v5.3.0 - guizang methodology absorption

v5.3.0 - guizang Methodology Absorption (5 blind spots fixed)

Inspired by op7418's methodology article: "做杂志，不做网页" / "限制不是阻碍，是底线" / "Skill 是一个产品"

新增功能

• Category Detection: Silent pre-Q1 content category identification (10 categories). Each category maps to default Mode + Palette + Key Visual Trait
• Screenshot Styling: Device frame components (macOS/iOS/Browser), material backgrounds, mode-specific shadow & radius rules
• Out of Scope declaration: 6 explicit boundaries — full article typesetting, video, pure image editing, fan content, hard-sell advertising, tutorials >15 cards
• Adaptive Font Size Rules: Title length → size/weight/line-clamp mapping with React utility function getTitleStyle()
• Design Decision "Why" Explanations: Added rationale to "The Larger, The Lighter" and "Aesthetic Guardrails"

变更

• multi-illustration.md Step C: added Category Detection before Q1
• image-sources.md: added Screenshot Styling section
• react-output-spec.md: added Adaptive Font Size Rules section
• aesthetics-guide.md: added "Why it works" / "Why restriction works" rationale blocks
• SKILL.md: added Out of Scope section before Mode Detection

v5.2.0 - guizang-skill absorption + image sources

v5.2.0 - guizang-skill Absorption + Image Sources

Added

• The Larger, The Lighter typography rule: Large text (56px+) uses weight 200-400, small text (10-12px) uses weight 500-650
• Dual Style System (Editorial vs Swiss): Two complete visual stances
  • 4 Swiss Accent Palettes: IKB Blue / Lemon Yellow / Lemon Green / Safety Orange
  • Swiss Gray Scale: 5-step calibrated premium gray
  • Style Identity Test: 4-rule Swiss test + 3-rule Editorial test
• Aesthetic Guardrails: Swiss mode limited to 4 accent palettes, Editorial mode limited to warm palettes, no cross-mode mixing
• Visual Rhythm Planning: hero/dark/light/accent theme classes across illustrations
• Swiss Gray Scale: Cross-accent unified premium gray system
• Image-Text Conflict Protection: Quiet zone test, subject mapping, object-position discipline, thumbnail test
• Image Sources: Free stock library integration - Pexels, Unsplash, Wallhaven
  • New reference file: references/image-sources.md

Changed

• Version bumped from 5.1.0 to 5.2.0
• SKILL.md Rules expanded from 21 to 27
• aesthetics-guide.md: added Phase 0.5, The Larger The Lighter, Aesthetic Guardrails, Image-Text Conflict Bans
• multi-illustration.md Step C: Q1 answers now map to Editorial/Swiss modes; added Visual Rhythm Planning

v5.1.0 - md2wechat-skill Absorption

v5.1.0 - md2wechat-skill Absorption

Added - md2wechat-skill Absorption (6 elements)

• 4-Purpose Decision Framework: Each illustration now tagged with purpose (attention/readability/memorability/conversion), driving silent design parameter adjustments
• 11 new illustration type templates: verdict-card, audience-fit-card, myth-fact-card, manifesto-card, bridge-card, callout-card, definition-card, cases-card, notice-card, series-card, subscribe-card
• AI Voice Decontamination (Category 8): 10 new anti-patterns (#39-48) for detecting and removing AI-generated text patterns
• Content Readiness Check: Pre-matching structural check in Step 1
• 3 new Brand DNA entries: ByteDance/????, ???/sspai, ????
• Uniqueness Constraint: Anti-pattern for type overuse + per-type max count table

Changed

• Version bumped from 5.0.0 to 5.1.0
• SKILL.md Step 1: added Content Readiness Check
• density-standards.md: expanded from 38 to 48 anti-pattern rules
• content-layout-mapping.md: Brand DNA Registry expanded from 10 to 13 brands
• multi-illustration.md: Step A table expanded from 10 to 20 extract types

v4.2.0 - Refined Multi-Illustration Templates

v4.2.0

Added

• 3 knowledge card sub-types: rule-card, checklist-card, cheatsheet-card
• 3 logic visualization sub-types: logic-chain, process-pipeline, version-timeline
• Cover/back-cover density gate exemption (≥6/15 vs ≥9/15)

Changed

• Logic visualization now distinguishes causal chains vs pipelines vs timelines
• Knowledge cards now have 4 sub-types with selection guide

v4.1.0 - Multi-Illustration Mode

v4.1.0

Added

• Multi-Illustration Mode: one article → multiple illustrations
• 5 new templates: cover, back-cover, quote-card, section-divider, logic-chain
• 3-dimension 15-point density gate per illustration (≥9/15)
• Two mandatory confirmation points: content plan + style plan
• WeChat public account size specs
• Content drives quantity, not templates

Changed

• SKILL.md restructured with Mode Detection
• Rules expanded to 15 (added #14 content-driven quantity, #15 per-illustration density gate)

v4.0.0 - Chart System + Brand Profile + Kami Tokens

v4.0.0

Added

• Chart system: 14 chart types with auto-selection decision tree
• Brand Profile system: 4-layer brand configuration
• Kami Full Token System: 20+ CSS variables
• Extended Anti-Patterns: 7 categories, 38 rules
• Writing quality gate: assertion-evidence, Impact formula
• Document type presets: 9 Kami-inspired templates
• Multi-language font stacks: Japanese + Korean
• Slide scaling formula
• Step 0: Brand Profile in execution flow

Changed

• Version 3.0.0 → 4.0.0
• Execution flow: 4 steps → 5 steps
• Rules: 10 → 13

v3.0.0 - React Design Draft Generator

v3.0.0

Added

• Three-dimension combination system: Layout × Style × Palette (5,086 combinations)
• 25 layout patterns
• 23 visual styles
• 12 color palettes
• 7 font presets with local-first CJK strategy
• 29 quick presets with keyword triggers
• 5-dimension density scoring (25-point scale)
• Kami Ten Invariants
• Pre-generation consultation + Post-generation edit guide
• React 4-piece output specification