VideoDB Video and Audio Workflows
video-db/skills
Ingests video, audio, live streams and desktop captures, indexes them for timestamped search, then edits, subtitles, dubs and clips the results.
A skill your agent uses when asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a math problem (数学题, geometry, algebra, functions, motion/行程 problems), from a problem screenshot…
Install Edu Math Video
$ npx skills add wy51ai/edulab --skill edu-math-video -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install wy51ai/edulab edu-math-video --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/wy51ai/edulab.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/edu-math-video .claude/skills/edu-math-video && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "edu-math-video" agent skill from https://github.com/wy51ai/edulab/tree/master/skills/edu-math-video into .claude/skills/edu-math-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "edu-math-video", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/wy51ai/edulab/tree/master/skills/edu-math-videoType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add wy51ai/edulab --skill edu-math-video -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install wy51ai/edulab edu-math-video --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/wy51ai/edulab.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/edu-math-video .agents/skills/edu-math-video && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "edu-math-video" agent skill from https://github.com/wy51ai/edulab/tree/master/skills/edu-math-video into .agents/skills/edu-math-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "edu-math-video", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add wy51ai/edulab --skill edu-math-video -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install wy51ai/edulab edu-math-video --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/wy51ai/edulab.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/edu-math-video .cursor/skills/edu-math-video && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "edu-math-video" agent skill from https://github.com/wy51ai/edulab/tree/master/skills/edu-math-video into .cursor/skills/edu-math-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "edu-math-video", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/wy51ai/edulab.git --path skills/edu-math-video--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add wy51ai/edulab --skill edu-math-video -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install wy51ai/edulab edu-math-video --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/wy51ai/edulab.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/edu-math-video .gemini/skills/edu-math-video && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "edu-math-video" agent skill from https://github.com/wy51ai/edulab/tree/master/skills/edu-math-video into .gemini/skills/edu-math-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "edu-math-video", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install wy51ai/edulab edu-math-videoInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add wy51ai/edulab --skill edu-math-video -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/wy51ai/edulab.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/edu-math-video .github/skills/edu-math-video && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "edu-math-video" agent skill from https://github.com/wy51ai/edulab/tree/master/skills/edu-math-video into .github/skills/edu-math-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "edu-math-video", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add wy51ai/edulab --skill edu-math-video -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install wy51ai/edulab edu-math-video --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/wy51ai/edulab.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/edu-math-video .opencode/skills/edu-math-video && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "edu-math-video" agent skill from https://github.com/wy51ai/edulab/tree/master/skills/edu-math-video into .opencode/skills/edu-math-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "edu-math-video", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
edu-math-video- GitHub stars
- 1.4k
- Token cost
- ~2.5k tokens
- SKILL.md length
- 640 words
- Files
- 34 (incl. scripts)
- Skills in repo
- 6
- Repo updated
- First seen
- Licence
- Apache-2.0
At a glance
A skill your agent uses when asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a math problem (数学题, geometry, algebra, functions, motion/行程 problems), from a problem screenshot…
- Works in 11 steps: 严格按下面 10 步顺序做，每步的"通过标准"满足了才能进入下一步。… → 先讲清楚题目。 视频第一幕必须展示原题图片（截图或… → tts 字段只能是能念出来的中文。 不能有阿拉伯数字、= + − × ÷ / ^… → …
- Asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a math problem (数学题
- SKILL.md covers 必须遵守的规则, 11 个步骤, 常见错误（真实发生过） and 参考文件
- Runs Python, Shell and JavaScript scripts from its folder; calls node, bash and npm; needs GLM_API_KEY
What it does
Edu Math Video is an agent skill from wy51ai/edulab. Use when asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a math problem (数学题, geometry, algebra, functions, motion/行程 problems), from a problem screenshot or text, with Chinese voice-over (GLM-TTS / 智谱 TTS), bilingual zh+en subtitles and hand-drawn canvas animation rendered to MP4.
Its SKILL.md is about 2.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 37 other files, including scripts (for example `examples/cone-parallel/anim.js`, `examples/cone-parallel/episode.json` and `examples/cone-parallel/script.json`).
It sits in Media & Creative, covering Text to speech and voice, Transcription and Translation. It works with Python. The licence is Apache-2.0.
When your agent uses it
- Asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a math problem (数学题
- Motion/行程 problems)
- From a problem screenshot
- With Chinese voice-over (GLM-TTS / 智谱 TTS)
Example prompts
- “/edu-math-video”
Requirements
- Python 3
- Node.js
- A Bash shell
- A credential in GLM_API_KEY
Workflow steps
11 steps, taken from the first numbered list in SKILL.md.
- 严格按下面 10 步顺序做，每步的"通过标准"满足了才能进入下一步。 不要跳步，不要"最后一起检查"。
- 先讲清楚题目。 视频第一幕必须展示原题图片（截图或 make_problem_png.py 生成的图），旁白逐条读条件，每读一条就在图上框出那一条。
- tts 字段只能是能念出来的中文。 不能有阿拉伯数字、= + − × ÷ / ^ √ ∠ △ ∥ ° ( ) 等符号：AC=3 写成 AC等于三，x² 写成 x的平方。字幕 zh 字段保留正常数学写法。--check 会报错拦住。
- 读音必须固定。 调 TTS 之前 --check 必须显示 未固定读音的生僻字/多音字: 0 且 script errors: 0。多音字的标注写在该字后面：长[cháng]、要[yào]求。不能写成 中点[zhōng]（那会把"点"读成 zhōng）。
- API Key 只能由用户提供。 缺 GLM_API_KEY 时先问用户：配置 key（音质最好，推荐），还是用兜底引擎（edge-tts / macOS say，见 reference/glm-tts-setup.md）。不要编造 key、模型名或接口地址，不要把 key…
- 所有画面内容在 y ≤ 860 以内（y≈914~1044 是字幕框）。图形放左半边 x 60~900，推导文字写在右边的横线板 board() 上。
- 先写分镜，再写动画。 storyboard.md 给每句旁白规划"指/动/留"，按 reference/visual-design.md 的动作表选动作；node render.mjs motion 必须通过。
- 先预览再花钱。 用 --preview + stills auto 检查版面，没问题再跑真正的 TTS。
- 看图验证。 每次渲染截图后都要打开 build/sheet.png 亲眼检查（用读图工具），不能只看命令有没有报错。汇报时写你实际看到了什么，不要凭印象全部打勾。
- 不要修改 engine.js，除非用户要求换风格；需要新图形就在 anim.js 里写函数。
- 不要把 node_modules、package*.json 复制进视频文件夹，依赖在 WS 根目录装一次，所有视频共享。
What it can do on your machine
Read from SKILL.md and the folder at commit 6a9c876. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Ships 5 files in scripts/ (Python, Shell and JavaScript, from the files we listed), which the agent can run.
Shell commands in SKILL.md call:
nodebashnpm
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md. Its commands use npm, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names these keys or tokens, usually read from environment variables:
GLM_API_KEY
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Edu Math Video loads about 2.5k tokens when it runs. Until then it costs about 80 tokens; SKILL.md has 640 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check: notes
The automated check noted patterns worth knowing about, such as sudo or a known installer.
- NoteMentions a .env fileSKILL.md:27
p4/srt 也输出到这里。**不要因为别的目录（包括本技能所在的项目）已经有 `.env`、`node_modules`、`pron.json` 就把 WS 换过去**，缺什么由 `new_video.sh` 和第 2 步补齐。`PROJ - NoteMentions a .env fileSKILL.md:59
- 默认音色是 `chuichui`（锤锤），用户可以在 `.env` 里用 `GLM_VOICE` 换。 - NoteMentions a .env fileSKILL.md:60
用户**配置（推荐放 `~/.config/math-problem-video/.env`，所有目录通用），然后继续。**不要去别的项目里找 `.env` 用，也不要因此换工作目录。** 用户没有 key 或不想配：`TTS_ENGINE - NoteMentions a .env fileSKILL.md:139
| 当前目录没有 `.env`/`node_modules`，就把项目建到有这些东西的别的目录（例如技能所在的项目） | WS 永远是用户当前目录；缺依赖由 new_video.sh 链接，缺 key 请用户配置 `~/.config/ma - NoteMentions a .env fileSKILL.md:159
glm-tts-setup.md)：智谱 GLM-TTS 注册、API Key、`.env`、音色、语速、报错处理。**引导用户配置时读这个。**
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
SKILL.md
The full file from wy51ai/edulab at commit 6a9c876, republished under its Apache-2.0 licence (© wy51ai). 640 words, ~2,482 tokens.
Download SKILL.mdSave it as
.claude/skills/edu-math-video/SKILL.md (or your agent's skills folder). This skill also uses 33 other files; get the full folder from GitHub.- name
- edu-math-video
- description
- Use when asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a math problem (数学题, geometry, algebra, functions, motion/行程 problems), from a problem screenshot or text, with Chinese voice-over (GLM-TTS / 智谱 TTS), bilingual zh+en subtitles and hand-drawn canvas animation rendered to MP4.
数学题讲解视频 (edu-math-video)
把一道数学题做成 16:9、1920×1080、带中文配音 + 中英双语字幕 + 手绘笔记本风格动画的 MP4。
流水线（每个视频一个文件夹）：
script.json ──build_audio.py──► build/timeline.json + build/mix.wav + ../<name>.srt
(旁白) (GLM-TTS + 合成音乐) │
anim.js + engine.js ◄─────────────────────────┘ (按旁白时间点驱动动画)
└──node render.mjs video──► ../<name>.mp4核心原则：画面跟着旁白走，而且图会"讲题"。 每句旁白，图上都有一个动作把这句推理演出来（线段滑过去重合、三角形叠上去、相机转到俯视、动点移动……），不是只多出一条线、一个标签。每个动画都用"第 k 句旁白开始的时刻"S.at(k, f) 定时，绝不写死秒数。
本技能自带全部代码，不要从别的项目复制文件：
template/：可直接运行的完整示例项目（直角三角形斜边中线题）。engine.js是通用引擎（不改），anim.js/script.json/episode.json/problem.png是每道题都要重写的。shared/pron.py+shared/pron.json：读音控制（GLM-TTS 没有 SSML/拼音输入）。examples/cone-parallel/：立体几何完整示例（anim.js + storyboard.md + script.json）：相机斜视↔俯视、侧面展开成扇形、内错角旋转、平面平移。做立体几何、圆、多问题时先读它。scripts/：new_video.sh（建项目）、setup_check.sh（查环境）、make_problem_png.py（文字题→图片+高亮框）、problem_boxes.py（截图→高亮框坐标）、contact_sheet.py（截图拼图审查）。
下文 SKILL = 本文件所在目录，WS = 用户当前所在的目录（你的工作目录，$PWD）：视频文件夹建在这里，mp4/srt 也输出到这里。不要因为别的目录（包括本技能所在的项目）已经有 .env、node_modules、pron.json 就把 WS 换过去，缺什么由 new_video.sh 和第 2 步补齐。PROJ = 本视频文件夹，PY = setup_check.sh 报告的 python（macOS 上一般是 /usr/bin/python3；PATH 里的 python3 可能没有 numpy）。
必须遵守的规则
退出码不是 0 就是没完成。 --check 退出码 1（哪怕只剩一个 ⚠ 多音字）、motion 退出码 1、截图里有问题，都不能写"只是警告、不影响"然后继续。修好再往下走。
- 严格按下面 10 步顺序做，每步的"通过标准"满足了才能进入下一步。 不要跳步，不要"最后一起检查"。
- 先讲清楚题目。 视频第一幕必须展示原题图片（截图或
make_problem_png.py生成的图），旁白逐条读条件，每读一条就在图上框出那一条。 tts字段只能是能念出来的中文。 不能有阿拉伯数字、= + − × ÷ / ^ √ ∠ △ ∥ ° ( )等符号：AC=3写成AC等于三，x²写成x的平方。字幕zh字段保留正常数学写法。--check会报错拦住。- 读音必须固定。 调 TTS 之前
--check必须显示未固定读音的生僻字/多音字: 0且script errors: 0。多音字的标注写在该字后面：长[cháng]、要[yào]求。不能写成中点[zhōng]（那会把"点"读成 zhōng）。 - API Key 只能由用户提供。 缺
GLM_API_KEY时先问用户：配置 key（音质最好，推荐），还是用兜底引擎（edge-tts / macOSsay，见 reference/glm-tts-setup.md）。不要编造 key、模型名或接口地址，不要把 key 打印出来或写进项目文件夹。 - 所有画面内容在 y ≤ 860 以内（y≈914
1044 是字幕框）。图形放左半边 x 60900，推导文字写在右边的横线板board()上。 - 先写分镜，再写动画。
storyboard.md给每句旁白规划"指/动/留"，按 reference/visual-design.md 的动作表选动作；node render.mjs motion必须通过。 - 先预览再花钱。 用
--preview+stills auto检查版面，没问题再跑真正的 TTS。 - 看图验证。 每次渲染截图后都要打开
build/sheet.png亲眼检查（用读图工具），不能只看命令有没有报错。汇报时写你实际看到了什么，不要凭印象全部打勾。 - 不要修改
engine.js，除非用户要求换风格；需要新图形就在anim.js里写函数。 - 不要把
node_modules、package*.json复制进视频文件夹，依赖在WS根目录装一次，所有视频共享。
11 个步骤
第 1 步：弄清题目，自己先把题解对
- 拿到题目文字（有截图就先读图）。先自己完整解一遍并验算，写下：已知条件清单、所求、关键思路（1~2 个）、每一步计算、最终答案、验算方法。
- 答案不确定或题目看不清时，问用户，不要猜。
- 通过标准： 你能用 3~6 个"幕"讲完，每一幕一句话说清目的；答案已验算。
第 2 步：建项目、查环境
bash SKILL/scripts/new_video.sh "$PWD" <ascii_folder_name> # WS 就是当前目录
bash SKILL/scripts/setup_check.sh WS/<ascii_folder_name>new_video.sh会把pron.py/pron.json链接到技能自带的共享词表，并在找得到时把node_modules链接到已有安装（不复制）；找不到才需要在WS下npm install。python 包用PY -m pip install --user numpy requests pypinyin pillow。new_video.sh报ERROR: ... is where the skill is installed说明你把 WS 设成了技能所在的项目，改用当前目录。- 默认音色是
chuichui（锤锤），用户可以在.env里用GLM_VOICE换。 - 缺
GLM_API_KEY：停下，按 reference/glm-tts-setup.md 引导用户配置（推荐放~/.config/math-problem-video/.env，所有目录通用），然后继续。不要去别的项目里找.env用，也不要因此换工作目录。 用户没有 key 或不想配：TTS_ENGINE=auto（默认）会自动改用兜底引擎，setup_check.sh会显示用的是哪个；装 edge-tts 前先征得用户同意。 - 通过标准：
setup_check.sh输出ALL OK。
第 3 步：测试 TTS（验证 key 和音色）
cd PROJ && PY build_audio.py --say "你好，我们来看一道数学题。"- 成功会打印
WROTE build/say_xxxx.wav和voice: ...（edge:/say:开头 = 兜底引擎）。401/403 = key 错或没余额，告诉用户，不要重试 10 次。 - 第一次给这个用户做视频时，把这个 wav 路径告诉用户试听，确认音色（音色表见 glm-tts-setup.md）。
- 通过标准： 生成了 wav。
第 4 步：准备题目图片 problem.png 和高亮框
二选一（细节见 reference/animation.md）：
- 只有文字：
PY SKILL/scripts/make_problem_png.py --text "完整题目" --mark "条件1" "条件2" ... "所求" --out PROJ/problem.png，它直接输出srcW/srcH/boxes。 - 有截图：复制到
PROJ/，必要时problem_boxes.py crop裁掉无关部分并存为problem.png；然后problem_boxes.py grid出坐标网格图，读出每个条件的像素框；再problem_boxes.py check画框确认。 - 把尺寸和框填进
anim.js顶部的PROBLEM。 - 通过标准： 你打开了
*.preview.png或*.check.png，文字完整（没有 ☒/□ 方框），每个框都恰好罩住对应文字。
第 5 步：写 episode.json 和 script.json
按 reference/script-writing.md 写。要点：
episode.json：title（视频标题）、output_name（输出文件名，如斜边中线_Median_to_Hypotenuse）、pop_scenes（要加"啵"音效的幕）。script.json：幕的数组，每幕{"scene": "英文id", "lines": [{"zh","tts","en"}, ...]}。第一幕intro读题，最后一幕outro回顾 + 报答案。- 每句旁白 ≤ 36 个汉字宽（否则字幕折两行），一句只讲一件事。
- 通过标准： JSON 合法；每一幕的 id 都有计划好的画面。
第 6 步：写分镜 storyboard.md
先读 reference/visual-design.md，再按 template/storyboard.md 的格式给每句旁白写一行：| 幕id 句号 | 旁白要点 | 指 | 动 | 留 | 板书 |。
- "指"：这句提到的元素怎么点亮。"动"：图上一个体现推理的动作，从动作表里选（相等→复制滑过去重合，全等→三角形叠上去，立体↔平面→相机转动，求长度→数字计数……）。"留"：动作后留下的标记。
- 除第一幕和最后一幕外，"动"不能空，也不能只写"出现/显示"。每幕至少一个连续运动。
- 先想清楚：只看图、不看板子，观众能不能看懂这一步为什么成立？不能就换动作。
- 先在 storyboard.md 顶部写图形清单：题目里所有的点、线段、平行/垂直/中点关系、已知长度，全部要画出来，位置按数据计算。
- 通过标准： 每句都有一行；图形清单完整；你能说出每一幕里"最能让人看懂"的那个动作。
第 7 步：读音与脚本检查，修到 0
cd PROJ && PY build_audio.py --check ; cat build/pron_report.txtERROR行：按提示改script.json（通常是 tts 里有数字/符号，或 anim.js 缺SC.<id>；第 8 步写完 anim.js 前后者会报错，属正常）。⚠ 多音/生僻行：按 reference/pronunciation.md 处理：该字后面加[拼音]，或反复出现的词加进pron.json的words，确认默认读音正确的词加进ok。- 通过标准：
未固定读音的生僻字/多音字: 0且script errors: 0，退出码 0。
Show full SKILL.md (261 more words)Show less
第 8 步：写 anim.js
按 storyboard.md 逐句实现，API 见 reference/animation.md。先通读 template/anim.js（示例）再动手，结构保持一致：
PROBLEM（第 4 步的数据）、TAGS（每幕左上角标签）、图形函数（如fig(st)）、SC.<id> = (lt, S) => {...}（每幕一个）。- 每个元素的出现时间 = 旁白提到它的时刻：
S.P(S.at(k, f), 时长)。 - 每句：
glow(..., bump(lt, at(k)))点亮提到的元素 → 分镜里的动作（slideSeg、movePoly、camTween、动点、countTo…）→ 留下标记。 - 右侧
board()+bl(行号, 文字, S.at(k, f), lt)逐行写推导（与图同色），最后stamp()盖章给答案。 - 通过标准：
PY build_audio.py --check无 ERROR。
第 9 步：免费预览：动作检查 + 版面检查
cd PROJ && PY build_audio.py --preview && node render.mjs motion && node render.mjs stills auto && PY SKILL/scripts/contact_sheet.py .motion在每句里取 6 帧，比较左半边图形区的变化：STATIC（没变化）必须修；每幕至少一句MOVE。不通过就回到分镜换更好的动作，不要只加一个无意义的晃动来凑数。stills auto在每句旁白快结束时截一帧（此时这句该出现的东西都应在画面上）。命令输出FAILED: page error说明 anim.js 有 JS 错误，先修。- 打开
build/sheet.png（多于 12 张时是sheet_1.png、sheet_2.png…）逐张检查，对照 reference/animation.md 的审查清单：没有重叠、没有出界、没有挡住字幕、每句提到的东西都画出来了、数值与解答一致。 - 有问题 → 改 anim.js → 重跑本步。
- 通过标准：
motion check passed；全部截图检查通过（在回复里逐幕说明你看到了什么）。
第 10 步：真正生成配音，自动听写核对，再检查一次画面
cd PROJ && PY build_audio.py && PY build_audio.py --asr && node render.mjs stills auto && PY SKILL/scripts/contact_sheet.py .- TTS 按文本缓存在
build/tts/，改了某句只会重新合成那一句。 --asr用智谱语音识别把每句配音转回文字，写入build/asr_report.txt（结果有缓存）。字母序列对不上的句子标 ✗，退出码 1：修改那句（见 pronunciation.md 的"字母"一节）后重跑这两条命令。报告里的汉字也要扫一遍，发现明显读错的字（不是同音字，也不是 ASR 把"三"写成"3"这种）就按第 7 步处理。- 真实时长与预览不同，重新看一遍 sheet。
- 通过标准： 打印
mix + srt written；--asr显示字母读错的句子: 0（没有 GLM key 时显示ASR SKIPPED：交付时告诉用户字母读音没有机器核对，列出含点名的句子请用户重点听）；截图检查通过。
第 11 步：渲染视频并交付
cd PROJ && node render.mjs video 6 # 6 = 并行浏览器页数（不是帧率！帧率固定 30）- 输出
WS/<output_name>.mp4和WS/<output_name>.srt（WS = 用户的当前目录）。2~3 分钟的视频约需几分钟。 - 交付时告诉用户：mp4/srt 路径、时长、各幕内容；说明你无法听音频，请用户试听读音，特别列出你标注过读音的字。用户反馈读错时：加到
pron.json或行内标注 → 再跑第 10、11 步（只会重合成改过的句子）。
常见错误（真实发生过）
| 错误做法 | 正确做法 |
|---|---|
当前目录没有 .env/node_modules，就把项目建到有这些东西的别的目录（例如技能所在的项目） | WS 永远是用户当前目录；缺依赖由 new_video.sh 链接，缺 key 请用户配置 ~/.config/math-problem-video/.env |
| 从别的视频文件夹（化学 chem 等）复制代码 | 用 new_video.sh 从本技能 template/ 建项目 |
编造 GLM_MODEL=glm-4-voice、api.zhipuai.cn 等配置 | 模型只有 glm-tts，地址已写在 build_audio.py；只需用户提供 GLM_API_KEY，可选 GLM_VOICE |
node render.mjs video 24 以为是 24fps | 参数是并行数；预览版面用 stills auto，不要反复渲染整段视频 |
中点[zhōng] | 标注只作用于紧挨着的前一个字：中[zhōng]点（其实"中点"不需要标） |
tts 里写 AC=3、60 − 7t、x² | AC等于三、六十减七t、x的平方 |
| 题目截图随便放在 (50,50) 400×400 | 用 PROBLEM + problemCard() 自动适配到标题下方的卡片里 |
用绝对秒数定时 if (lt > 12.5) | S.at(k, f)：第 k 句开始后 f 比例处 |
| 渲染完不看画面就交付 | 每次都看 build/sheet.png |
tts 里手动写 C、O 或 C,O 来分开字母 | 不用管：pron.py 自动把相邻字母拆成 C O（实测逗号、顿号会多出停顿） |
--check 还有 ⚠ 就调 TTS | 修到 0；否则会花钱合成出错误读音 |
为了"保险"把常见词都加进 pron.json 或行内标注 | 只标会读错的字；替换成同音字会破坏断句（直搅三搅形），pron.py 已经只在上下文会读错时才替换 |
| 一个分句二十多个字不加标点 | 在换气处加逗号，分句 ≤ 18 字（--check 会 warn），否则 TTS 自己在词中间停 |
题目图片里 x² 显示成 x☒ 却没发现 | 审查时放大看文字；make_problem_png.py 已自动换字体 |
用编辑工具改 JSON 后出现弯引号 “，解析失败 | JSON 语法引号必须是英文 "；见 troubleshooting.md |
| 图一次画好，之后每句只多一条线/一个标签，解释全在板子上 | 每句都有"指→动→留"；相等就滑过去重合，立体↔俯视用 camTween；render.mjs motion 必须通过 |
| 需要俯视图时在旁边另画一张图 | 同一个 3D 模型 camTween 转到 pitch = π/2 |
| 一句旁白塞三四个知识点 | 一句一个点，画面同步出现一个元素 |
参考文件
- reference/glm-tts-setup.md：智谱 GLM-TTS 注册、API Key、
.env、音色、语速、报错处理。引导用户配置时读这个。 - reference/script-writing.md：
script.json格式、幕的设计、数学式子的口语写法对照表。 - reference/pronunciation.md：读音控制原理、标注语法、
pron.json、数学常见多音字。 - reference/visual-design.md：讲解动画怎么设计：指→动→留→连、推理→动画动作表、立体转俯视、圆锥展开、分镜格式。写分镜前必读。
- reference/animation.md：engine.js API、场景写法、版面坐标、几何/函数/立体/行程题画法、审查清单。
- reference/troubleshooting.md：报错与处理。
© wy51ai, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 33 other files (scripts) in skills/edu-math-video of wy51ai/edulab.
- SKILL.md
- examples/cone-parallel/anim.js
- examples/cone-parallel/episode.json
- examples/cone-parallel/problem.png
- examples/cone-parallel/script.json
- examples/cone-parallel/storyboard.md
- reference/animation.md
- reference/glm-tts-setup.md
- reference/pronunciation.md
- reference/script-writing.md
- reference/troubleshooting.md
- reference/visual-design.md
- scripts/contact_sheet.py
- scripts/make_problem_png.py
- scripts/new_video.sh
- scripts/problem_boxes.py
- scripts/setup_check.sh
- … and 17 more
Open the folder on GitHubat commit 6a9c876
Compare with similar skills
Edu Math Video next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Edu Math Video this skillwy51ai/edulab | 1.4k | — | ~2.5k | Automated safety check: Notes | Apache-2.0 | |
| VideoDB Video and Audio Workflowsvideo-db/skills | 127 | — | ~5.5k | Automated safety check: Notes | MIT | |
| Proofreadsonilo-ai/skills | 115 | — | ~4.4k | Automated safety check: Notes | MIT | |
| CLI Anything VideocaptionerHKUDS/CLI-Anything | 52k | — | ~1.2k | Automated safety check: Pass | Apache-2.0 | |
| Video Translatorshang-zhu/violin | 1.1k | — | ~1k | Automated safety check: Notes | MIT | |
| Speech Engineelevenlabs/skills | 482 | — | ~2.5k | Automated safety check: Warn | MIT |
Similar skills
Proofread
sonilo-ai/skills
Transcribe a video with Sonilo and translate the transcript into editable .srt files — one per target language, plus the detected source language — so the wording can be read and corrected before…
CLI Anything Videocaptioner
HKUDS/CLI-Anything
AI-powered video captioning — transcribe speech, optimize/translate subtitles, and burn them into video via the stable VideoCaptioner backend.
Video Translator
shang-zhu/violin
Dub a video into another language and generate subtitles using the default Together + Cartesia stack.
Speech Engine
elevenlabs/skills
Add real-time voice conversations to a custom agent runtime with ElevenLabs Speech Engine.
Video Translation
NoizAI/skills
Translate and dub videos from one language to another, replacing the original audio with TTS while keeping the video intact.
More from wy51ai/edulab
Edu Chem Video
wy51ai/edulab
A skill your agent uses when asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a chemistry problem (化学题: 氧化还原配平 双线桥 电子守恒, 物质的量计算, 化学平衡 三段式 平衡常数 转化率 反应速率, 离子反应, 电化学, 溶液 滴定…
Edu Analytic Geometry
wy51ai/edulab
把一道解析几何题解成一个自包含的交互教学网页：左栏题面 + 动态控制台（一个 可变参数滑块驱动实时重算的几何量 + 理论范围/定值指示），中栏 KaTeX 分步解析，右栏 2D Canvas 动态几何画板（椭圆/双曲线/抛物线/圆 + 动直线/动点 + 向量 + 标注 + 画笔涂鸦）。
Edu Solid Geometry
wy51ai/edulab
把一道立体几何题解成一个自包含的交互教学网页：左侧 MathJax 分步解析， 右侧 Three.js 可交互 3D 模型（分步高亮 + 镜头切换）。支持三种入口——给定文字题目、 随机出题、上传题目图片识别后解题。覆盖正方体/长方体、棱锥/棱柱、圆柱/圆锥上的线面角、 二面角、异面直线夹角、点到平面距离、体积等题型，统一用"建系+向量法"，并由 sympy 精确 计算驱动（答案、3D…
Edu Physics Video
wy51ai/edulab
A skill your agent uses when asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a physics problem (物理题: mechanics/力学 受力分析 牛顿定律 斜面 传送带 板块 平抛 圆周 能量 动量, optics/光学 折射…
Edu Chem Reaction
wy51ai/edulab
把一个化学反应做成自包含的微观 3D 交互演示网页：左/上为 Three.js 可交互分子动画 （拖滑块看断键·成键·原子重组，分步高亮），右为 KaTeX 反应方程 + 分步讲解 + 原子守恒计数 + 可选能量-反应进程曲线。支持三入口——给定文字反应/方程、随机出题、上传图片识别后演示。
Related
Works with
Categories
Questions about Edu Math Video
What does Edu Math Video do?
A skill your agent uses when asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a math problem (数学题, geometry, algebra, functions, motion/行程 problems), from a problem screenshot…. Edu Math Video is an agent skill from wy51ai/edulab. Use when asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a math problem (数学题, geometry, algebra, functions, motion/行程 problems), from a problem screenshot or text, with Chinese voice-over (GLM-TTS / 智谱 TTS), bilingual zh+en subtitles and hand-drawn canvas animation rendered to MP4.
When should I use Edu Math Video?
Edu Math Video fits situations like: asked to make an explainer / walkthrough video (讲解视频、解题视频、例题精讲、微课) for a math problem (数学题; motion/行程 problems); from a problem screenshot; with Chinese voice-over (GLM-TTS / 智谱 TTS).
How do I install Edu Math Video in Claude Code?
Run `npx skills add wy51ai/edulab --skill edu-math-video -a claude-code`. Or copy the skill folder (skills/edu-math-video in wy51ai/edulab) into .claude/skills/edu-math-video in your project. Claude Code loads it when a task matches its description.
How do I install Edu Math Video in Codex?
Run `npx skills add wy51ai/edulab --skill edu-math-video -a codex`. Or copy the skill folder (skills/edu-math-video in wy51ai/edulab) into .agents/skills/edu-math-video in your project. Codex loads it when a task matches its description.
Can I use Edu Math Video in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add wy51ai/edulab --skill edu-math-video -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/edu-math-video, .gemini/skills/edu-math-video, .github/skills/edu-math-video and .opencode/skills/edu-math-video in your project.
What does Edu Math Video need to run?
Going by SKILL.md and its folder, Edu Math Video needs Python, a shell and JavaScript for the scripts in its folder, the command-line tools its instructions call (node, bash and npm) and credentials named GLM_API_KEY. Our summary lists: Python 3; Node.js; A Bash shell; A credential in GLM_API_KEY.
Does Edu Math Video access the network?
SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Is Edu Math Video safe to install?
Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
What licence does Edu Math Video use?
Edu Math Video is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Edu Math Video use?
About 2.5k tokens (SKILL.md is roughly 9.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Edu Math Video?
Skills that share tags, products or a category with Edu Math Video: VideoDB Video and Audio Workflows (video-db/skills, 127 stars), Proofread (sonilo-ai/skills, 115 stars), CLI Anything Videocaptioner (HKUDS/CLI-Anything, 52k stars) and Video Translator (shang-zhu/violin, 1.1k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Edu Math Video?
wy51ai (a GitHub user) maintains it in wy51ai/edulab, which has 1,406 GitHub stars. The repository holds 6 skills in this directory. The repository was last updated on October 10, 2026.
Source: wy51ai/edulab on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.