMOUS RATIFY/PRESERVE/LEAVE rather than CHANGE. The session's value is in formalizing the implicit architecture as explicit canon (canon-document amendments) and surfacing any consistency-checks (e.g., the Joe-prose italic consistency-check from this session). Canonical case: Chat 48 fifth-cycle Phase 2B typographic-register deliberation produced 5 RATIFY/PRESERVE/LEAVE + 1 CONSISTENCY-CHECK from 6 candidates; the consistency-check (Joe-prose italic) verified clean — zero manuscript changes resulted; canon-document amendments landed at `auden-and-the-decade.md` v0.7 -> v0.8 + `architecture-seven-books.md` v0.54 -> v0.55 + `chat-advisor-vera/SKILL.md` v7 -> v8. Pattern: at saturation, an extraordinary-session typically produces small canon-documentation updates rather than manuscript changes. The session's near-unanimity is not a sign of panel-mis-calibration; it's the saturation-state signal that the architecture has been built well enough that no typographic intervention is needed. **Second canonical use of extraordinary-session-as-saturation-confirmation** (the first being Chat 47 fourth-cycle two-stage editorial pass per §9.44; the difference: §9.44 audits against a criterion-framework, §9.50 ratifies a new canonical category). *(Cf. §9.23 chat-advisor-canon-and-Code-side-canon-both-can-be-silent; §9.24 whole-book-pass-as-saturation-confirmation; §9.44 two-stage-whole-book-editorial-pass-as-saturation-confirmation; §9.18 architectural-defense-overrides-near-aphorism-flag.)* **§12 extended to Chat 48 with quantitative summary:** Chat 48 = 6 commits across 4 work-streams in one calendar day (the most active production-day in project history). Commit-by-commit: (i) `861b971` first-cycle: production-statistics report given a committed source at `reports/` + METHODOLOGY v0.3 -> v0.4 header drift fix; (ii) `21d2933` first-cycle continued: METHODOLOGY v0.4 -> v0.5 incremental bump (9 precedents §9.36-§9.44 formalized); (iii) `37e6b2b` second-cycle: ChatGPT 5-item external-review pack (Pinkas typo + word-break-fix-v1 + copyright page + new Akhmatova translation + PDF outline bookmarks; -1w manuscript 75,945 -> 75,944); (iv) `3276980` third-cycle: word-break-fix-v2 (cell-based) + citation italic preservation; (v) `0da8f5f` fourth-cycle: Okan 5-item visual-demand pack (justified text + Prologue dateline italic + epigraph wording revisions + copyright wording flexibility + fiction disclaimer; -1w manuscript 75,944 -> 75,943); (vi) [this commit] fifth/sixth-cycle: Phase 2B typographic-register deliberation (5 RATIFY + 1 CONSISTENCY-CHECK pass) + canon-document cascade (chat-advisor-vera v7 -> v8, auden-and-the-decade v0.7 -> v0.8, architecture-seven-books v0.54 -> v0.55, METHODOLOGY v0.5 -> v0.6) + auto-memory refresh + HANDOFF/README refresh + Chat 49 opening prompt + handoff zip generation. Manuscript: 75,944 -> 75,943 (-1w from `` token removal in Vera-prologue.md). Cumulative protocol uses at Chat 48 close: external-input-fix-protocol 26 (UNCHANGED), surgical-style-pass 9 (UNCHANGED), extended-masters-session 6 -> 7 (Chat 48 Phase 2B = 7th canonical), triplet-meter-pass 1 (UNCHANGED). Cumulative skill cascades at Chat 48 close: 4 skill bumps (chat-advisor-vera v7 -> v8 + auden-and-the-decade v0.7 -> v0.8 + architecture-seven-books v0.54 -> v0.55 + METHODOLOGY v0.5 -> v0.6) plus 1 reviews artifact (`chat48-phase-2b-typographic-register-deliberation.md`). **End-of-edits posture deepened further.** Systematic-edit phase remains CLOSED. Read-only / reading-and-evaluating posture preserved with three active sub-protocols (external-input-fix-protocol with drafting-panel + extended-masters-session as ratification-and-refinement + extended-masters-session as saturation-extraordinary-session per §9.50). The book is structurally complete + production-PDF-clean + canon-fully-ratified-typographically + Google-Books-distribution-ready (per Chat 48 fourth-cycle copyright-page wording flexibility for free distribution). Document grew from ~1010 lines (v0.5) to ~1180 lines (v0.6); v0.5 spirit preserved; new content additive in §9 sub-sections (§9.45-§9.50) + §12 extension. Author-ratified Chat 48 close. Mirror to `~/.claude/skills/` post-session-close (no skill file; this is pack-root document). - **v0.5** (Chat 48 first-cycle housekeeping, 2026-05-03) — Incremental bump formalizing the 9 methodological precedents that surfaced during Chat 47 second/third/fourth cycles but were deferred for v0.5 incorporation per the Chat 48 opening prompt's pending-precedents queue. **Major additions: cross-file V-line propagation discipline + external-review pack as single-panel-multi-item operation + author-discretionary voting threshold canonization + Code-opinion-as-panel-voice protocol canonization + surgical-style-pass specialization via derived skill (triplet-meter-pass v0.1 -> v0.2 canonized) + external-derived-skill installation discipline + 98/2-ratio validates author's self-policed discipline + implement-or-leave-without-asking-back framing + two-stage whole-book editorial pass as saturation-confirmation.** Bundled with Chat 48 first-cycle methodological-bookkeeping work: production-statistics report given a committed source at `book-depository/reports/` (prose source + matplotlib + fpdf2 build script + 8 chart PNGs) producing the rebuilt `Vera-production-statistics-report.pdf` at Chat 47 absolute close numbers; METHODOLOGY v0.3 -> v0.4 header drift fix (Chat 47 first-cycle commit had landed the v0.4 version-log entry but left the document-header line at v0.3; surgical fix at line 5). **§9 sub-section additions (precedents formalized):** - **§9.36 cross-file V-line propagation discipline** (Chat 47 second-cycle). When an external-review-ratified replacement of a load-bearing line at one canonical site lands in one file, the same line may be re-rendered as a cross-reference at one or more other files where the same architectural anchor surfaces; the replacement is not complete until all cross-file instances are propagated. Canonical case: Marginalia V-line replacement at the prologue's load-bearing line — the Chat 47 second-cycle's panel-ratified surgical/insertion fix-pass propagated the same revised V-line to the Oksana-prologue-1970 interstitial cross-reference in one binding commit + logged the propagation discipline in the reviews artifact `chat47-second-cycle-external-review-panel.md`. Precedent: a load-bearing canonical line is not a single-file artifact; cross-file references inherit replacement obligations and the propagation must be logged. *(Cf. typesetter-brief grep invariants for canonical line counts; cf. §9.32 architectural-coherence-over-cycle-iteration.)* - **§9.37 external-review pack debate as single-panel-multi-item operation** (Chat 47 second-cycle). When an external-review pack surfaces multiple distinct flags at once, the appropriate Code-side routing is a single masters panel deliberation that debates all items together as a multi-item operation — not N parallel single-item panels. The unified panel applies consistent voicing across items, sees cross-item architectural relationships (cross-file propagation, temporal-consistency between flagged sites, register-coherence), and produces one ratification record in a single reviews artifact. Canonical case: Chat 47 second-cycle's 5-item external-review panel (Marginalia V-line + Joe bread-nod + Vira-encounter cup-placed + Halloran chyron Vera-interior glimpse + Nina Pattern B extension) deliberated as one panel + produced 5 IMPLEMENT verdicts + 1 cross-file propagation note + 1 LEAVE on general trim-candidates, all in one bundled commit + one reviews artifact. Precedent: external-review packs are panel-units, not item-units, for ratification purposes; one panel + N items + N verdicts + 1 implementation cycle + 1 reviews record. *(Cf. §9.20 roundtable-as-diagnostic-not-edit-generative; §9.39 Code-opinion-as-panel-voice protocol.)* - **§9.38 author-discretionary voting threshold for IMPLEMENT-vs-LEAVE on author-directed external-review** (Chat 47 second-cycle). When the author has explicitly authorized an "implement if the panel sees fit" framing on an external-review pack (distinct from neutral LEAVE-vs-CHANGE adjudication), the panel's voting threshold for IMPLEMENT shifts: the panel operates with the author's pre-authorization at the threshold rather than requiring unanimous architectural-defense for LEAVE. Practical effect: the panel implements unless a strong architectural-defense case is voiced on a specific item. Canonical case: Chat 47 second-cycle's 5-item panel implemented all 5 items because the author had directed "implement if they see fit" pre-authorization; only the meta-item "general trim-candidates" was LEAVE'd because the panel's architectural-defense case (Master 1 V compression already saturated; further trim risks body-pulse signature) was strong enough to override the implement-by-default threshold. Precedent: author-authorization framing of an external-review pack changes the panel's default voting posture from architectural-defense-LEAVE-default to author-authorization-IMPLEMENT-default; the LEAVE outcome remains available but requires a stronger panel-level case. *(Cf. §9.43 implement-or-leave-without-asking-back framing.)* - **§9.39 Code-opinion-as-panel-voice protocol** (Chat 47 second-cycle). Code can express its own opinion within a masters panel deliberation as a distinct panel-voice — alongside the masters' literary voicings, not only as neutral orchestrator/scribe. The Code-voice is grounded in Code's full canon access (per-book reference files, grep invariants, antagonists.md, auden-and-the-decade.md, prior precedents in METHODOLOGY) and provides architectural-integrity verifications that the masters' voicings may not have surfaced. Canonical case: Chat 47 second-cycle's 5-item panel record includes Code's voice on each item alongside Master 1 V/Master 2 C/Master 3 S/Master 4 D/Master 5 A/Master 6 P voicings; Code's verifications (cross-file V-line propagation note per §9.36, Halloran-chyron temporal-consistency check across the 4-anchoring-site architecture) shaped 2 of the 5 implementation decisions. Precedent: Code is not just orchestrator at saturation; Code is a panel-voice with full-canon-access whose architectural verifications complement the masters' literary/aesthetic voicings; the Code-voice should be logged in reviews artifacts alongside the masters' voicings, not redacted as orchestration meta-text. *(Cf. §9.18 architectural-defense-overrides-near-aphorism-flag for Code's per-book reference-file architectural defense; cf. §9.32 architectural-coherence-over-cycle-iteration.)* - **§9.40 specialization-of-surgical-style-pass via derived skill** (Chat 47 third-cycle). The `surgical-style-pass` parent skill (project-agnostic, single-pattern audit methodology with six-stage staged diagnostic + classification rubric + report/Code-prompt artifact pair) can be specialized into a derived skill for a specific recurring construction that warrants its own discrimination criterion + replacement playbook + canonical examples. The derived skill inherits the parent's six-stage methodology + classification rubric + report/Code-prompt artifacts unchanged; it specializes the diagnostic strategy, the discrimination criterion, the replacement playbook, and the body-pulse-keystone audit specifically for its target pattern. Canonical case: `triplet-meter-pass` v0.1 (specialization of `surgical-style-pass` for the three-beat declarative cluster *X is X. X is not X. X is.* and its near-relatives); installed at Chat 47 third-cycle; first canonical invocation produced 5 IMPLEMENT + 2 LEAVE + 1 DEFER from the chat-advisor session 35 Stage 2 report's 7-item slate. Precedent: when a single recurring construction (causal connectors, *the kind of* hedges, body-language clichés, the X-is-X triplet, em-dash explanations, *which is* gloss, etc.) warrants its own focused audit with pattern-specific discrimination criterion + replacement playbook, a derived-skill specialization is the right unit, not a one-off parent-skill invocation. *(Cf. §9.41 external-derived-skill installation discipline; §9.42 98/2-ratio validates author's self-policed discipline.)* - **§9.41 external-derived-skill installation discipline** (Chat 47 third-cycle). When a specialization-of-surgical-style-pass derived skill arrives from an external session (chat-advisor or other) the install discipline mirrors the bundle-install-with-verification-first discipline of §9.22: verify the skill's pattern-recognition strategy + discrimination criterion + replacement playbook against the manuscript before applying any prose changes; install the skill file at `book-depository/skills//SKILL.md` + mirror to `~/.claude/skills/`; run the first invocation as a Code-side adjudication of the report's findings, applying §9.39 Code-opinion-as-panel-voice + §9.43 implement-or-leave-without-asking-back. The post-invocation skill-version update (e.g., operating-target revision per actual Vera precedent) is a distinct event from the installation — the same skill can be bumped after its first canonical invocation reveals where the predicted defaults diverged from disciplined-prose actuals. Canonical case: `triplet-meter-pass` v0.1 install at Chat 47 third-cycle (verify chat-advisor session 35 Stage 2 report's 7-item slate against per-book reference files + Master 6 P-protected zones + body-pulse signature keystones; install skill file + mirror; first invocation 5 IMPLEMENT + 2 LEAVE + 1 DEFER) followed by v0.1 -> v0.2 update at Chat 47 fourth-cycle from external author-uploaded version with operating-target revision 70/30 -> 95/5 per Vera precedent. Precedent: external-derived-skills install via verify-then-canonize-then-invoke; the install is itself a sub-protocol distinct from the skill's first invocation; post-invocation skill-version-update is its own discrete event. *(Cf. §9.22 bundle-install-with-verification-first-discipline; §9.40 specialization-of-surgical-style-pass via derived skill.)* - **§9.42 98/2-ratio validates author's self-policed discipline** (Chat 47 third-cycle). When a single-pattern audit predicts a generic operating-target ratio (e.g., 70/30 accept/cut for `triplet-meter-pass` v0.1's predicted distribution against generic prose) but the manuscript scores far higher on the keep-side of that ratio (e.g., 98/2 actual at Vera), the divergence validates the author's self-policed discipline — it is not a mis-counting + does not warrant tightening the audit. Practical implication: the operating-target default for the skill should be revised toward the actual-observed disciplined-prose ratio so that future invocations against similarly-disciplined manuscripts don't generate false-positive flags. Canonical case: `triplet-meter-pass` v0.1 predicted 70/30 (accept/cut) based on hypothetical generic-prose distribution; Vera precedent at Chat 47 third-cycle measured 98/2 actual (3 cuts + 2 ambiguous + 2 architectural-decisions + 1 deferred from 203 candidates across the manuscript); v0.2 update at Chat 47 fourth-cycle revised the operating-target default to 95/5, matching the disciplined-prose norm + acknowledging that the v0.1 70/30 prediction was an unsupported hypothesis. Precedent: a manuscript that scores higher on a discipline-defending audit than the predicted default has validated the author's discipline; the skill's default should be updated to reflect the discipline-norm rather than retain the discipline-failure-prediction; the generic-prose default and the disciplined-prose actual are different operating regimes and the skill's default should declare which regime it serves. *(Cf. §9.40 specialization-of-surgical-style-pass via derived skill; cf. §9.15 some-apparent-improvements-ARE-the-architecture-not-the-flaw.)* - **§9.43 implement-or-leave-without-asking-back framing** (Chat 47 third-cycle). When a panel adjudicates an external-review or audit-class report against the manuscript and produces per-item verdicts (IMPLEMENT / LEAVE / DEFER), the panel's verdicts can be applied directly without round-robin author-confirmation on each item — provided: (a) the author has pre-authorized the panel's adjudication framing per §9.38 author-discretionary voting threshold; (b) the panel has full architectural-defense canon access per §9.18 + §9.39; (c) per-item reasoning is logged in the reviews artifact for author retroactive review; (d) the operation is bundled into a single binding commit with diff summary. The implement-or-leave-without-asking-back framing is appropriate at saturation when the panel-author trust has been established across many cycles; it accelerates throughput at end-of-edits without sacrificing architectural integrity. Canonical case: Chat 47 third-cycle's `triplet-meter-pass` adjudication produced 5 IMPLEMENT + 2 LEAVE + 1 DEFER from the 7-item slate; author had pre-authorized "implement what the panel sees fit"; per-item reasoning logged in `reviews/chat47-third-cycle-triplet-meter-pass-implementation.md`; author retroactively confirmed all verdicts post-implementation. Precedent: at saturation with full panel canon access + author pre-authorization, the panel's verdicts can apply directly; round-robin author-confirmation per item is not required; the reviews artifact is the audit-trail substitute. *(Cf. §9.38 author-discretionary voting threshold; cf. §9.20 roundtable-as-diagnostic-not-edit-generative.)* - **§9.44 two-stage whole-book editorial pass as saturation-confirmation at end-of-edits** (Chat 47 fourth-cycle). The two-stage whole-book editorial pass — Stage 1 (a 9-criterion targeted audit by 3 named master-pairs operating in parallel) followed by Stage 2 (the full v0.31 framework adjudicated by Master 3 S-lead with all 8 masters) — is the canonical saturation-confirmation protocol at end-of-edits. The two stages distinguish: Stage 1 surfaces any HIGH-conviction findings against a focused 9-criterion subset chosen for the saturation-state question (testimony-multiplicity coherence, city-register coherence, women's-sovereignty floor, etc.); Stage 2 adjudicates against the full 42+-criterion (or current-version) framework. A two-stage outcome of zero implementations confirms the manuscript has held against both targeted + comprehensive audit; it is the second canonical use of eight-master per-book read at end-of-edits (the first being Chat 43 close per §9.24 whole-book-pass-as-saturation-confirmation). Canonical case: Chat 47 fourth-cycle's two-stage pass — Stage 1 (Master 7 G + Master 8 P + Master 5 A; 9-criterion targeted audit) returned 0 HIGH-conviction findings + ~12 LEAVE-confirmations; Stage 2 (Master 3 S-lead, full v0.31 framework, all 8 masters) returned 0 implementations + ~30 LEAVE-confirmations across the full criterion set; combined two-stage result confirmed saturation; logged in `reviews/chat47-fourth-cycle-two-stage-editorial-pass.md`. Precedent: two-stage whole-book editorial pass with zero-implement outcome is the canonical end-of-edits saturation signal; the two-stage structure (targeted then comprehensive) discriminates between zone-specific saturation and global saturation; runs as the second canonical use of eight-master per-book read at end-of-edits. *(Cf. §9.24 whole-book-pass-as-saturation-confirmation for the first canonical instance at Chat 46 fourth-cycle; cf. §9.14 pure-LEAVE deliberation outcome class.)* **§12 extended to Chat 48 first-cycle with quantitative summary:** Chat 48 first-cycle = methodological-bookkeeping commit bundling (a) production-statistics report given a committed source at `book-depository/reports/` (production-statistics.md prose source + build_production_statistics.py matplotlib + fpdf2 generator + 8 chart PNGs at `charts/01-08-*.png`) producing the rebuilt `Vera-production-statistics-report.pdf` at `~/Downloads/` updated to Chat 47 absolute close numbers; (b) METHODOLOGY v0.3 -> v0.4 header drift fix (Chat 47 first-cycle commit had landed the v0.4 version-log entry but the document-header line remained at v0.3 — surgical fix at line 5); (c) METHODOLOGY v0.4 -> v0.5 incremental bump formalizing the 9 deferred precedents from Chat 47 second/third/fourth cycles in §9.36-§9.44 (this entry). Manuscript UNCHANGED at 75,945 words; 87 catches + 0 candidates UNCHANGED. Stream 4 fresh `wc -w` measurement at v0.5 bump time: 1,107,170 -> 1,182,827 (+75,657 across Chats 46-47; reviews-dominant growth: reviews/ 804,316 + skills/ 290,059 + top-level docs 51,874 + design/ 30,463 + agent-pitch/ 6,115). Updated four-ratio table at Chat 47 absolute close: prompt-to-final 187,237 / 75,945 = ~2.47x (was 2.46x); generated-to-final 30,277,448 / 75,945 = ~398.7x (was 398.9x); artifact-to-final 1,182,827 / 75,945 = ~15.57x (was 14.60x; reviews-artifact growth dominates); survival 75,945 / 30,277,448 = ~0.2508% (was 0.2506%; asymptote holds at ~0.25%). Process-residue per word of manuscript change at Chats 46-47 = ~688x (vs ~468x at Chats 44-45 vs ~5x at Chats 38-43); saturation phase change deepens further. Cumulative protocol uses at Chat 48 first-cycle: external-input-fix-protocol 26 (UNCHANGED), surgical-style-pass 9 (UNCHANGED), extended-masters-session 6 (UNCHANGED), triplet-meter-pass 1 (UNCHANGED; v0.1 first canonical invocation at Chat 47 third-cycle + v0.2 install update at Chat 47 fourth-cycle per §9.41 external-derived-skill installation discipline). **Saturation-state methodological-housekeeping pattern.** v0.5 is the first incremental-bump-after-major-bump in the METHODOLOGY series — v0.4 was a major bring-current covering ~860 lines + 17 precedents at Chat 47 close; v0.5 adds 9 precedents that were in operational practice across Chat 47 cycles but not yet documented. The pattern: at saturation, the methodology document accumulates precedents at a similar rate to the manuscript's stability — small, additive, verifiable; new precedents arrive in operational practice and are logged at the next housekeeping cycle. The v0.4 -> v0.5 increment matches: 9 precedents in operational practice -> ~150 lines added -> no body restructure -> header paragraph minimally updated. This is the methodological-saturation-housekeeping rhythm that the publication-prep posture sustains. Document grew from ~860 lines (v0.4) to ~1010 lines (v0.5); v0.4 spirit preserved; new content additive in §9 sub-sections (§9.36-§9.44) + §12 extension. Author-ratified Chat 48 first-cycle. Mirror to `~/.claude/skills/` post-session-close (no skill file; this is pack-root document). - **v0.4** (Chat 47 housekeeping carry-forward, 2026-05-03) — Extension covering Chat 45 sixth-cycle through Chat 47 W2 final halo Ch XI rewrite. **Major additions: 17 methodological precedents formalized + read-only posture transition documented + drafting-panel sub-protocol canonized + author-directed extended-masters-session routing canonized.** **§9 sub-section additions (precedents formalized):** - **§9.18 architectural-defense-overrides-near-aphorism-flag** (Chat 45 sixth-cycle, 8th canonical precedent in METHODOLOGY series). When a chat-advisor surgical-style-pass or editorial-pass produces candidates that read as register-different from the prevailing voice and the candidate is high-conviction at the chat-advisor's framing, Code's pre-flight architectural review against per-book reference files (especially `auden-and-the-decade.md` for Auden anchors) takes precedence. Canonical case: B7 Ch XI close *"the world stays defenseless under its night"* CAND-2; flagged as the only true near-aphorism by chat-advisor session 32; Code identified as canonical Book VII Auden cadence-echo; eight-master panel ratified 8/0 LEAVE on architectural-defense grounds. Precedent: chat-advisor's silence on a register-classification is not evidence of non-protection. *(Cf. chat-advisor-vera v6 Tier 1 "Chat-advisor canon is full but not total".)* - **§9.19 continuity-correction-class-vs-signature-class discrimination** (Chat 46 first-cycle). External-critic flag B6-04 photograph-line "gave→left for / ten→nine" ratified 8/0 as continuity-correction class (small-anchor-text correction matching architectural fact: Karen-as-mother + nine-not-ten time-elapsed) — distinct from signature-class flag (which would propose a register-level rewrite). The discrimination determines whether the External-Input-Fix-Protocol's surgical 3-round panel applies (continuity-class) or the Extended-Masters-Session 9-round protocol applies (signature/architectural-class). - **§9.20 roundtable-as-diagnostic-not-edit-generative** (Chat 46 second-cycle). The Joe-position 2-round + 8-master roundtable 5-round bundled deliberation produced **5 LEAVE-confirmations + 0 edits**. The roundtable's value at saturation is diagnostic (architecture verified intact, masters' concerns surfaced and self-yielded under panel pressure) — not edit-generative. Pattern: a saturation-state roundtable that yields zero edits is the protocol working as designed, not failing. - **§9.21 self-yielded-master-concerns-LEAVE-under-panel-pressure** (Chat 46 second-cycle). Three masters surfaced concerns at Round 1; on Round 2-3 deliberation, each self-yielded their concern under panel-collective scrutiny. Precedent: in 8-master panel deliberation at saturation, individual master-concerns that fail to gather panel-support are correctly classified as LEAVE-class — the master themselves recognizes the architecture's integrity once the full canon is reviewed. The 8-master panel's collective intelligence exceeds individual master's first-impression at saturation. - **§9.22 bundle-install-with-verification-first-discipline** (Chat 46 third-cycle). Chat-advisor-bundle v3-3 → v3-4 install (Phase 2A typographic-register reconciliation): the install verified the bundle's typographic findings against the manuscript BEFORE applying any prose changes. Verification surfaced six register-distinctions pending extended-masters-session adjudication; bundle install applied the canon-only updates (chat-advisor-vera v5 → v6 with Tier 1 extraction-fidelity discipline) without touching manuscript prose. Precedent: bundle installs at saturation must verify-then-apply; canon-only updates are routine; manuscript-touching candidates are deferred to appropriate-class protocol. - **§9.23 chat-advisor-canon-and-Code-side-canon-both-can-be-silent** (Chat 46 third-cycle). Six typographic-register findings surfaced by chat-advisor session 33 (two-register I-voice, "Keep your face still" italic refrain, Auden cadence-echo vs verse-quotation, Joe's notes as italic interior-speech, embedded-historical-voice, matrilineal Cyrillic chiasmus typographic register) are pending extended-masters-session adjudication. Both the chat-advisor canon and the Code-side canon are SILENT on Vera-specific typographic-register canon — a category neither has yet ratified. The silence is not absence-of-architecture; it's pre-ratification. Pattern: when both canons are silent, the appropriate routing is extended-masters-session for new-canon-category establishment, not unilateral canon-add by either side. - **§9.24 whole-book-pass-as-saturation-confirmation** (Chat 46 fourth-cycle). The first canonical whole-book multi-agent editorial pass (3 parallel agents distributing ~76K words; book-editorial v0.31 retroactively documents the methodology). Aggregate findings: 7 HIGH-conviction catches + 6 MEDIUM-conviction candidates across the three agents + ~30 LEAVE-confirmations. Code-side architectural verification ratified 7 of 7 catches and downgraded all 6 candidates to LEAVE on architectural-defense grounds. Saturation signature: implementations:LEAVE-after-Code-classification ratio ≈ 7:9 = 0.78, confirming saturation-state expectation that most findings classify as LEAVE under architectural defense. The whole-book pass is the canonical pre-condition for the systematic-edit-phase CLOSE event + read-only-posture transition. - **§9.25 audiobook-cold-read-surfaces-singular-pattern-breaks** (Chat 46 fifth-cycle). Audiobook listening-mode catches what silent-eye proofreading misses; the audiobook ear holds rhythm, the rhythm holds the refrain, and the refrain registers the break. Five author-flagged sites at Chat 46 fifth-cycle were all SINGLE INSTANCES in 75,825 words — isolated tics against established refrains, not manuscript-wide saturation problems. Audiobook-cold-read is now a canonical sub-session trigger under the read-only posture; the trigger is established by the Chat 46 fifth + sixth cycles + the Chat 47 W2 final halo Ch XI rewrite (audiobook-cold-read of Wednesday-night chapter triggered the workshop that produced the W2 final). - **§9.26 Master 1 V-supervised-parallel-agent-sweep-with-zero-new-catches-outcome** (Chat 46 fifth-cycle). Two parallel general-purpose sweep agents under Master 1 V-supervision returned ZERO additional catches across all 5 fifth-cycle pattern-classes. Confirms systematic-edit phase's READ-ONLY closure; sweep distinguished body-without-conscious-will refrain (8 canonical at Chat 47 close) from off-register mental-instance tic (1 site fixed). Discrimination work IS the sweep's value at saturation. - **§9.27 pattern-class-as-family-architecture** (Chat 46 fifth-cycle). 5 cold-read flags representing 5 distinct pattern-classes; on sweep prove to each be represented by single instance at saturation. Pattern-class identification is reusable for future cold-read flag categorization but manuscript at saturation does not produce additional catches per class. The architecture itself is the saturation; the singular pattern-break is what the body's-pace audiobook ear catches. - **§9.28 architectural-canon-finding-logged-for-future-Tier-2** (Chat 46 fifth-cycle). The body-without-conscious-will refrain at 7 sites (8 at Chat 47 close) was identified at Chat 46 fifth-cycle and deferred to Tier 2 incorporation in chat-advisor-vera at the next housekeeping cycle. The deferral is the protocol working: canon-discoveries surface at fix-pass time + are logged for documentation cascade at housekeeping-cycle time. Chat 47 housekeeping completes the deferred Tier 2 incorporation in chat-advisor-vera v7. - **§9.29 first-canonical-instance-of-read-only-with-occasional-cold-read-fixes-pattern** (Chat 46 fifth-cycle). The systematic-edit-phase-CLOSED posture from Chat 46 fourth-cycle is now OPERATIONAL through Chat 46 fifth-cycle; this is the canonical first-instance of the read-only-with-author-directed-cold-read-fixes pattern documented in book-editorial v0.31 Whole-Book Pass Mode + posture lines. - **§9.30 multi-author-message-progressive-scope-expansion** (Chat 46 sixth-cycle). 5 cascading author-directives within single cycle absorbed iteratively (jazz-add to music; 3rd Halloran reference; ex-musician reference; subtle conspiracy-connection hint). The external-input-fix-protocol's design accommodates within-cycle scope-expansion; the protocol is not single-message-bound. Pattern: when author surfaces additional flags during a cycle, the cycle absorbs them rather than closes-and-reopens. - **§9.31 drafting-panel-within-external-input-fix-protocol** (Chat 46 sixth-cycle). For substantive prose insertions (vs surgical edits), named master-pair drafts inline; full panel ratifies at compressed Round 4. Distinct from extended-masters-session 9-round protocol; appropriate for cold-read-derived insertion-class work. Canonical first instances: B1-05 music-background drafting panel (Master 1 V + Master 6 P); B1-06 + B4-08 Halloran chyron drafting panel (Master 2 C + Master 3 S). Both drafted inline + 8/0 ratified at compressed Round 4. - **§9.32 architectural-coherence-over-cycle-iteration** (Chat 46 sixth-cycle). The Halloran-3-mention timing-scan caught a panel-first-draft inconsistency (initial B1-06 draft "ANALYST'S TESTIMONY EXPECTED THIS WEEK" inconsistent with B7's "second day testimony"); re-deliberated and corrected within-cycle. Pattern: an external-input-fix-protocol cycle includes architectural-coherence verification BETWEEN draft-ratification and implementation; coherence-failures trigger within-cycle re-deliberation. - **§9.33 special-editorial-pass-as-canon-discovery-vs-catch-finding** (Chat 46 sixth-cycle). At saturation, the Master 5 A/Master 4 D/Master 2 C special editorial pass (8 invented factors beyond standard 42) surfaced primarily skill-augmentation-candidates (4 augmentations + 1 catch) rather than prose-catches. The pass's value at saturation is canon-discovery for future skill-file bumps (now formalized in book-editorial v0.31). - **§9.34 subtle-hint-architecture-for-arch-antagonist-conspiracy-connection** (Chat 46 sixth-cycle). Single-chyron substitution at B1-06 ("PRIVATE-INTELLIGENCE CONTRACTORS NAMED") provides connection-language without hard-push. Reader catches on close-read; Vera registers nothing. The subtle-hint pattern preserves antagonist-canon ("Halloran never knows Vera exists" + "Vera never knows Halloran's name") while connecting institutional-spheres at the chyron-grammar level. Now canonized in `antagonists.md` v0.9 Halloran 3-news-TV-chyron architecture. - **§9.35 author-built-substantive-revision-via-chat-advisor-workshop-routed-through-extended-masters-session-as-author-directed-ratification-and-refinement** (Chat 47 W2 final halo Ch XI rewrite). When the author builds substantive prose-revision work via the chat-advisor's iterative workshop process and arrives at Code with the change-decision already made, the appropriate Code-side routing is extended-masters-session (8 masters + 3 reader-personas) operated as ratification-and-refinement (not LEAVE-vs-CHANGE adjudication). The masters cannot block the author's change-decision; their role is: (a) verify the author-built prose preserves architectural integrity against the full Code-side canon (which the chat-advisor canon does not have full access to per CAND-2 + §9.18 precedents); (b) surface line-level refinement-candidates the author may accept or decline; (c) flag concerns the author should know about. First canonical case: Chat 47 W2 final halo Ch XI rewrite — extended-masters 8/0 RATIFY with 5 refinement-candidates surfaced; author accepted A-1 (back-yard verification fragment), declined D-1/V-1/C-1; P-1 (Master 8 P hedge-verification) resolved by grep. Manuscript delta: B7-10 published 925w → W2 final 822w (-103w), Vera-complete.md 75,960 → 75,857 (-103w). Page count UNCHANGED at 289 interior / 291 with-covers. Distinct from §9.31 drafting-panel-within-external-input-fix-protocol: that protocol is for substantive insertions into existing chapters at cold-read scale; §9.35 is for full-chapter-replacement at architectural-rupture scale where the author has done the building. **§12 extended to Chats 46-47 with quantitative summary:** Chat 46 = 6 sub-cycles + 35 surgical edits + 2 substantive prose insertions + new canonical architectures (Halloran 3-news-TV-chyron + Vera music-background + body-without-conscious-will refrain at 7→8 sites) + 5 cumulative skill bumps + 1 bundle install + 1 systematic-edit-phase-CLOSE event + complete documentation cascade; manuscript 75,835 → 75,960 (+125w net across all 6 sub-cycles). Chat 47 = 1 W2 final halo Ch XI rewrite + 5 housekeeping bumps (METHODOLOGY v0.3 → v0.4 + book-editorial v0.30 → v0.31 + chat-advisor-vera v6 → v7 + antagonists v0.8 → v0.9 + typesetter-brief v3.7 → v3.8) + agent-pitch + production-stream-metrics + covers verification + Phase 2B chat-advisor session 33 candidates artifact; manuscript 75,960 → 75,857 (-103w). Cumulative external-input-fix-protocol uses: 26 → 26 (unchanged at Chat 47; W2 routed via extended-masters-session not external-input-fix-protocol per §9.35); cumulative extended-masters-session uses: 3 → 4 (Chat 47 W2 = 4th canonical). **Skill cascade documented at v0.4:** book-editorial v0.29 → v0.31 (with retroactive ratification of unlogged Chat 46 fourth-cycle work + Chat 47 augmentations); chat-advisor-vera v6 → v7 (4th positive canon-rule + Tier 2 manuscript-state header refresh); antagonists v0.8 → v0.9 (Halloran 3-chyron architecture canonized); typesetter-brief v3.7 → v3.8 (W2 metrics); production-stream-metrics v0.4 unchanged at Chat 47 close; surgical-style-pass v1 unchanged. **End-of-edits posture deepened.** Systematic-edit phase remains CLOSED; read-only / reading-and-evaluating posture preserved; author-directed cold-read fixes accommodated via external-input-fix-protocol (drafting-panel sub-protocol for substantive insertions per §9.31) + extended-masters-session (per §9.35 for chapter-rewrite-class work). Two canonical sub-protocols now serve the read-only posture, distinguished by scale and class. Document grew from ~660 lines (v0.3) to ~860 lines (v0.4); v0.3 spirit preserved; new content additive in §9 sub-sections + §12 extension. Author-ratified Chat 47 close. Mirror to `~/.claude/skills/` post-session-close (no skill file; this is pack-root document). - **v0.3** (Chat 45 fifth-cycle close, 2026-05-01) — Extension covering Chat 44 (4 sub-cycles) + Chat 45 (5 sub-cycles) methodology. **Major additions:** §9.11 architectural triangulation (independent audits same-conclusion, canonical Chat 45 third + fourth cycles); §9.12 logged-dissent-not-blocking (Master 2 C Charlotte-bedroom millimeter-dissent + Master 7 G testimony-multiplicity dissent at Chat 45 first + fifth); §9.13 room-vs-need distinction (Chat 45 second-cycle four-additions deliberation, including the father-touch panel-cannot-converge → writer-decides outcome at LEAVE per "he has to be quiet"); §9.14 pure-LEAVE deliberation outcome class (zero-implementation outcomes are protocol working as designed at saturation; three canonical instances); §9.15 "Some apparent improvements ARE the architecture, not the flaw" (Chat 44 Item 8 older-women + Chat 45 first-cycle Edit 2 chant + Chat 45 fourth-cycle Pattern A+B; pattern of recognition-not-correction across external-reader cycles producing positive canon-rules); §9.16 bundle install with intelligent reconciliation (Chat 45 fifth-cycle bundle-v3-2 with stale "Pending micro-trim candidates" section preserved + STATUS NOTE annotation); §9.17 three-verdict-mix outcome from taste-class concerns (Chat 45 fifth-cycle three-concerns deliberation producing LEAVE / LEAVE / ADOPT with logged dissent). **§12 extended to Chats 27-45** with full Chat 44 + 45 quantitative summary: 9 sub-cycles + 20 surgical edits + 22 LEAVE-confirmations + 3 canon bumps + 1 bundle install + manuscript -165w net (76,000 → 75,835); confirmations-to-implementations ratio 5.5:1 in Chat 45 (signature of saturation); three positive canon-rules in chat-advisor-vera SKILL.md Tier 2 (older-women + Pattern A+B + back-matter sequencing); end-of-edits posture deepened. **Skill cascade documented at v0.3:** typesetter-brief v2.5 → v3.2 (across Chats 44-45 cycle entries); external-input-fix-protocol v0.1 (Chat 43 close: 9 uses → Chat 45 fifth-cycle close: 19 uses, +10 across both chats); surgical-style-pass v1 (7 → 8 uses); extended-masters-session v0.1 (3 → 3 uses unchanged); chat-advisor-vera v1 → v3 → v4 → v5 (bundle-v3 install + Code's Pattern A+B addition + bundle-v3-2 install). **Architecture-of-architecture observation logged.** Document grew from ~570 to ~660 lines; v0.2 spirit preserved; new content additive in §9 sub-sections + §12 extension. Author-ratified Chat 45 fifth-cycle close. Mirror to `~/.claude/skills/` post-session-close (no skill file; this is pack-root document). - **v0.2** (Chat 43 close, 2026-04-29) — Comprehensive bring-current covering ~28 chats of methodological development through the manuscript's completion and publication-prep arc. Major additions: §0 published colophon (seven-paragraph piece) + two dedications + working-signature triple preserved as project's internal signature; §1 origin extended with eight-master expansion (Master 7 G + Master 8 P) + chat-first-Code-migrated split's evolution into two-agent workflow; §2.1 design-stage ratifications across Chats 25-41 + architectural-rupture decisions formalized via `extended-masters-session`; §2.2 Code role evolved into orchestration-than-execution; §2.3 chat-advisor upgraded to two-tier (`chat-advisor-vera-SKILL.md` v1); §2.4 six masters extended to eight (Master 7 G + Master 8 P integration formalized); §2.5 external-critic rounds + deflation-tier framework; §2.6 Opus 4.7 specified + Agent-tool delegation patterns; §3 full skill-file inventory at Chat 43 close (versions: architecture v0.54, character v0.11, locations-timelines v0.7, voice-and-body v0.26, book-editorial v0.28, typesetter-brief v2.5, external-input-fix-protocol v0.1, surgical-style-pass v1, extended-masters-session v0.1, chat-advisor-vera v1); §4 Mode E canonical sites + Matrilineal Cyrillic Chiasmus + Cyrillic-exception list + matrilineal-axis chronological spine + front-matter 8-element architecture + three architectural spines verified at end-of-edits; §5 new protocols (5.9 two-agent workflow + external-input-fix-protocol; 5.10 three-architectural-spine triad audit via Agent-tool delegation; 5.11 cross-skill-file canon-hygiene); §6 Three Moves held; §7 design decisions + architectural-rupture decisions + end-of-edits declaration as additional human-direction-core moments; §8 prose-level discipline confirms-before-risky-actions + commit-each-milestone-and-push standing orders; §9 new sub-sections (9.5 confirmed sixth category; 9.7 matrilineal-axis retcon and skill-file drift; 9.8 LEAVE-confirmation as legitimate audit outcome; 9.9 production-stream-metrics survival ratio; 9.10 architectural-rupture-class decision discipline); §10 some open questions answered, two new ones surfaced (skill-file canon-hygiene routine vs incident-driven; eight extended masters' relationship to six); §11 unchanged; **NEW §12** — the publication-prep arc Chats 27-43 (posture, triggers, work-streams, quantitative summary, end-of-edits declaration). Document grew from 368 lines to ~570 lines; the v0.1 spirit and structure preserved; new content additive rather than restructuring. Author-ratified Chat 43 close. Mirror to `~/.claude/skills/` post-session-close. - **v0.1** (Chat 15 open, 2026-04-20) — Initial draft. Pre-Book-IV-drafting task. Captures the method as it stands at Chat 14 close: the apparatus (Okan / Code / Chat / Six Masters / Third-party LLMs / Anthropic infrastructure / three-platform architecture), the skill system as evolving artifact, the architectural system's load-bearing rules, the protocols of organized dissent (masters rounds, pre-consults, pin rounds, writing-room rotation, Master 1 V Compression trio, order-of-operations, External-Critic Input Protocol, catch logging), the Three Moves pre-drafting stance, the human-direction core, the prose-level discipline, observed patterns and findings (composite-score compression, "the body" drift, ChatGPT 5.4 canon-divergence pattern, masters false-convergence, Sub 3 tentative sixth category, Vera chat register), open questions of the method, and the colophon's ambient Gramsci stance. Document lives at pack root. Will update as the method evolves through Books IV–VII drafting.