This is a content visualization tool that takes text or URLs and renders them as PNG cards in seven different styles. The default long card adapts to content length, but you can also cast content as infographics, multi-card sets, editorial sketchnotes that follow a problem-to-insight arc, manga-style comics, whiteboard diagrams, or big-font cards for social media. It's built around a Chinese metaphor of casting (铸) content into visual molds. The interesting bit is the taste.md reference file that sets design constraints across all modes, explicitly rejecting AI-generated aesthetics like Inter font, pure black, and centered hero layouts. Files drop into your Downloads folder. If you regularly turn writing into shareable images, especially for Chinese platforms, this handles the templating and screenshot automation.

skills.sh installs
7.1K
GitHub stars
7.7K
last updated
Aug 21, 2026

Install

$ npx -y skills add lijigang/ljg-skills --skill ljg-card --agent claude-code

Installs into .claude/skills of the current project.

Files

ljg-card:铸

内容进去,PNG 出来。生成图负责把思想变成可见动作,HTML 负责把话说准确;-f 只设计原文的呈现,不改动原文,也不生成新图。

-l、-w 承担知识讲解时,还要接通「具体情形 → 判别依据 → 有条件的关系 → 具体结果」。读者应能从正文里的动作或状态看出对象凭什么属于某个概念,再用关系推出眼前的结果;只把例子放在术语旁边,连接仍然缺失。先确定目标读者要完成什么判断,再决定材料和画面;下上结构不规定版面上下,也不要求每张卡从故事开头。文学、叙事与审美表达按自身目的组织,不强加习题。

成品在内容自身的结论、问题或余韵处结束,之后只留署名与来源。不要追加「阅读边界」、免责声明、制作说明或验收说明,也不要换个标题保留这类尾注。会改变理解的原文条件、分歧与限制,就近融入相应论证;提炼范围、材料缺口和验证状态放在交付回复中。-f 原稿已有的段落仍按全文保真合同保留。

参数

参数模具尺寸图像角色
-l(默认)长图 · 静线叙事1080 × auto0–3 幅;有理解增益时用 1 个安静主场景,必要时追加连续视觉拍点
-f全文 · 原文排版1080 × auto生成图固定为 0;只保真呈现原稿自带图片
-c漫画1080 × auto缺口驱动、同案重跑的漫画分镜
-w白板 · 纵向论证1080 × auto嵌入承重推理节点的局部手绘动作;不设顶部主视觉

未给参数时使用 -l。

必读顺序

-l、-c、-w 每次执行都必须依次 Read:

  1. references/taste.md
  2. references/image-generation.md
  3. 当前 mode 文件
  4. 当前 HTML 模板

不得跳过共享图像协议直接写提示词,也不得把一种 mode 的图像语法套给另一种。

-l、-w 用于知识讲解时,在第 3 步后、读取模板前补读 references/learning-design.md:其中包含读者任务、下上连接、来源边界、换案检验,以及四组失败与修复对照。-f 不进入这条路线。

-f 依次 Read:

  1. references/taste.md
  2. references/mode-full.md
  3. assets/full_template.html

-f 不生成新图;只有原稿自带图片时才补读 references/image-generation.md 的资产身份、来源保真与本地落盘合同。

-f 字体优先级

Sponsors

Details

Repository
github.com/lijigang/ljg-skills
Categories
Frontend Development, Code Review & Quality
First seen
Apr 16, 2026

More from lijigang/ljg-skills

All 25
1starchild-ai-agent avatarorderly-onboardingstarchild-ai-agent/official-skillsOrderly Network onboarding: omnichain perps infra, MCP server, SDK and CLI quickstart. Also indexes executed Orderly fills into Starchild trade analytics. Use w7.1Kskills.sh installs
2openai avatarfigma-implement-designopenai/skillsTranslates Figma designs into production-ready application code with 1:1 visual fidelity. Use when implementing UI code from Figma files, when user mentions "im7Kskills.sh installs
3anthropics avatararchitectureanthropics/knowledge-work-pluginsCreate or evaluate an architecture decision record (ADR). Use when choosing between technologies (e.g., Kafka vs SQS), documenting a design decision with trade-7Kskills.sh installs
4pixijs avatarpixijs-mathpixijs/pixijs-skillsUse this skill when working with coordinates, vectors, matrices, shapes, hit testing, or layout rectangles in PixiJS v8. Covers Point/ObservablePoint, Matrix (26.9Kskills.sh installs
5wshobson avatarblock-no-verify-hookwshobson/agentsConfigure a PreToolUse hook to prevent AI agents from skipping git pre-commit hooks with --no-verify and other bypass flags. Use when setting up Claude Code pro6.9Kskills.sh installs
6clerk avatarclerk-expo-patternsclerk/skillsExpo / React Native patterns with Clerk — SecureStore token cache, OAuth deep linking, useAuth in native, Expo Router protected routes, push notifications with 6.8Kskills.sh installs