K2F Idml — AI-Native Aesthetic Design

K2F IDML — AI 原生美学设计

Purpose

目的

Teach agents to author

content_and_design.md

for publications meant to open in InDesign: explicit copy and assets, spread-level story beats, typographic roles, and art direction—not a slide deck pasted into pages or a wall of undifferentiated body text.

教授 Agent 为需在 InDesign 中打开的出版物编写

content_and_design.md

：提供明确的文案与素材、跨页级叙事节拍、字体排印角色和艺术指导，而不是把幻灯片演示稿粘贴到页面，或堆砌一整墙未经区分的正文。

Built on K2F

基于 K2F 构建

This skill produces print-ready packages as

.K2F

documents exportable to IDML. Follow the k2f skill for authoring: unpack or init a workspace, edit the semantic tree, run pack/verify, then export. K2F keeps agents on meaning (stories, figures, hierarchy roles)—not frame coordinates—while the engine compiles layout and typography into an editable InDesign handoff.

Before you build, install the core skill:

npx skills add suzheng/k2f --skill k2f

.

此 skill 可生成印刷就绪的

.K2F

文档，并导出为 IDML。创作时遵循 k2f skill：解包或初始化工作区，编辑语义树，运行 pack/verify，然后导出。K2F 让 Agent 专注于语义（故事、图示、层级角色），而非坐标框；引擎会将版面与字体排印编译成可在 InDesign 中编辑的交付文件。

构建前，先安装核心 skill：

npx skills add suzheng/k2f --skill k2f

。

When to Use

适用场景

New or refreshed brochures, magazines, catalogs, annual reports, program books, or brand lookbooks from briefs or draft copy.
Restructuring long marketing copy into scannable editorial spreads before a designer refines in InDesign.
Delivering native text and images to production—not a flattened PDF or full-page screenshot workflow.

根据 brief 或草稿文案，新建或更新宣传册、杂志、产品目录、年度报告、会议手册或品牌样册。
在设计师继续于 InDesign 中细化之前，将冗长的营销文案重构为可快速浏览的编辑型跨页。
向制作团队交付原生文本和独立图片，而不是采用扁平化 PDF 或整页截图工作流。

Workflow

工作流

Step 1. Generate the content_and_design.md.

第 1 步：生成 content_and_design.md。

One blueprint merging reader path, page/spread roles, typed content blocks, and design intent the engine can interpret. Use labeled sections; map spreads later in k2f.

Sub-steps

Publication scenario — Audience, occasion (launch, conference, retail shelf), reading mode (skim vs study), approximate page count, binding or fold (saddle, perfect, tri-fold), and what must stay verbatim (brand claims, legal, pricing). Drop stories that do not serve the scenario.
Document skeleton — Ordered parts: cover (and optional wrap), front matter (TOC, letter, credits), body sections/chapters, back matter (index, colophon, contact). Note which spreads are opening, feature, text-heavy, gallery, or closing so rhythm is not uniform page-to-page.
Content inventory — Per spread or section, list blocks with stable ids and roles: display headline, deck/subhead, body (with approximate word budget), pull quote (+ attribution), sidebar/fact box, figure (+ caption + credit), table (title + column semantics), label/caption, CTA or contact strip. Mark must-run user copy vs agent-drafted placeholders the user must replace.
Information architecture — One dominant idea per opening spread; support with secondary blocks, not competing headlines. Sequence: hook → proof → detail → action. Pair every figure with caption on the same spread; keep related sidebar beside the paragraph it annotates. For catalogs, group by category with consistent product fields (name, spec, price, SKU)—do not mix narrative essays between SKU grids.
Design intent — Tone (luxury quiet, newsroom energetic, institutional sober), typographic hierarchy (display vs deck vs body vs caption sizes and weights—relative, not points), density (airy feature spreads vs tight reference pages), column intent (single narrative column vs multi-column body vs full-bleed hero), color (dominant neutrals + one accent for rules, labels, CTAs), imagery (subjects, crop style, negative space, when text overlays photo vs sits beside). State contrast target (readable captions on photos; no low-contrast gray body).

Rules & constraints

Spread thinking: plan left–right pairs as a unit; avoid orphan headlines at the foot of a spread.
Role discipline: each block has one role; do not stack three headline-sized lines without a clear primary.
Word budgets: cap body per spread; move overflow to the next spread or a sidebar—never shrink type in the blueprint.
Verbatim gate: trademarks, disclaimers, prices, and legal lines only from user sources.
Designer handoff: prefer editable text and discrete figures over describing “designed as one image.”

Common failure modes

Slide-deck pacing (title + bullets every page) → assign editorial roles and varied spread types.
Uniform template fatigue → alternate feature, text, and gallery spreads.
Captions separated from figures → lock figure+caption pairs in the inventory.
Missing entry points → add pull quotes, decks, or section openers for skimmers.
Generic stock-photo direction → specify subject, mood, and crop per slot.
Catalog noise → separate storytelling spreads from SKU grids; consistent field order.
Invented prices, awards, or compliance text → flag gaps; never fabricate.

Quality checklist

Skeleton lists every major part; each body section has spread roles assigned.
Every figure/table has caption, credit when needed, and same-spread pairing.
Headlines are unique and ranked (one primary per spread).
Word budgets sum to the stated page count without cramming.
Verbatim copy is quoted or flagged; no silent invention of legal or pricing.
Design intent covers tone, type roles, density, columns, color, and imagery without coordinates.
Closing spread delivers the scenario CTA (contact, QR context, next step).

创建一份融合读者路径、页面/跨页角色、类型化内容区块和设计意图的蓝图，供引擎解读。使用带标签的分节；稍后在 k2f 中映射跨页。

子步骤

出版场景 — 受众、使用场合（发布活动、会议、零售货架）、阅读模式（略读 vs 精读）、大致页数、装订或折叠方式（骑马订、无线胶订、三折），以及必须逐字保留的内容（品牌声明、法律信息、价格）。删除不服务于该场景的故事。
文档骨架 — 按顺序列出各部分：封面（可选环绕封面）、前置内容（目录、致辞、鸣谢）、正文部分或章节，以及后置内容（索引、版权页、联系方式）。标明哪些跨页属于开篇、专题、文字密集、图集或收尾，避免页面之间的节奏完全一致。
内容清单 — 按跨页或章节列出内容区块，为其设置稳定的 id 和角色：展示标题、导语/副标题、正文（附大致字数预算）、醒目引文（附署名）、侧栏/信息框、图片（附图注与来源）、表格（标题与列语义）、标签/说明文字、CTA 或联系信息条。标记必须逐字保留的用户文案，以及由 Agent 起草且必须由用户替换的占位内容。
信息架构 — 每个开篇跨页只设一个主导概念；用次级区块加以支撑，避免多个标题相互争夺注意力。顺序为：吸引点 → 佐证 → 细节 → 行动。每张图片都应与图注配对并置于同一跨页；相关侧栏应紧邻其所注释的段落。对于产品目录，按类别分组并保持一致的产品字段（名称、规格、价格、SKU）；不要在 SKU 网格之间穿插叙事性文章。
设计意图 — 基调（奢华静谧、新闻编辑室式活力、机构沉稳感）、字体排印层级（展示标题、导语、正文、图注的相对字号与字重——使用相对关系，不使用具体磅值）、信息密度（疏朗的专题跨页 vs 紧凑的参考页）、栏式意图（单栏叙事 vs 多栏正文 vs 满版主视觉）、色彩（以中性色为主，搭配一个用于分隔线、标签和 CTA 的强调色）、图像（主题、裁切方式、留白，以及何时将文字叠加在照片上、何时置于照片旁）。说明对比度目标（照片上的图注应清晰可读；正文不得使用低对比度灰色）。

规则与约束

跨页思维：将左右页对视为一个整体进行规划；避免标题孤零零地落在跨页底部。
角色纪律：每个区块只承担一种角色；不要在没有明确主标题的情况下连续堆叠三行标题级文字。
字数预算：为每个跨页的正文设置上限；溢出的内容移至下一跨页或侧栏——绝不要在蓝图中缩小字号。
逐字保留校验：商标、免责声明、价格和法律声明只能取自用户提供的来源。
设计师交付：优先使用可编辑文本和独立图片，不要将其描述成“设计为一张整图”。

常见失败模式

幻灯片式节奏（每页都是标题 + 项目符号）→ 为内容分配编辑角色，并采用多样化的跨页类型。
模板疲劳、千篇一律 → 交替使用专题、文字和图集跨页。
图注与图片分离 → 在内容清单中锁定图片与图注的配对关系。
缺少阅读入口 → 增加醒目引文、导语或章节开篇页，帮助略读者快速进入内容。
泛泛的 stock photo 指导 → 为每个图片位明确主题、氛围和裁切方式。
产品目录噪声 → 将叙事型跨页与 SKU 网格分开；保持字段顺序一致。
编造价格、奖项或合规文本 → 标记缺口；绝不虚构。

质量检查清单

骨架列出了所有主要部分；每个正文部分都已分配跨页角色。
每张图片/表格都有说明文字，必要时附来源，并与对应对象位于同一跨页。
标题各不相同且层级明确（每个跨页仅有一个主标题）。
字数预算之和与既定页数相符，不会造成过度拥挤。
必须逐字保留的文案已加引号或明确标记；未擅自补写法律或价格信息。
设计意图涵盖基调、字体排印角色、信息密度、栏式、色彩和图像，且不依赖具体坐标。
收尾跨页包含场景所需的 CTA（联系方式、QR 码使用情境、下一步操作）。

Step 2. Use the content_and_design.md and k2f skill to generate a K2F file and export to an appropriate file format.

第 2 步：使用 content_and_design.md 和 k2f skill 生成 K2F 文件并导出为适当的文件格式。

Follow k2f to instantiate spreads from the blueprint, bind content roles from the inventory, verify, and export the IDML package for InDesign.

遵循 k2f，根据蓝图实例化跨页，按内容清单绑定内容角色，执行 verify，并导出可在 InDesign 中使用的 IDML 包。

k2f-idml

Original

Translation