gpt-image-2-style-library

어떤 스킬인가요?

Choose GPT-Image2 / gpt-image-2 visual styles and industrial prompt templates from the awesome-gpt-image-2 style library. Use when an agent needs to create, rewrite, classify, or improve image-generation prompts with repository-backed templates, categories, style tags, scene tags, pitfalls, and example cases.

설치 방법

프로젝트 폴더에서 아래 명령어를 실행하고, 설치할 에이전트를 선택하세요.

설치 명령을 직접 실행해야 적용됩니다. 지원 에이전트와 필요한 권한·라이선스는 제작자의 안내를 확인하세요.

스킬 원문 파일

display_name: GPT-Image2 Style Library
short_description: Choose GPT-Image2 templates, styles, and prompt constraints.
default_prompt: Use the GPT-Image2 style library to create a production-ready image prompt.

#!/usr/bin/env node
import { cpSync, existsSync, mkdirSync, rmSync } from 'node:fs';
import { dirname, join } from 'node:path';
import { homedir } from 'node:os';
import { fileURLToPath } from 'node:url';

const skillName = 'gpt-image-2-style-library';
const args = process.argv.slice(2);
const commandOrTarget = args[0] || 'install';
const allowedCommands = new Set(['install', 'sync']);

let command = commandOrTarget;
let targetArgs = args.slice(1);
if (!allowedCommands.has(commandOrTarget)) {
  command = 'install';
  targetArgs = args;
}

const usage = [
  'Usage: gpt-image-2-style-library install [all|codex|claude-code|agents]',
  'Examples:',
  '  gpt-image-2-style-library install all',
  '  gpt-image-2-style-library install claude-code',
  '  gpt-image-2-style-library install codex'
].join('\n');

if (!allowedCommands.has(command)) {
  console.error(usage);
  process.exit(1);
}

const packageRoot = dirname(dirname(fileURLToPath(import.meta.url)));
const entries = ['SKILL.md', 'agents', 'assets', 'references'];
const targetDefinitions = {
  codex: {
    label: 'Codex',
    root: join(process.env.CODEX_HOME || join(homedir(), '.codex'), 'skills')
  },
  'claude-code': {
    label: 'Claude Code',
    root: join(process.env.CLAUDE_HOME || join(homedir(), '.claude'), 'skills')
  },
  agents: {
    label: 'Shared agent skills',
    root: join(process.env.AGENTS_HOME || join(homedir(), '.agents'), 'skills')
  }
};
const targetAliases = {
  all: Object.keys(targetDefinitions),
  codex: ['codex'],
  claude: ['claude-code'],
  'claude-code': ['claude-code'],
  agents: ['agents'],
  shared: ['agents']
};

for (const entry of entries) {
  const source = join(packageRoot, entry);
  if (!existsSync(source)) {
    throw new Error(`Missing package entry: ${entry}`);
  }
}

function selectedTargets(rawTargets) {
  const names = rawTargets.length ? rawTargets : ['all'];
  const selected = new Set();

  for (const name of names) {
    const normalized = name.toLowerCase();
    const matches = targetAliases[normalized];
    if (!matches) {
      console.error(usage);
      throw new Error(`Unknown target: ${name}`);
    }
    for (const match of matches) selected.add(match);
  }

  return [...selected].map((name) => targetDefinitions[name]);
}

for (const targetDefinition of selectedTargets(targetArgs)) {
  const target = join(targetDefinition.root, skillName);
  mkdirSync(targetDefinition.root, { recursive: true });
  rmSync(target, { recursive: true, force: true });
  mkdirSync(target, { recursive: true });

  for (const entry of entries) {
    cpSync(join(packageRoot, entry), join(target, entry), { recursive: true });
  }

  console.log(`Installed ${skillName} for ${targetDefinition.label}: ${target}`);
}

{
  "name": "gpt-image-2-style-library",
  "version": "1.0.4",
  "description": "Agent skill for choosing GPT-Image2 visual styles, prompt templates, categories, and scene tags.",
  "type": "module",
  "license": "MIT",
  "repository": {
    "type": "git",
    "url": "git+https://github.com/freestylefly/awesome-gpt-image-2.git",
    "directory": "agents/skills/gpt-image-2-style-library"
  },
  "homepage": "https://github.com/freestylefly/awesome-gpt-image-2#agent-skill",
  "bugs": {
    "url": "https://github.com/freestylefly/awesome-gpt-image-2/issues"
  },
  "keywords": [
    "agent-skill",
    "claude-code",
    "codex",
    "codex-skill",
    "gpt-image-2",
    "prompt",
    "image-generation",
    "style-library"
  ],
  "bin": {
    "gpt-image-2-style-library": "bin/install.mjs"
  },
  "files": [
    "SKILL.md",
    "agents",
    "assets",
    "bin",
    "references"
  ],
  "publishConfig": {
    "access": "public"
  }
}

# GPT-Image2 Style Library Reference

Generated from `data/style-library.json`. Use this file as the detailed index for choosing GPT-Image2 prompt templates, visual styles, categories, and scene tags.

## Selection Rules

- Match explicit product types to template categories first, such as product, poster, UI, infographic, brand, photography, character, or document.
- Match visual words to style tags next, such as realistic, 3D, illustration, classical, brand, poster, or UI.
- Match context words to scene tags next, such as commerce, education, social, food, travel, story, history, tech, or creative.
- If a request is vague, offer 2-3 strong template directions and ask the user to choose before writing the final prompt.
- Final output should include the selected template name, a copyable GPT-Image2 prompt, and concise constraints for text, aspect ratio, layout, and negative details.

## Template Index

### UI Screenshot System / UI 截图系统

- ID: `ui-screenshot-system`
- Category: UI & Interfaces
- Styles: UI
- Scenes: Tech, Social
- Tags: UI, Dashboard, Screenshot
- Cover: `/images/case17.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-ui
- Example cases: case 17, case 2, case 4

Use when:
- EN: Use for app screens, dashboards, social screenshots, and live interface mockups.
- ZH: 用于 App 截图、仪表盘、社媒截图和直播界面。

Guidance:
  - Lock platform, aspect ratio, layout hierarchy, and exact visible text.
  - Specify UI chrome such as status bars, tabs, action rows, or comment layers.
  - 锁定平台、比例、层级和画面文字。
  - 明确状态栏、Tab、操作区、评论层等 UI 元素。

Pitfalls:
  - Avoid vague platform names and generic app mockups.
  - Constrain text readability and platform-specific details.
  - 避免平台描述过泛。
  - 约束文字可读性和平台特征。

### Infographic Engine / 信息图引擎

- ID: `infographic-engine`
- Category: Charts & Infographics
- Styles: Infographic, Charts
- Scenes: Education, Tech
- Tags: Infographic, Chart, Education
- Cover: `/images/case334.png`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-infographic
- Example cases: case 334, case 1, case 8

Use when:
- EN: Use for explainer graphics, technical diagrams, timelines, and knowledge cards.
- ZH: 用于解释图、技术图解、时间线和知识卡片。

Guidance:
  - Define 3-5 modules, information flow, visual hierarchy, and short labels.
  - Use color groups, arrows, icons, and clean spacing to reduce clutter.
  - 定义 3-5 个模块、信息流、层级和短标签。
  - 用色块、箭头、图标和留白控制复杂度。

Pitfalls:
  - Avoid long paragraphs inside the image.
  - Limit module count before adding visual detail.
  - 避免把长段正文塞进画面。
  - 先限制模块数量，再补视觉细节。

### Scientific Scale Diagram / 科学尺度缩放图

- ID: `scientific-scale-diagram`
- Category: Charts & Infographics
- Styles: Infographic, Charts, Realistic
- Scenes: Education, Tech
- Tags: Infographic, Chart, Education
- Cover: `/images/case341.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-infographic
- Example cases: case 341

Use when:
- EN: Use when the topic needs micro-to-macro scale comparison and labeled detail windows.
- ZH: 用于需要从微观到宏观展示尺度变化的科普主题。

Guidance:
  - Use 6-8 scale frames and keep each label short.
  - Show units, magnification, and distinct scale detail.
  - 使用 6-8 个尺度框，每个标签保持短句。
  - 展示单位、倍率和不同尺度的细节。

Pitfalls:
  - Avoid making every scale frame visually identical.
  - Avoid generic magnifying glass icon layouts.
  - 避免所有尺度框长得一样。
  - 避免通用放大镜式布局。

### Poster Layout System / 海报排版系统

- ID: `poster-layout-system`
- Category: Posters & Typography
- Styles: Poster
- Scenes: Commerce, Social
- Tags: Poster, Typography, Campaign
- Cover: `/images/case345.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-poster
- Example cases: case 345, case 5, case 10

Use when:
- EN: Use for event posters, movie posters, covers, and social campaign visuals.
- ZH: 用于活动海报、电影海报、封面和社媒传播视觉。

Guidance:
  - Lock subject, headline, layout, palette, and aspect ratio.
  - Make the title hierarchy and primary visual clear.
  - 锁定主体、标题、版式、配色和比例。
  - 突出标题层级和主视觉。

Pitfalls:
  - Avoid mixed moodboards or process sheets when asking for one finished poster.
  - Constrain extra text and decorative symbols.
  - 需要成品海报时，避免生成拼贴展示板。
  - 约束多余文字和装饰符号。

### Sports Campaign Poster / 运动商业 Campaign

- ID: `sports-campaign-poster`
- Category: Posters & Typography
- Styles: Poster, Realistic
- Scenes: Commerce, Fashion
- Tags: Poster, Campaign, Typography
- Cover: `/images/case350.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-poster
- Example cases: case 350, case 3

Use when:
- EN: Use for sports brand campaigns, athlete posters, and product-led sport visuals.
- ZH: 用于运动品牌 Campaign、运动员海报和运动产品视觉。

Guidance:
  - Define sport, athlete pose, hero prop, title, and brand palette.
  - Use dramatic light, clean composition, and readable data overlays.
  - 定义运动项目、姿态、核心道具、标题和品牌色。
  - 使用强光影、干净构图和可读数据层。

Pitfalls:
  - Avoid wrong equipment and noisy collage.
  - Keep the athlete and hero prop visually dominant.
  - 避免错误运动器材和杂乱拼贴。
  - 让运动员和核心道具占据主导。

### Conceptual Typography Poster / 概念字体海报

- ID: `conceptual-typography-poster`
- Category: Posters & Typography
- Styles: Poster
- Scenes: Creative, Social
- Tags: Typography, Poster, Style
- Cover: `/images/case355.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-poster
- Example cases: case 355

Use when:
- EN: Use when the exact title must become the main visual structure.
- ZH: 用于标题文字需要成为主视觉结构的海报。

Guidance:
  - Make typography the hero and spell the title exactly.
  - Tie human figures, objects, or landscapes to the title meaning.
  - 让字体成为画面主角，并保证标题拼写准确。
  - 人物、物体或风景需要服务标题含义。

Pitfalls:
  - Avoid default word art, unrelated icons, and misspelled title text.
  - Limit the color system to a restrained palette.
  - 避免默认字效、无关图标和标题错字。
  - 控制配色数量，保持克制。

### Ink Double Exposure Poster / 水墨双重曝光海报

- ID: `ink-double-exposure-poster`
- Category: Posters & Typography
- Styles: Poster, Illustration, Classical
- Scenes: Story, History
- Tags: Poster, Classical, Style
- Cover: `/images/case359.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-poster
- Example cases: case 359

Use when:
- EN: Use for poetic portrait posters, ink atmospheres, and layered cultural visuals.
- ZH: 用于诗意人像海报、水墨氛围和文化主题视觉。

Guidance:
  - Blend portrait silhouette, ink texture, atmosphere, and negative space.
  - Keep composition quiet, premium, and readable.
  - 融合人像剪影、水墨质感、氛围和留白。
  - 保持构图克制、高级、可读。

Pitfalls:
  - Avoid cheap fantasy collage and overloaded scenery.
  - Use subtle text or no text unless required.
  - 避免廉价奇幻拼贴和景物堆叠。
  - 非必要时减少文字。

### Nature Science Poster / 自然科普海报

- ID: `nature-science-poster`
- Category: Posters & Typography
- Styles: Poster, Infographic
- Scenes: Education
- Tags: Poster, Education, Style
- Cover: `/images/case339.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-poster
- Example cases: case 339

Use when:
- EN: Use for natural subjects that need a premium, clean science poster feel.
- ZH: 用于自然主题的高级、干净科普海报。

Guidance:
  - Use a clear subject, minimal copy, soft shadows, and disciplined whitespace.
  - Keep the scientific label short and visible.
  - 使用清晰主体、少量文案、柔和阴影和充足留白。
  - 让科普标签短而清楚。

Pitfalls:
  - Avoid heavy advertising language.
  - Avoid dense encyclopedia blocks.
  - 避免广告感太重。
  - 避免密集百科正文。

### Product Commerce Visual / 商品商业视觉

- ID: `product-commerce-visual`
- Category: Products & E-commerce
- Styles: Product, Realistic
- Scenes: Commerce, Food
- Tags: Product, Commerce, Packaging
- Cover: `/images/case373.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-product
- Example cases: case 373, case 358

Use when:
- EN: Use for product hero shots, packaging visuals, detail pages, and sales layouts.
- ZH: 用于商品主图、包装视觉、详情页和销售卖点排版。

Guidance:
  - Define product, selling points, material, scene, lighting, and layout blocks.
  - Separate hero product, benefit labels, and supporting props.
  - 定义商品、卖点、材质、场景、光线和版块。
  - 区分主商品、卖点标签和辅助道具。

Pitfalls:
  - Avoid random props that weaken product recognition.
  - Constrain packaging text and claim wording.
  - 避免无关道具削弱商品识别。
  - 约束包装文字和卖点表达。

### Personalized Beauty Report / 个性化美妆报告

- ID: `personalized-beauty-report`
- Category: Products & E-commerce
- Styles: Product, UI
- Scenes: Commerce, Fashion
- Tags: Product, Layout, Style
- Cover: `/images/case353.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-product
- Example cases: case 353

Use when:
- EN: Use for beauty recommendations, skin reports, shopping assistants, and lifestyle product cards.
- ZH: 用于美妆推荐、肤质报告、导购助手和生活方式商品卡片。

Guidance:
  - Use a report-like hierarchy with diagnosis, recommendation, and product cards.
  - Keep product images, labels, and ratings aligned.
  - 使用诊断、推荐和商品卡片的报告层级。
  - 对齐商品图、标签和评分。

Pitfalls:
  - Avoid medical claims and unreadable dense notes.
  - Keep recommendation logic simple.
  - 避免医疗化结论和难读小字。
  - 保持推荐逻辑清楚。

### Brand Identity Package / 品牌身份包

- ID: `brand-identity-package`
- Category: Brand & Logos
- Styles: Brand
- Scenes: Commerce
- Tags: Brand, Logo, Identity
- Cover: `/images/case354.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-brand
- Example cases: case 354

Use when:
- EN: Use for logo systems, brand boards, visual identity kits, and application mockups.
- ZH: 用于 Logo 系统、品牌板、VI 套件和应用样机。

Guidance:
  - Define brand name, positioning, palette, typography, logo usage, and touchpoints.
  - Ask for a coherent board with aligned applications.
  - 定义品牌名、定位、配色、字体、Logo 用法和触点。
  - 要求视觉板中的应用统一对齐。

Pitfalls:
  - Avoid unrelated logo variants and inconsistent palettes.
  - Keep brand text accurate.
  - 避免无关 Logo 变体和混乱配色。
  - 保持品牌文字准确。

### Brand Touchpoint Board / 品牌触点视觉板

- ID: `brand-touchpoint-board`
- Category: Brand & Logos
- Styles: Brand, Product
- Scenes: Commerce, Social
- Tags: Brand, Identity, Campaign
- Cover: `/images/case362.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-brand
- Example cases: case 362

Use when:
- EN: Use for multi-touchpoint campaign boards and brand rollout previews.
- ZH: 用于多触点 Campaign 展示和品牌落地预览。

Guidance:
  - Specify touchpoint list, shared visual rules, and mockup arrangement.
  - Use one palette and one typography logic across all panels.
  - 指定触点清单、统一视觉规则和样机排列。
  - 让所有面板共享配色和字体逻辑。

Pitfalls:
  - Avoid mixing many unrelated campaign styles.
  - Limit touchpoints if readability drops.
  - 避免混入多个无关 Campaign 风格。
  - 可读性下降时减少触点数量。

### Architecture & Space / 建筑与空间

- ID: `architecture-space`
- Category: Architecture & Spaces
- Styles: Architecture
- Scenes: Travel, Commerce
- Tags: Architecture, Interior, Map
- Cover: `/images/case331.png`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-architecture
- Example cases: case 331, case 11

Use when:
- EN: Use for interiors, architecture renders, city maps, spatial plans, and environment concepts.
- ZH: 用于室内、建筑表现、城市地图、空间规划和环境概念图。

Guidance:
  - Define viewpoint, scale, material, lighting, and spatial function.
  - For maps, specify landmarks, labels, border decoration, and visual accuracy level.
  - 定义视角、尺度、材质、光线和空间功能。
  - 地图需要指定地标、标签、边框装饰和准确度。

Pitfalls:
  - Avoid impossible perspectives unless the output is conceptual.
  - Lock map label language and relative placement.
  - 概念图之外要避免不合理透视。
  - 锁定地图标签语言和相对位置。

### Realistic Photography / 写实摄影

- ID: `realistic-photography`
- Category: Photography & Realism
- Styles: Photography, Realistic
- Scenes: Fashion, Commerce
- Tags: Photography, Realistic, Lens
- Cover: `/images/case377.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-photo
- Example cases: case 377

Use when:
- EN: Use for portraits, street photos, product photography, and cinematic realism.
- ZH: 用于人像、街拍、商品摄影和电影感写实。

Guidance:
  - Specify camera distance, lens, light source, texture, background, and motion.
  - Use believable imperfections for documentary realism.
  - 指定机位、镜头、光源、质感、背景和动作。
  - 加入可信的小瑕疵增强纪实感。

Pitfalls:
  - Avoid over-polished plastic skin unless commercial beauty is required.
  - Add negative constraints for hands, text, and anatomy when needed.
  - 商业美妆之外，避免过度磨皮。
  - 需要时加入手部、文字、结构类负面约束。

### Street Accident Moment / 街头意外瞬间摄影

- ID: `street-accident-moment`
- Category: Photography & Realism
- Styles: Photography, Realistic
- Scenes: Travel, Social
- Tags: Photography, Realistic, Scene
- Cover: `/images/case376.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-photo
- Example cases: case 376

Use when:
- EN: Use for candid street moments, accidental spills, documentary phone shots, and fast action.
- ZH: 用于街头抓拍、意外泼洒、手机纪实和快速动作。

Guidance:
  - Describe the exact moment, camera height, motion blur, and street context.
  - Add negative constraints for staged poses and fake ad lighting.
  - 描述具体瞬间、机位高度、运动模糊和街景。
  - 加入避免摆拍和广告棚拍感的限制。

Pitfalls:
  - Avoid too-clean compositions.
  - Keep the event plausible and grounded.
  - 避免画面过于干净。
  - 让事件看起来可信。

### Illustration & Art Style / 插画与艺术风格

- ID: `illustration-art-style`
- Category: Illustration & Art
- Styles: Illustration
- Scenes: Story, Creative
- Tags: Illustration, Art, Style
- Cover: `/images/case346.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-illustration
- Example cases: case 346, case 6

Use when:
- EN: Use for anime, watercolor, ink, decorative art, and style experiments.
- ZH: 用于动漫、水彩、水墨、装饰画和风格实验。

Guidance:
  - Define composition, subject, palette, brush material, mood, and rendering depth.
  - For reference images, state what must be preserved.
  - 定义构图、主体、配色、笔触材质、情绪和完成度。
  - 参考图任务需要说明保留哪些特征。

Pitfalls:
  - Avoid style-only prompts without composition.
  - Lock character identity when using references.
  - 避免只写风格，不写构图。
  - 使用参考图时锁定角色识别。

### Character Design Sheet / 角色设定表

- ID: `character-design-sheet`
- Category: Characters & People
- Styles: Character, Illustration
- Scenes: Story
- Tags: Character, Pose, Style
- Cover: `/images/case347.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-character
- Example cases: case 347

Use when:
- EN: Use for character sheets, pose grids, action breakdowns, and identity references.
- ZH: 用于角色设定表、动作网格、动作拆解和一致性参考。

Guidance:
  - Define identity anchors, outfit, proportions, pose count, and sheet layout.
  - Keep face, hairstyle, and costume details consistent.
  - 定义身份锚点、服装、比例、动作数量和版式。
  - 保持脸、发型和服装细节一致。

Pitfalls:
  - Avoid changing costume details between poses.
  - Limit pose count if the sheet becomes crowded.
  - 避免不同动作里服装细节变化。
  - 画面拥挤时减少动作数量。

### 3D Collectible Toy / 3D 收藏玩具

- ID: `3d-collectible-toy`
- Category: Characters & People
- Styles: 3D, Character
- Scenes: Commerce, Creative
- Tags: Character, 3D, Style
- Cover: `/images/case378.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-character
- Example cases: case 378

Use when:
- EN: Use for premium collectible figures, avatar toys, blind-box characters, and 3D display renders.
- ZH: 用于高级收藏玩具、头像公仔、潮玩角色和 3D 展示图。

Guidance:
  - Preserve face and outfit anchors from the reference.
  - Specify material, packaging, base, lighting, and collectible scale.
  - 保留参考图中的脸和服装锚点。
  - 指定材质、包装、底座、光线和收藏比例。

Pitfalls:
  - Avoid generic toy bodies without identity details.
  - Keep packaging text minimal and accurate.
  - 避免没有身份细节的通用玩具。
  - 包装文字保持少量且准确。

### Scene Storytelling / 场景叙事

- ID: `scene-storytelling`
- Category: Scenes & Storytelling
- Styles: Scenes, Illustration
- Scenes: Story, Social
- Tags: Scene, Story, Storyboard
- Cover: `/images/case330.png`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-scene
- Example cases: case 330

Use when:
- EN: Use for storyboards, worldbuilding, live scenes, and emotional narrative frames.
- ZH: 用于分镜、世界观、直播场景和情绪叙事画面。

Guidance:
  - Define who, where, when, conflict, emotion, and camera framing.
  - Use scene details to support narrative rather than decoration.
  - 定义人物、地点、时间、冲突、情绪和机位。
  - 让场景细节服务故事。

Pitfalls:
  - Avoid generic fantasy backgrounds.
  - Keep narrative cues visible in the frame.
  - 避免通用幻想背景。
  - 让故事线索在画面里可见。

### History & Classical Themes / 历史与古风题材

- ID: `history-classical-themes`
- Category: History & Classical Themes
- Styles: History, Classical, Illustration
- Scenes: History, Story
- Tags: History, Classical, Scroll
- Cover: `/images/case375.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-history
- Example cases: case 375, case 338

Use when:
- EN: Use for ancient Chinese themes, scrolls, dynasty clothing, poetry visuals, and historical scenes.
- ZH: 用于古风题材、长卷、朝代服饰、诗词视觉和历史场景。

Guidance:
  - Specify dynasty, clothing system, object references, layout format, and cultural mood.
  - Use scroll, album page, or poster format deliberately.
  - 指定朝代、服饰制度、器物参考、版式和文化气质。
  - 明确长卷、册页或海报形式。

Pitfalls:
  - Avoid mixing dynasties when historical accuracy matters.
  - Constrain random modern props.
  - 需要历史准确时，避免朝代混搭。
  - 约束随机现代物件。

### Document & Publishing / 文档与出版物

- ID: `document-publishing`
- Category: Documents & Publishing
- Styles: Documents, Infographic
- Scenes: Education, Tech
- Tags: Document, Publishing, Layout
- Cover: `/images/case360.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-document
- Example cases: case 360

Use when:
- EN: Use for white papers, manuals, encyclopedic plates, report pages, and publication systems.
- ZH: 用于白皮书、手册、百科图鉴、报告页面和出版系统。

Guidance:
  - Define page size, columns, table of contents, figure system, and typography hierarchy.
  - Use readable headings, tables, labels, and page rhythm.
  - 定义页面尺寸、分栏、目录、图表系统和字体层级。
  - 使用可读标题、表格、标签和页面节奏。

Pitfalls:
  - Avoid tiny dense text.
  - Keep charts and captions aligned to the page grid.
  - 避免密集小字。
  - 让图表和说明对齐页面网格。

### Concept Product Breakdown / 概念产品研发拆解

- ID: `concept-product-breakdown`
- Category: Other Use Cases
- Styles: Other Use Cases, Product
- Scenes: Creative, Tech
- Tags: Creative, R&D, Special
- Cover: `/images/case370.jpg`
- Template source: https://github.com/freestylefly/awesome-gpt-image-2/blob/main/docs/templates.md#tpl-other
- Example cases: case 370, case 361

Use when:
- EN: Use for experimental prompt tasks, R&D boards, exploded diagrams, and unusual visual systems.
- ZH: 用于实验型任务、研发视觉板、拆解图和特殊视觉系统。

Guidance:
  - Define the artifact type, components, labels, material logic, and final presentation format.
  - Use clear callouts and a controlled technical style.
  - 定义产物类型、组件、标签、材质逻辑和展示格式。
  - 使用清晰标注和受控技术风格。

Pitfalls:
  - Avoid unspecified mixed tasks.
  - Keep labels short and component relationships visible.
  - 避免任务边界过泛。
  - 标签要短，组件关系要清楚。

## Categories

- UI & Interfaces: UI 与界面 | Apps, websites, dashboards, social screenshots, and product interfaces.
- Charts & Infographics: 图表与信息可视化 | Infographics, knowledge maps, technical explainers, and structured diagrams.
- Posters & Typography: 海报与排版 | Event posters, covers, type-driven visuals, and strong layout compositions.
- Products & E-commerce: 商品与电商 | Product shots, detail pages, packaging, selling points, and ads.
- Brand & Logos: 品牌与标志 | Logos, identity systems, brand touchpoints, and campaign visuals.
- Architecture & Spaces: 建筑与空间 | Architecture renders, interiors, city maps, and spatial concepts.
- Photography & Realism: 摄影与写实 | Portraits, phone photography, film texture, and commercial photography.
- Illustration & Art: 插画与艺术 | Illustration, art styles, material experiments, and decorative images.
- Characters & People: 人物与角色 | Character design, pose references, cards, and 3D toys.
- Scenes & Storytelling: 场景与叙事 | Storyboards, narrative scenes, livestream frames, and worldbuilding.
- History & Classical Themes: 历史与古风题材 | Classical scrolls, historical figures, traditional themes, and poetry visuals.
- Documents & Publishing: 文档与出版物 | White papers, manuals, encyclopedic plates, and publishing layouts.
- Other Use Cases: 其他应用场景 | Creative experiments, special tasks, mixed workflows, and practical cases.

## Styles

- 3D: 3D | Keywords: 3d, toy, render, 玩具
- Architecture: 建筑 | Keywords: None
- Brand: 品牌 | Keywords: brand, logo, identity, 品牌, 标志
- Character: 角色 | Keywords: character, avatar, pose, 角色, 人物
- Characters: 人物 | Keywords: None
- Charts: 图表 | Keywords: None
- Classical: 古典 | Keywords: classical, dynasty, history, 古风, 历史
- Documents: 文档 | Keywords: None
- History: 历史 | Keywords: None
- Illustration: 插画 | Keywords: illustration, painting, watercolor, 插画, 绘画
- Infographic: 信息图 | Keywords: infographic, diagram, 信息图, 图解
- Other Use Cases: 其他应用场景 | Keywords: None
- Photography: 摄影 | Keywords: None
- Poster: 海报 | Keywords: poster, cover, typography, 海报, 封面
- Product: 商品 | Keywords: product, packaging, 商品, 包装
- Products: 商品 | Keywords: None
- Realistic: 写实 | Keywords: photo, realistic, camera, 写真, 写实
- Scenes: 场景 | Keywords: None
- UI: 界面 | Keywords: ui, interface, dashboard, 界面, 截图

## Scenes

- Creative: 创意 | Keywords: None
- Tech: 科技 | Keywords: ai, rag, tech, data, 技术, 数据
- Commerce: 商业 | Keywords: product, brand, ad, campaign, 商品, 商业, 广告
- Education: 教育 | Keywords: guide, atlas, science, learning, 学习, 科普
- Social: 社媒 | Keywords: social, x , wechat, 朋友圈, 社媒
- Fashion: 时尚 | Keywords: fashion, clothing, portrait, 服饰, 写真
- Food: 食品饮品 | Keywords: food, drink, coffee, tea, 餐厅, 咖啡, 茶
- Travel: 旅行 | Keywords: city, map, street, 城市, 地图, 街头
- Story: 叙事 | Keywords: story, scene, world, 故事, 场景
- History: 历史 | Keywords: history, dynasty, ancient, 历史, 古希腊, 唐

---
name: gpt-image-2-style-library
description: Choose GPT-Image2 / gpt-image-2 visual styles and industrial prompt templates from the awesome-gpt-image-2 style library. Use when an agent needs to create, rewrite, classify, or improve image-generation prompts with repository-backed templates, categories, style tags, scene tags, pitfalls, and example cases.
---

# GPT-Image2 Style Library

Use this skill to turn a user's image-generation intent into a production-ready GPT-Image2 prompt using the awesome-gpt-image-2 style library.

## Example Output

![City life system map example](assets/city-life-system-map.png)

Example request: `用 gpt-image-2-style-library 技能生成城市生命系统图谱`

## Reference

- Read `references/style-library.md` before choosing a template or style.
- The reference is generated from `data/style-library.json` in the repository.
- Prefer the reference over memory when template names, categories, covers, or style tags matter.

## Workflow

1. Detect the user's language and answer in that language.
2. Identify the user's target output: product, poster, UI, infographic, brand, photo, illustration, character, scene, history, document, or special task.
3. Match the request in this order: template category, visual style tag, scene tag, then nearest example cases.
4. If one template is clearly strongest, use it directly. If several are plausible, present 2-3 options with short reasons and ask the user to choose.
5. Build the final prompt with these blocks:
   - subject and task
   - composition and layout
   - visual style and materials
   - text and label requirements
   - aspect ratio and output format
   - constraints and negative details
6. Include the selected template name and any useful example case IDs.

## Output Defaults

- Provide a copyable prompt first.
- Keep constraints concrete: exact text, aspect ratio, readable labels, layout hierarchy, and avoided artifacts.
- For Chinese requests, write the final prompt in Chinese unless the user asks for English.
- For English requests, write the final prompt in English unless the user asks for Chinese.
- When the user asks for multiple concepts, reuse one template and vary subject, composition, palette, and scene.

## Maintenance

When the source repository changes, run:

```bash
npm run generate:style-skill
```

To install the skill into the local Codex skill folder, run:

```bash
npm run install:skill
```