@demig0d2/skills 1.0.1 → 1.1.0

package/skills/book-creator/SKILL.md

@@ -0,0 +1,848 @@
+---
+name: book-creator
+version: 1.0.0
+description: |
+  Complete end-to-end book creation system for Vivid (Dheelep N). One skill that does
+  everything: concept expansion, research aggregation, structure options, full intake
+  questionnaire, table of contents, chapter writing (guided or autonomous), 7-dimension
+  chapter audit, DNA-aware humanizer pass, continuity tracking, and KDP-ready DOCX output.
+  Triggers when user says "write a book," "start a book," "I have a book idea," "create a
+  manuscript," "write chapter," "humanize this," "audit this chapter," or uploads a concept
+  document. All scripts and references are bundled — install this one skill and the entire
+  pipeline works without installing anything else.
+allowed-tools:
+  - Read
+  - Write
+  - Edit
+  - Bash
+  - Glob
+  - WebSearch
+  - AskUserQuestion
+---
+
+# Book Creator — Complete Book Production System for Vivid
+
+One skill. Every tool needed to go from a seed idea to a KDP-ready published manuscript.
+No other skills needed. Everything is bundled.
+
+---
+
+## SCRIPTS (all in `scripts/`)
+
+### Pre-writing
+```bash
+python scripts/validate_concept.py <concept.md>          # validate concept doc completeness
+python scripts/validate_concept.py <concept.md> --strict # stricter word count minimums
+```
+
+### During writing
+```bash
+python scripts/word_count.py <manuscript.md>                        # per-chapter word counts
+python scripts/word_count.py <manuscript.md> --target medium        # vs KDP targets
+python scripts/toc_extract.py <manuscript.md>                       # verify structure
+python scripts/toc_extract.py <manuscript.md> --markdown            # markdown TOC output
+
+python scripts/log_manager.py init "Book Title"                     # init continuity log
+python scripts/log_manager.py add-chapter 1 "Chapter Title"         # log chapter
+python scripts/log_manager.py add-fact "Established fact"           # log narrative fact
+python scripts/log_manager.py add-insight 1 "Core insight"          # log delivered insight
+python scripts/log_manager.py add-metaphor 1 "metaphor used"        # retire metaphor
+python scripts/log_manager.py add-thread "Open thread description"  # track open thread
+python scripts/log_manager.py close-thread 0                        # resolve thread
+python scripts/log_manager.py show                                   # full log view
+python scripts/log_manager.py summary                               # quick status
+python scripts/conflict_check.py <chapter.md>                       # check for contradictions
+```
+
+### Per-chapter quality
+```bash
+python scripts/dna_scan.py <chapter.md>                  # identify protected DNA constructions
+python scripts/scan_ai_patterns.py <chapter.md>          # flag all 25 AI patterns
+python scripts/scan_ai_patterns.py <chapter.md> --summary # quick count view
+python scripts/score_report.py --chapter "Title"          # interactive audit scoring
+python scripts/score_report.py --chapter "Title" --output audit.md  # save report
+```
+
+### Research
+```bash
+python scripts/bank_formatter.py research_bank.json          # view research bank
+python scripts/bank_formatter.py research_bank.json --validate # check quality
+```
+
+### Final output
+```bash
+python scripts/kdp_check.py <manuscript.docx>              # KDP pre-flight check
+python scripts/kdp_check.py <manuscript.docx> --trim 6x9   # with trim size
+```
+
+---
+
+## REFERENCES (all in `references/`)
+
+- `references/kdp_specs.md` — KDP trim sizes, margins, typography, page count estimates
+- `references/patterns_quick_ref.md` — All 25 AI pattern trigger words and before/after fixes
+- `references/thinkers_reference.md` — Curated philosophers and quotes for Vivid's genres
+- `references/log_schema.md` — Continuity log JSON schema and full CLI reference
+
+Read the relevant reference at the start of each phase:
+- Concept phase → thinkers_reference.md
+- Writing phase → kdp_specs.md + log_schema.md
+- Humanizer phase → patterns_quick_ref.md
+- Output phase → kdp_specs.md
+
+---
+
+## PIPELINE MAP
+
+```
+INPUT: SEED IDEA / ROUGH CONCEPT / FULL CONCEPT DOCUMENT
+              ↓
+   ┌─────────────────────────────────┐
+   │  PHASE 1: CONCEPT               │ ← skip if full concept doc already provided
+   │  Expand seed → concept doc      │
+   │  Validate: validate_concept.py  │
+   └─────────────────────────────────┘
+              ↓
+   ┌──────────────────────────────────────┐
+   │  PHASE 2: STRUCTURE OPTIONS          │
+   │  3 proposals → user picks one        │
+   └──────────────────────────────────────┘
+              ↓
+   ┌───────────────────────────────────────────┐
+   │  PHASE 3: INTAKE QUESTIONNAIRE            │
+   │  26 questions → all decisions locked      │
+   └───────────────────────────────────────────┘
+              ↓
+   ┌───────────────────────────────────────────────────┐
+   │  PHASE 4: RESEARCH BANK                           │
+   │  Concept + TOC → quotes, thinkers per chapter     │
+   │  Validate: bank_formatter.py --validate           │
+   └───────────────────────────────────────────────────┘
+              ↓
+   ┌─────────────────────────────────────────┐
+   │  PHASE 5: TABLE OF CONTENTS             │
+   │  Full TOC → user approves               │
+   │  Verify: toc_extract.py                 │
+   └─────────────────────────────────────────┘
+              ↓
+   ┌────────────────────────────────────────────────────────────┐
+   │  PHASE 6: CHAPTER LOOP  (repeats for every chapter)        │
+   │                                                            │
+   │  6A. Continuity Brief ←───────────────────────────────┐   │
+   │      log_manager.py show + conflict_check.py           │   │
+   │        ↓                                               │   │
+   │  6B. Chapter Writer (Guided or Autonomous)             │   │
+   │        ↓                                               │   │
+   │  6C. Chapter Auditor                                   │   │
+   │      dna_scan.py → scan_ai_patterns.py → score_report.py   │
+   │        ↓ PASS                                          │   │
+   │  6D. Humanizer Pass                                    │   │
+   │      DNA protection → strip 25 patterns → DNA rewrite  │   │
+   │        ↓                                               │   │
+   │  6E. Continuity Log Update ────────────────────────────┘   │
+   │      log_manager.py + word_count.py                        │
+   └────────────────────────────────────────────────────────────┘
+              ↓
+   ┌──────────────────────────────────────────────────────┐
+   │  PHASE 7: KDP DOCX OUTPUT                            │
+   │  Format → kdp_check.py → validate → deliver          │
+   └──────────────────────────────────────────────────────┘
+```
+
+---
+---
+
+# PHASE 1 — CONCEPT EXPANDER
+
+## TRIGGER
+Activate Phase 1 when the user provides:
+- A vague idea, single title, rough paragraph, or loose bullet points
+- Anything too thin to build structure options from
+
+If a complete concept document is uploaded → **skip to Phase 2**.
+After generating the concept document → run `python scripts/validate_concept.py <file>`.
+Do not proceed to Phase 2 if validation scores below 80% or any required field fails.
+
+## PROCESS
+
+### Step 1 — Seed Extraction
+Identify:
+- Core emotional or intellectual question at the heart of the idea
+- Implied audience and their pain
+- Genre fit (Philosophy/Essay | Self-help/Motivational | Non-fiction/Technical)
+- Natural tension or paradox (this is the engine)
+
+### Step 2 — Clarifying Questions (only if seed has under 3 usable signals)
+Ask ONE round, max 4 questions:
+```
+1. Who is this for — what is their core pain or question?
+2. What should they feel or understand by the last page that they don't now?
+3. Is there a personal story at the root of this idea?
+4. Any existing books in this space that this should be different from?
+```
+
+### Step 3 — Concept Document Output
+
+```
+BOOK CONCEPT DOCUMENT
+═══════════════════════════════════════════════════════
+
+WORKING TITLE: [compelling title + optional subtitle]
+ALTERNATE TITLES: [2 other options]
+
+LOGLINE (one sentence, under 25 words):
+[Entire premise. Banned words: "journey," "transformation," "discover"]
+
+THE CORE QUESTION:
+[A genuine question, not a topic.]
+
+THE READER:
+  Who they are:             [demographic + psychographic]
+  What they feel right now: [current emotional state]
+  What they've tried:       [what hasn't worked]
+  What they secretly want:  [real desire beneath surface problem]
+
+THE PROMISE:
+[Specific — what they'll have after finishing. Not "a better life."]
+
+GENRE:    [Vivid's category]
+TONE:     [raw/confessional | philosophical | instructional-personal]
+REGISTER: [intimate/direct | authoritative/grounded | exploratory]
+
+THE UNIQUE ANGLE:
+[What Vivid's lived experience adds that a generic author couldn't.]
+
+THE CENTRAL TENSION OR PARADOX:
+[The paradox the book lives in.]
+
+EMOTIONAL ARC:
+  Start:  [reader's emotional state on page 1]
+  Middle: [what they confront — harder before better]
+  End:    [what shifts — not solved, but transformed]
+
+THEMATIC CLUSTERS (7–10):
+  1. [Theme + one sentence on territory]
+  2–10. [...]
+
+SIGNATURE ELEMENTS:
+[Recurring devices or structural choices.]
+
+COMPARABLE BOOKS:
+[2–3 real comp titles + how this is different from each]
+
+ESTIMATED SCOPE: [short / medium / full-length]
+
+POTENTIAL CHAPTER TERRITORIES (8–14):
+[Possible chapter-level questions or themes]
+
+RISKS / BLIND SPOTS:
+[What could make this fall flat?]
+
+═══════════════════════════════════════════════════════
+```
+
+### Quality Checks
+- [ ] Logline under 25 words, no banned terms
+- [ ] Core question is a genuine question, not a topic
+- [ ] Unique angle is specific to Vivid's lived experience
+- [ ] Thematic clusters are distinct
+- [ ] Emotional arc has a real, difficult middle
+- [ ] `validate_concept.py` scores ≥ 80%
+
+→ Proceed to Phase 2.
+
+---
+---
+
+# PHASE 2 — STRUCTURE OPTIONS
+
+Present 3 structure options. User picks one.
+
+```
+STRUCTURE OPTIONS
+═══════════════════════════════════════════════════════
+
+OPTION A — [Name, e.g., "Focused & Tight"]
+Parts: [N] | Chapters: [N] | ~[X]–[Y]k words | ~[Z] pages
+Logic: [why this fits the concept]
+Chapter sketch:
+  Introduction: [title]
+  Ch.1: [title] | Ch.2: [title] | ...
+
+OPTION B — [Name]
+[same format]
+
+OPTION C — [Name]
+[same format]
+
+═══════════════════════════════════════════════════════
+Which structure resonates? You can mix elements.
+```
+
+Wait for user choice. Confirm hybrid if mixing.
+→ Proceed to Phase 3.
+
+---
+---
+
+# PHASE 3 — INTAKE QUESTIONNAIRE
+
+Lock all 26 decisions. Ask all at once. Use defaults for skipped answers, flag them.
+
+```
+BOOK INTAKE QUESTIONNAIRE
+═══════════════════════════════════════════════════════
+
+A. FORMATTING & PUBLISHING
+ 1. Trim size: 5×8 / 6×9 / 8.5×11 / other?
+ 2. Font preference: [default: Georgia 11pt]
+ 3. Line spacing: 1.15 / 1.5 / double?
+ 4. Chapter titles: numbers only / names only / both?
+ 5. Part dividers? Yes / No
+ 6. Headers: book title left + chapter title right? Or other?
+ 7. Page numbers: bottom center / bottom outside / top outside?
+
+B. VISUAL ELEMENTS
+ 8. Text-only or images? → If images: AI prompts / supply files? → Color or B&W?
+ 9. Chapter opener decoration? Yes / No / Specify
+10. Pull quotes or callout boxes? Yes / No
+
+C. FRONT & BACK MATTER
+11. Foreword or Preface? I draft / You write / Skip
+12. Dedication? Yes (text?) / No
+13. "How to Use This Book" section? Yes / No
+14. About the Author: existing bio / write new / skip
+15. "Stay in Touch" page? Yes / No
+16. Acknowledgments? Yes / No
+
+D. CHAPTER STRUCTURE
+17. "My Story + My Reflection" dual-section format? Yes / No / Modified
+18. For technical books: exercises, code blocks, case studies? Yes / No
+19. Reflection questions at chapter end? Yes / No
+20. Epigraph per chapter? Yes / No
+
+E. WRITING MODE
+21. GUIDED (you provide notes per chapter) or AUTONOMOUS (I write from concept alone)?
+
+F. VOICE & CONTENT
+22. Personal stories to include? [list or upload]
+23. Topics OFF LIMITS?
+24. Desired total length? [or defer to structure]
+25. Author name on cover? [real / pen name / confirm "Dheelep N"]
+26. Specific quotes or thinkers to include or avoid?
+
+═══════════════════════════════════════════════════════
+```
+
+Confirm: "All inputs locked. Building Research Bank and TOC."
+→ Proceed to Phase 4.
+
+---
+---
+
+# PHASE 4 — RESEARCH AGGREGATOR
+
+Build the Research Bank. Save as `research_bank.json`.
+Validate after building: `python scripts/bank_formatter.py research_bank.json --validate`
+Fix any flagged issues before proceeding.
+
+Read `references/thinkers_reference.md` before selecting any sources.
+
+## Research categories per chapter
+
+**A. Philosophical Anchors** (max 1 per chapter appears in final text)
+Format: `[Thinker/Work] — [Core idea] — [Chapter connection]`
+
+Priority thinkers for Vivid's genres:
+- Solitude / inner life: Rilke, Thoreau, Pascal, Kierkegaard, Montaigne
+- Philosophy of identity: Epictetus, Marcus Aurelius, Sartre, Camus, Beauvoir
+- Psychology of self: Viktor Frankl, Carl Rogers, Nathaniel Branden
+- Eastern frameworks: Stoic practice, non-attachment, witness consciousness
+- Modern psychology: attachment theory, self-determination theory, ACT
+
+**B. Empirical Anchors** (1–2 per chapter max)
+Format: `[Finding in plain English] — [Source] — [Application]`
+
+**C. Quotes** (under 30 words, not overexposed)
+Format: `"[Quote]" — [Author, Work]`
+
+**D. Conceptual Oppositions**
+Best argument against the chapter's central claim.
+
+## Research Bank output format
+
+```
+RESEARCH BANK — [Book Title]
+═══════════════════════════════════════════════════════
+
+GLOBAL ANCHORS:
+[3–5 ideas thematic to the whole manuscript]
+
+CHAPTER [N]: [Title]
+─────────────────────────────────────
+Theme: [one sentence]
+Philosophical anchors: [items]
+Quotes: ["quote" — Author]
+Research: [finding + source]
+Opposition: [counterargument]
+Writing notes: [how to deploy in Vivid's style]
+═══════════════════════════════════════════════════════
+```
+
+→ Proceed to Phase 5.
+
+---
+---
+
+# PHASE 5 — TABLE OF CONTENTS
+
+Build full TOC. Verify with `python scripts/toc_extract.py <manuscript.md>` after writing begins.
+Do NOT start writing until TOC is explicitly approved.
+
+```
+TABLE OF CONTENTS — [Book Title]
+═══════════════════════════════════════════════════════
+
+FRONT MATTER
+  Half-title | Title | Copyright | Dedication | How to Use | TOC
+
+BODY
+  Introduction: [Title]
+  PART I: [Title]
+    Chapter 1: [Title]
+    Chapter 2: [Title]
+  ...
+
+BACK MATTER
+  Epilogue | About the Author | Stay in Touch | Acknowledgments
+
+─────────────────────────────────────────────────────
+Estimated: ~[X]k words / ~[Y] pages at [trim size]
+═══════════════════════════════════════════════════════
+```
+
+→ Proceed to Phase 6 loop.
+
+---
+---
+
+# PHASE 6 — CHAPTER LOOP
+
+Run all 5 sub-phases for every chapter. Loop until all chapters complete.
+
+---
+
+## PHASE 6A — CONTINUITY BRIEF
+
+**Initialize log on Chapter 1:**
+```bash
+python scripts/log_manager.py init "Book Title"
+```
+
+**Before every chapter — run:**
+```bash
+python scripts/log_manager.py show
+python scripts/conflict_check.py <previous_chapter.md>
+```
+
+Output the brief:
+
+```
+CONTINUITY BRIEF — Before Writing Chapter [N]: [Title]
+═══════════════════════════════════════════════════════
+ESTABLISHED (relevant to this chapter):
+• [facts, tone decisions, commitments]
+
+OPEN THREADS TO RESOLVE HERE:
+• [threads needing payoff]
+
+METAPHORS TO AVOID (retired):
+• [list]
+
+INSIGHTS NOT TO REPEAT:
+• [already delivered]
+
+TONE REMINDERS:
+• [chapter-specific considerations]
+
+WORD COUNT GUIDANCE:
+• Prior avg: [X] words. Target [Y–Z] for this chapter.
+═══════════════════════════════════════════════════════
+```
+
+---
+
+## PHASE 6B — CHAPTER WRITER
+
+### GUIDED MODE
+1. Present chapter title + 2–3 sentence writing brief
+2. Ask: "What key points, stories, or insights should I include?"
+3. Wait for input → write using inputs + Style DNA below
+
+### AUTONOMOUS MODE
+Write without waiting. Use concept doc, Research Bank, TOC, questionnaire answers, Style DNA, Continuity Brief.
+Checkpoint every 2–3 chapters: "Chapters [X–Y] complete. Any direction changes?"
+
+### Chapter length targets
+| Scope | Words per chapter |
+|---|---|
+| Short (under 20k total) | 1,500–2,000 |
+| Medium (20–50k) | 2,000–3,500 |
+| Full-length (50k+) | 3,500–5,000 |
+
+### Default structure (self-help / philosophy)
+
+**My Story** — raw first-person narrative. Past tense. Immersive. Confessional. No hedging.
+
+**My Reflection and Insights** — philosophical synthesis. Present tense. Oscillates I/you.
+Moves from personal pain outward to universal truth.
+
+### Pain-to-Transformation Arc (every chapter, no exceptions)
+```
+1. ENTER THE PAIN       — at least 2–3 paragraphs. don't rush.
+2. WRONG ATTEMPTS       — show the failing strategies. readers recognize themselves here.
+3. THE REALIZATION      — arrives from inside the experience, not external advice.
+4. THE SHIFT            — perspective change, not a tidy fix.
+5. THE LANDING          — one sentence that reframes everything. does not summarize.
+```
+
+### Opening rule
+NEVER open with: definition / "In today's world" / statistics paragraph
+DO open with: scene / confession / question that makes the reader stop
+
+### Closing rule
+End on a single reframe — not a summary. The closing line should feel earned.
+
+---
+
+## PHASE 6B-DNA — VIVID'S WRITING STYLE DNA
+*Extracted from Master of Being Alone by Dheelep N*
+
+### Voice — Raw, Personal, Never Preachy
+- Leads with vulnerability. Arrives at insight through pain, not theory.
+- Never announces wisdom. Lets experience speak.
+- Addresses reader directly: "you," "my friend," rhetorical questions.
+- Honest about failure, ambiguity, not-yet-knowing.
+- Always in-process — never above the reader.
+
+Sample: *"I won't sugarcoat it. The path to mastering solitude is not a gentle stroll through a sunlit meadow."*
+
+### Sentence Rhythm
+- Short punchy sentences land the blow: "It didn't work." / "I failed every time."
+- Long sentences build emotional complexity then release.
+- Semicolons used naturally for rhythm.
+- Rhetorical questions as pivots, not decoration.
+- No 2+ consecutive sentences of the same structure.
+
+### Paragraph Density
+- 3–7 sentences per paragraph. Never a wall of text.
+- Each paragraph has one central beat.
+- No filler. No restatement.
+
+### Metaphors & Imagery
+- Physical, grounded, sensory. Maps emotional states to physical sensations.
+- Protected examples: "knife that cut me deeper every day" / "beggar asking for scraps of attention" / "invisible chains beginning to loosen" / "silence pressed in like a physical weight"
+- Never recycle: journey, storm, turning point, crossroads, chapter of life
+
+### Philosophical Integration
+- Philosophy is embodied, not academic.
+- Arrives at conclusions through experience, then names them simply.
+- Uses paradox: "The more independent you become, the more connected you feel."
+- Max 1 external citation per chapter.
+
+### Signature Phrases (use sparingly, never formulaically)
+- "But here's..." — pivot to deeper truth
+- "The hardest part wasn't..." — reframe of the obvious
+- "Not [surface X], but [deeper Y]"
+- "That is what [X] did to me from the inside."
+- Direct reader address mid-paragraph: "...and you start to see."
+- Single standalone closing sentence that reframes everything
+
+### Vocabulary Register
+- Accessible but emotionally elevated.
+- Protected elevated words: "insidious," "imperceptibly," "agonizing," "desolate," "conspicuously," "grotesque," "suffocating," "unraveling"
+- Avoids jargon, buzzwords, self-help clichés, passive voice.
+
+### Chapter Titles
+Short. Direct. Declarative or interrogative. The title is a promise.
+Examples: "Why It Hurts So Much" / "Stop Chasing Connection" / "Aloneness Is Not the Enemy"
+
+### Hard Anti-Patterns (never write these)
+- Generic motivational fluff: "you've got this," "believe in yourself"
+- Academic distance: "it can be argued that," "research suggests"
+- Bullet-point wisdom inside narrative sections
+- Rushed resolution (pain → fix in under 3 paragraphs)
+- More than 1 external quote per chapter
+- Any sentence that could appear on LinkedIn
+- Arriving at conclusions too cleanly
+
+### Genre adjustments
+
+**Philosophy / Essay**
+- Chapters can be shorter (1,500–2,500w) and denser
+- "My Story" optional — pure essay acceptable on some chapters
+- Tone: contemplative, probing, occasionally unsettling
+
+**Self-Help / Motivational**
+- Always use My Story + My Reflection structure
+- Personal anecdotes are the engine — never strip for general advice
+- Avoid: 3-step formulas, numbered tips in narrative, toxic positivity
+
+**Non-Fiction / Technical**
+- Structure: The Problem → The Concept → How It Works → In Practice
+- Vivid's voice still applies: open with real scenario, not a definition
+- Tone: clear, direct, grounded — never textbook flat
+
+---
+
+## PHASE 6C — CHAPTER AUDITOR
+
+Run after every chapter draft. Scripts first, then qualitative audit.
+
+**Run scripts:**
+```bash
+python scripts/dna_scan.py <chapter_file>          # identify protected constructions
+python scripts/scan_ai_patterns.py <chapter_file>  # flag AI patterns
+python scripts/score_report.py --chapter "Title"   # generate scored report
+```
+
+**Score 7 dimensions (1–5 each):**
+
+| Dimension | What to check |
+|---|---|
+| Voice Authenticity | Opens with scene/confession, immersive My Story, no LinkedIn sentences |
+| Pain-to-Transformation Arc | Pain entered fully, wrong attempts shown, realization from within, perspective shift |
+| Sentence Rhythm | Varied length, short punchy sentences, rhetorical questions as pivots |
+| Metaphor & Imagery | Physical/grounded, specific, no recycled images, min 2 strong images |
+| Structural Integrity | Sections demarcated, one beat per paragraph, within word count |
+| AI Pattern Contamination | Clean of all 25 patterns (5=clean, 1=heavy) |
+| Reader Resonance | Specific pain named, "how did he know" moment, author in-process |
+
+**Audit report format:**
+```
+CHAPTER AUDIT — [Title]
+Words: [X] / Target: [Y]
+═══════════════════════════════════════════════════════
+Voice Authenticity          [X/5]
+Pain-to-Transformation Arc  [X/5]
+Sentence Rhythm             [X/5]
+Metaphor & Imagery          [X/5]
+Structural Integrity        [X/5]
+AI Pattern Contamination    [X/5]
+Reader Resonance            [X/5]
+───────────────────────────────────
+OVERALL                     [X/35]
+VERDICT: PASS / CONDITIONAL PASS / REVISE
+
+REQUIRED FIXES (≤3): [specific line callouts + fix direction]
+RECOMMENDED (4): [optional]
+STRONGEST MOMENTS: [2–3 lines to protect]
+═══════════════════════════════════════════════════════
+```
+
+**Verdicts:**
+- 28–35 → PASS → proceed to 6D
+- 21–27 → CONDITIONAL PASS → proceed to 6D, fix after
+- Under 21 → REVISE → back to 6B with fix list
+
+---
+
+## PHASE 6D — HUMANIZER PASS
+
+Only runs on PASS or CONDITIONAL PASS. Never on REVISE.
+
+**Always run scripts first:**
+```bash
+python scripts/dna_scan.py <chapter_file>         # Step 1 — find protected constructions
+python scripts/scan_ai_patterns.py <chapter_file> # Step 2 — find AI patterns to eliminate
+```
+
+Read `references/patterns_quick_ref.md` for the full trigger word list.
+
+### Step 1 — DNA Protection
+
+Tag all protected constructions BEFORE eliminating anything.
+
+| Protected | Rule |
+|---|---|
+| Negative parallelisms: "Not the silence, but what it forced me to face" | Grounded in specific experience → protect |
+| Parallel peaks: "The freedom to love... The freedom to walk..." | Earned accumulation → protect |
+| Weighted em dashes around emotional phrases | "That thought — that single image —" → protect |
+| Signature phrases | "But here's..." / "The hardest part wasn't..." → always protect |
+| Elevated vocabulary naming felt experience | "insidious," "agonizing," "suffocating" → protect |
+| Physical imagery: "pressed down on my chest" | Specific + grounded → protect |
+
+### Step 2 — Eliminate AI Patterns (25 patterns — unprotected text only)
+
+**Content:** Significance inflation · Notability claims · Superficial -ing endings · Promotional language · Vague attributions · Challenges/Future sections
+
+**Language:** AI vocabulary (delve, tapestry, nuanced, embark, realm, foster, transformative, impactful, robust, unpack, holistic, synergy, leverage, navigate-as-metaphor) · Copula avoidance (serves as → is) · Negative parallelisms (generic) · Rule of three (generic) · Synonym cycling · False ranges
+
+**Style:** Em dash overuse (generic connectors) · Boldface overuse · Inline-header lists · Title case headings · Emojis · Curly quotes
+
+**Communication:** Chatbot artifacts (I hope this helps, let me know, of course!) · Knowledge-cutoff disclaimers · Sycophantic tone
+
+**Filler:** Filler phrases (in order to, due to the fact that, at this point in time, it is important to note, at its core, in today's world, in conclusion) · Hedging overload · Generic positive conclusions · Hyphen overuse (cross-functional, data-driven, decision-making)
+
+### Step 3 — Vivid Style Rewrite
+
+Actively reshape toward his DNA — not passive editing:
+
+- **Voice check:** Does it sound like confessing or presenting? → Rewrite as confessing.
+- **Arc check:** Pain → insight too fast (under 3 paragraphs)? → Insert wrong attempt + cost.
+- **Rhythm check:** All sentences same length? → Add one very short sentence after a long accumulation.
+- **Imagery check:** Abstract emotional claim with no physical sensation? → Ground it in the body.
+- **Opening check:** Does it still drop the reader in cold? → Re-read and verify.
+- **Closing check:** Does it reframe or summarize? → Must reframe.
+
+**Six rewrite techniques:**
+1. **Confession Opening** — replace general claim with personal scene
+2. **Slow the Arc** — insert wrong attempt → why it failed → cost → then insight
+3. **Ground the Abstract** — add specific personal detail before the claim + physical image after
+4. **Punch Landing** — replace soft ending with one short direct sentence
+5. **Reader Address** — shift to "you" for one paragraph in Reflection sections
+6. **Reframe Close** — replace summary ending with single sentence that reframes everything
+
+### Step 4 — Self-Audit
+
+Ask: *"What makes this still obviously AI-generated or generic?"*
+- [ ] Rhythm too tidy? (even pacing, no variation)
+- [ ] Still no opinions — only neutral reporting?
+- [ ] Could any section appear in a press release? → rewrite
+- [ ] Author above the experience rather than inside it?
+- [ ] Any metaphors still recycled?
+- [ ] Opening still hooks? Closing still reframes?
+- [ ] Does this sound like Vivid — or just "good human writing"?
+
+### Step 5 — Output
+1. Draft rewrite (after DNA protection + pattern elimination)
+2. Self-audit bullets (what still reads AI or generic)
+3. Final version (after style rewrite + self-audit fixes) ← the deliverable
+4. Changes summary: Stripped / Protected / Added
+
+---
+
+## PHASE 6E — CONTINUITY LOG UPDATE
+
+After each finalized chapter:
+
+```bash
+python scripts/log_manager.py add-chapter [N] "[Title]"
+python scripts/log_manager.py add-fact "[new established fact]"
+python scripts/log_manager.py add-insight [N] "[core insight delivered]"
+python scripts/log_manager.py add-metaphor [N] "[metaphor used — now retired]"
+python scripts/log_manager.py add-thread "[any open thread introduced]"
+python scripts/word_count.py manuscript.md --target [short|medium|full]
+python scripts/log_manager.py summary
+```
+
+**Contradiction check before closing:**
+```bash
+python scripts/conflict_check.py <current_chapter.md>
+```
+
+If conflicts detected → surface with resolution options, do NOT auto-resolve.
+
+Confirm: "Log updated after Chapter [N]. [X] items added. [Y] threads open."
+→ Loop back to Phase 6A for next chapter.
+
+---
+---
+
+# PHASE 7 — KDP DOCX OUTPUT
+
+Run pre-flight check first:
+```bash
+python scripts/kdp_check.py <manuscript.docx> --trim [5x8|6x9|8.5x11]
+```
+
+Read `references/kdp_specs.md` for all trim dimensions, margin requirements,
+typography standards, and page count estimates before generating the DOCX.
+
+Use the docx skill at `/mnt/skills/public/docx/SKILL.md` for generation.
+
+## KDP Trim Dimensions (DXA — 1440 DXA = 1 inch)
+
+| Trim | Width | Height | Outside | Inside (gutter) | Top | Bottom |
+|---|---|---|---|---|---|---|
+| 5×8 | 7,200 | 11,520 | 864 | 1,008 | 1,008 | 864 |
+| 6×9 | 8,640 | 12,960 | 864 | 1,080 | 1,080 | 864 |
+| 8.5×11 | 12,240 | 15,840 | 1,008 | 1,260 | 1,260 | 1,008 |
+
+## Typography
+
+```javascript
+// Body: Georgia 11pt, 1.15 spacing
+{ font: "Georgia", size: 22, lineSpacing: { line: 276, lineRule: "auto" } }
+// Chapter title: Georgia 18pt bold centered
+{ font: "Georgia", size: 36, bold: true, alignment: CENTER, spacing: { before: 720, after: 480 } }
+// Section heading: Georgia 14pt bold
+{ font: "Georgia", size: 28, bold: true, spacing: { before: 480, after: 240 } }
+// Part title: Georgia 20pt bold centered, own page
+{ font: "Georgia", size: 40, bold: true, alignment: CENTER, spacing: { before: 2880, after: 2880 } }
+```
+
+## Headers & Footers
+- Even pages: book title, left-aligned
+- Odd pages: chapter title, right-aligned
+- Chapter first page: no header
+- Page numbers: bottom outside
+
+## Front Matter (in order)
+Half-title → Title → Copyright → Dedication → How to Use → TOC → blank if needed
+
+## Chapter Layout
+- Each chapter: new page (PageBreak before heading)
+- Body: justified, first paragraph no indent, subsequent paragraphs 360 DXA indent
+- Section headings: bold, left-aligned, 240 DXA space above
+
+## Validation
+```bash
+python scripts/office/validate.py manuscript.docx
+```
+If fails: unpack → fix XML → repack.
+
+---
+---
+
+# STYLE OVERRIDE
+
+If user uploads a new sample and says "match this style instead":
+1. Read sample (minimum 3,000 words)
+2. Extract style DNA using the 13-point framework from Phase 6B-DNA
+3. Flag: "Style overridden for this session. Using [book title] as reference."
+Vivid's embedded DNA remains the permanent default for new sessions.
+
+---
+
+# ERROR HANDLING
+
+| Situation | Action |
+|---|---|
+| Concept too vague | Run Phase 1 clarifying questions before proceeding |
+| validate_concept.py below 80% | Fix required fields before Phase 2 |
+| Skipped questionnaire answers | Use defaults, flag at top of affected chapters |
+| Chapter audit REVISE verdict | Return to Phase 6B — do NOT humanize a failing chapter |
+| Chapter over/under word count | Flag → ask: expand / trim / proceed as-is |
+| conflict_check.py finds conflicts | Surface with resolution options — do NOT auto-resolve |
+| kdp_check.py fails | Fix reported issues before delivering |
+| User skips a phase | Proceed from current state, flag gap and downstream impact |
+| Style override sample under 3,000 words | Ask for more text — sample too short for reliable extraction |
+
+---
+
+# QUICK REFERENCE
+
+```
+PHASE 1  Concept Expander      — seed → concept doc → validate_concept.py
+PHASE 2  Structure Options     — 3 proposals → user picks
+PHASE 3  Intake Questionnaire  — 26 questions → all decisions locked
+PHASE 4  Research Aggregator   — concept + TOC → bank → bank_formatter.py --validate
+PHASE 5  Table of Contents     — TOC → user approves → toc_extract.py
+PHASE 6A Continuity Brief      — log_manager.py show + conflict_check.py
+PHASE 6B Chapter Writer        — Guided or Autonomous + Style DNA
+PHASE 6C Chapter Auditor       — dna_scan + scan_ai_patterns + score_report
+PHASE 6D Humanizer Pass        — protect DNA → strip 25 patterns → rewrite → self-audit
+PHASE 6E Log Update            — log_manager + word_count + conflict_check
+PHASE 7  KDP DOCX Output       — kdp_check → format → validate → deliver
+```