Chinese DOCX Typesetting Baseline

Locked design-token baseline for Chinese Word/DOCX documents and WeChat-native (公众号) article layouts: body 16px, section head 18px, title 23px, deep-blue #1a3a5c headings, no HTML tables/cards/backgrounds. Use when typesetting a Chinese DOCX or WeChat article, when html-to-docx output loses styles or fails on Windows, or when successive documents must stay visually consistent. Trigger keywords: 中文 docx 排版, Word 中文文档美化, html to docx 样式丢失, 16 18 23 px 基线, 深蓝标题, 公众号排版, 微信文章排版, WeChat article typesetting, 字号基线, font size baseline, Windows html-to-docx 失败, DOCX 样式. Baseline changes must update this skill first.

Purpose: one locked baseline instead of per-task improvisation. Every Chinese DOCX produced from HTML follows the same token set, so successive documents look consistent and conversion failures have a known recovery path.

Design Tokens (locked)

Element	Token
Title (H1)	23px, deep blue #1a3a5c, bold
Section head (H2)	18px, deep blue #1a3a5c, bold
Body	16px, dark gray/black
Heading color	#1a3a5c
WeChat-native layout	vertical info bars + plain editorial style; NO HTML tables, cards, backgrounds, popup grids

Workflow

1. [Deterministic] Pick the target style: WeChat-native article (strict no-table/no-card rules) or generic Word document (tables allowed, same font tokens).
2. [Deterministic] Apply tokens: set title 23px / section 18px / body 16px, heading color #1a3a5c in the source HTML or converter config before conversion.
3. [Deterministic] Convert with the html-to-docx pipeline.
4. [Deterministic] Windows failure recovery: if the bundled converter fails on Windows, follow the html-to-docx recovery skill: pre-process the HTML (strip unsupported CSS constructs), retry once; if still failing, rebuild the document programmatically (e.g. python-docx) applying the same tokens paragraph by paragraph.
5. [Deterministic] Verify the output: reopen the DOCX; confirm heading sizes 23/18/16 present, heading color #1a3a5c applied, CJK glyphs render correctly (no tofu), text remains editable (no flattened images).

Hard Rules

• Never flatten text to images; the DOCX must stay fully editable.
• WeChat-native layouts: no HTML tables, cards, background colors, or popup grids — vertical info bars only.
• Baseline changes (sizes, colors) must update this skill first, then the documents. No per-task deviation.
• Headings stay anchored to topic content (who / scene / problem), never filename-style titles.

Failure Handling

• Converter crashes on Windows twice: switch to programmatic rebuild (python-docx) with the same tokens; do not debug the converter beyond two attempts.
• Styles lost after conversion: verify the HTML source carries inline styles or a converter-compatible stylesheet before blaming the converter; fix the source, reconvert once.
• CJK font issues: confirm the converter's font-fallback setting includes a Chinese font before changing the baseline.

Output Format

Report: target style chosen → conversion path used (direct / recovery / programmatic) → verification checklist results (sizes, color, editability) → output file path.