ai-teaching-media

AI Teaching Media Integrated Skill

Consolidate "content understanding -> illustration/infographic -> animated GIF/video -> cover/final video" into one interlinked skill pack. Users can install this directory to use all 6 sub-skills and chain them into a complete teaching production pipeline according to the scenario.

Directory Structure

ai-teaching-media/
├── SKILL.md                      # 总入口与路由
├── README.md                     # 使用说明
├── ai-image-generator/           # ① 通用生图执行层
├── tech-article-diagram/         # ② 技术长文智能插图
├── edu-subject-infographic/      # ③ 学科知识点竖版信息图
├── edu-teaching-animation/       # ④ 教学动图 + 配音教学视频
├── short-video-cover/            # ⑤ 短视频 3:4 封面
└── article-explainer-video/      # ⑥ 长文转章节解说视频

Overview of the 6 sub-skills

Sub-skill	Function	Typical Input	Output
ai-image-generator	Image-generation execution layer (called by other sub-skills)	prompt / reference image	PNG
tech-article-diagram	Intelligent illustration for long technical articles (7 styles)	Markdown long article	16:9 illustration
edu-subject-infographic	Subject knowledge point infographic	Knowledge point name	9:16 infographic
edu-teaching-animation	Teaching animated GIF + voiceover video	Subject concept	Silent MP4 / voiceover MP4
short-video-cover	Vertical short video cover	Spoken script	3:4 cover image
article-explainer-video	Long-article chapter explainer video	Technical article	1080p MP4

How to choose a capability (routing)

After receiving a user request, first determine the intent, then enter the corresponding sub-skill's SKILL.md：

1. Image generation only / batch image generation / image editing with reference image
   -> Read ai-image-generator/SKILL.md
2. Add illustrations, flowcharts, architecture diagrams, comparison diagrams to long technical articles
   -> Read tech-article-diagram/SKILL.md
   -> When generating images, call ai-image-generator
3. Create a knowledge point infographic / knowledge card / subject diagram
   -> Read edu-subject-infographic/SKILL.md
   -> When generating images, call ai-image-generator
4. Teaching animated GIF, silent loop, voiceover teaching video
   -> Read edu-teaching-animation/SKILL.md
5. WeChat Channels/Douyin/Xiaohongshu 3:4 cover
   -> Read short-video-cover/SKILL.md
   -> When generating images, call ai-image-generator
6. Turn the entire article into a chapter explainer video
   -> Read article-explainer-video/SKILL.md
   -> Reuse tech-article-diagram + ai-image-generator

If the user's goal is a "full set of teaching output", do not produce only a single-point result; chain them according to the pipeline below.

Recommended chained pipeline

Pipeline A: Knowledge point -> full set of teaching assets

Applicable: the user provides a knowledge point (such as "Pythagorean theorem" or "photosynthesis")

1) edu-subject-infographic
   产出 9:16 学科信息图（可先确认色系/年级）
2) edu-teaching-animation
   同主题产出：无声循环动图 + 配音教学视频
3) short-video-cover
   用口播文案/标题生成 3:4 短视频封面

Delivery suggestions:

• Infographic PNG
• Silent animated GIF MP4 (embedded in WeChat official accounts)
• Voiceover teaching video MP4
• Short video cover PNG

Pipeline B: Long technical article -> illustrations + explainer video

Applicable: the user provides a long technical article / tutorial Markdown

1) tech-article-diagram
   识别 4-5 个插图位，生成 16:9 概念/流程/对比/架构图
2) article-explainer-video
   文章分镜 → 章节插图 → TTS → HyperFrames 成片
3) short-video-cover（可选）
   为成片做竖版封面

Delivery suggestions:

• Illustration list + PNG
• Chapter explainer video MP4
• Optional short video cover

Pipeline C: Spoken/short video publishing package

Applicable: the user already has a spoken script or teaching video script

1) short-video-cover
   生成 3:4 封面
2) 若还缺讲解视频：
   - 单概念 → edu-teaching-animation
   - 长文讲解 → article-explainer-video

Execution principles

1. Route first, then execute: always read the corresponding sub-skill's SKILL.md, do not skip sub-skill constraints.
2. Keep the same theme consistent: for the same knowledge point/article, infographics, animated images, and videos should reuse the same set of theme colors and terminology as much as possible.
3. Underlying image generation uniformly goes through ai-image-generator: other skills handle planning and prompts, and do not reimplement the API.
4. Cross-skill script paths: sub-skills call each other via relative paths, for example:
   • ../ai-image-generator/scripts/generate.py
   • ../tech-article-diagram/scripts/inject_style.py
5. Explain first when dependencies are missing: when the image API key / Minimax / Node+ffmpeg is not ready, inform the user first; do not fail silently.

Environment dependencies

Capability	Dependency
Image generation type (①②③⑤, and the illustrations for ⑥)	MULERUN_API_KEY or APIMART_API_KEY or ATLASCLOUD_API_KEY or AGNES_API_KEY
Voiceover video (④⑥)	MINIMAX_API_KEY(optional) or free edge-tts（pip install edge-tts）
Video rendering (④⑥)	Node.js ≥ 22、ffmpeg、HyperFrames（npx obtain)

Quick trigger examples

# 单点
帮这篇文章加几张插图，用 notebook 风格
做一张“声现象”学科信息图
把勾股定理做成教学动图和配音视频
给这段口播做个 3:4 短视频封面
把这篇文章转成章节解说视频

# 串联
围绕“光合作用”产出全套教学资产：信息图 + 动图 + 配音视频 + 封面
把这篇 Claude Code 教程做成插图版文章，并再出一版解说视频和封面

Sub-skill entry points

• ai-image-generator/SKILL.md
• tech-article-diagram/SKILL.md
• edu-subject-infographic/SKILL.md
• edu-teaching-animation/SKILL.md
• short-video-cover/SKILL.md
• article-explainer-video/SKILL.md

When starting a task: first choose the above route or pipeline based on the user's goal, then enter the corresponding sub-skill and execute strictly.