vibe-knowledge-film skill
A skill turns a topic or script into a drawn explainer through a brief, storyboard, sample segment and local rendering.
- 1,095 ★ on the repository
- Last commit
Extended access
Why this level
- Scripts write project files and run a local browser and FFmpeg; font downloads use the network.
- Some study examples contain third-party subtitles and visual work that must not be claimed as original.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/vibe-knowledge-filmDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Scripts write project files and run a local browser and FFmpeg; font downloads use the network.
- Some study examples contain third-party subtitles and visual work that must not be claimed as original.
This is third-party code. Review the repository files before installing.
What it does
vibe-knowledge-film prepares a short educational video from a topic or script. The agent agrees the content and selects one of two visual styles, then creates subtitles, a storyboard and a separate code project. The engine draws frames with Canvas and JavaScript, captures them through a local browser and assembles a video. A sample segment comes before full production, allowing the user to check style and delivery before rendering the whole piece.
Who it is for. For creators and teachers who need a drawn video to explain a complex idea or behavior.
Good fit when
- Explain a scientific topic through a visual sequence.
- Illustrate a behavioral mechanism with a short fable.
- Check delivery on a short sample before committing to a full video.
- Keep a film as a separate editable code project.
Not a fit when
- You need photorealistic generation without a local browser.
- Chrome and FFmpeg cannot be installed on the machine.
- You intend to publish third-party study fragments as your own work.
Example request
Prepare an explanation of this topic, propose a style and show a 30-second sample before rendering the full film.Limitations
Rendering requires the Python dependencies, Playwright, installed Google Chrome and FFmpeg; without them the film code cannot become a video. Create each film project outside the skill directory so outputs stay separate from engine files during later edits. Background music is supplied separately by the user, who must check its reuse rights. Some filled briefs, subtitles and visual examples come from third-party short films for study; they must not be published as your own work.
How to disable. Remove the agent skill and stop its render process. Separately created film projects remain on disk.
FAQ
Which visual worlds are available?
Nocturne for difficult underlying concepts and Fable for psychological or behavioral patterns.
How is the world selected in the project command?
With `--world`: C for the first world or B for the second.
Can I start with a topic instead of a script?
Yes. It accepts a topic and agrees the content before generating the film.
How is a project created?
From the skill directory, run `python scripts/new_film.py /path/to/film-project --world C`. Replace the example path with your project’s absolute path outside the skill directory.
Why are fonts absent from the repository?
Download them with `python engine/get_fonts.py` before the first render. They use SIL Open Font License.
Discussion
No questions or reviews yet.
Related
SkillsA director skill: it interviews you about a video idea and turns it into a second-by-second shot script for HyperFrames
A skill that makes your agent film a short product video: concept, edit, sound and render on Remotion or HyperFrames
Vibe Vision Video is a video skill that keeps the project editable so its author or another agent can continue working on it.
A skill that finds B-roll and literal sources for lines, shows them in a Storyboard and hands over clips after you approve
A Claude Code and Codex skill that turns a topic into a narrated motion-graphics explainer video on Remotion with subtitles and chapters
A skill that turns a topic into a narrated 4K video: research, script, local TTS, Remotion composition and an FFmpeg mix
README in short
Examples show openings for a protein-folding film and a behavioral fable. Engine commands and presentation rules are organized into reference files.
README badge
Original SKILL.md
--- name: vibe-knowledge-film description: > 抖音「Vibe知识大赏」式的无旁白知识短片:2–5 分钟、16:9、没有人声,观众读底部一行字幕,同时看屏上把这句话演出来。 核心是讲解设计:给每个概念找一个外行认得的形象,画得像它本身(不用点、线、圈代替),句子里说到什么屏上就是什么、动词真的发生; 提问、命名、结论不硬配图。全片由代码逐帧画出(Canvas + seek(t) + Playwright + ffmpeg)。 内含四条参考片的逐句拆解、六种画法、找形象的方法、分镜表格式、两套外观、渲染引擎和合成配乐。 用户说"做一条 vibe 知识大赏的视频""像《信息论·惊奇之学》《运气的形状》《超常刺激》《自指》那样的片子""代码直出的知识短片""没有旁白的字幕科普片""把这个概念直观地讲清楚", 或者给了一个题目、一份文案要做成这种片子时使用。只做 16:9 横屏。有人声口播的讲解片不用这个。 --- # Vibe 知识短片 一种没有旁白的知识短片:观众读底部的一行字幕,同时看屏上把这句话**演出来**。 做这种片子的人是科普作者兼导演。色、字、辉光、片名卡都是皮;片子好不好,看的是观众能不能一眼认出屏上的东西、看见它在变、看完讲得出"因为……所以……"。 ## 配图不是讲解 这个 skill 做过一条失败的《折叠》(讲蛋白质和 AlphaFold):全片的主体是一串星点连成的折线。 字幕说"只换掉第 6 个,形状就变了,这就是镰刀型贫血",折线换了个卷法、变成玫红色。 颜色、字体、节奏都像参考片,可观众从头到尾没见过蛋白质是个什么东西、错一个字母到底坏在哪。 那串星点换给任何一段文案都成立——它是配图。 重做以后,同一段内容是:镜头从血管里的红细胞钻进去,找到一串珠子;第 6 颗变了色;拉出来,细胞弯成了镰刀; 再钻进去,珠串卷成一团抓住氧分子;变色的那颗在表面成了黏点;一团粘一团连成长杆;长杆从里面把细胞顶弯。 两版的对照图是 `examples/fold_old_vs_new.jpg`,重做的成片在 `examples/fold_C/`。 **动手前先读 `references/visualization.md`。** 四条参考片逐句的"文案 → 屏上是什么 → 观众为什么懂"、六种画法、什么时候不画、 怎么给概念找形象、分镜表怎么写,都在那里。本文件只是流程。 交付物:一个工程目录(`storyboard.md` 分镜表 + `film.js` + 渲染出的 `renders/final.mp4`),加上事实表。 ## 八条规矩 1. **先找形象,再写文案。** 每个概念先找到一样外行认得的东西来演它;句子里要有屏上看得见的东西和动作。画不出来的句子,改句子。 2. **屏上的主体叫得出日常名字,而且长得像它本身。** 一串有光泽的珠子、一个中间凹下去的红细胞、一条带刻度的数轴、一句古诗。 点、线、圈、柱子、方框不算;一个圆圈加一个标签也不算。 3. **句子里的动词,屏上真的做。** "砍掉一半"就盖住一半,"粘住"就真的卡上去。数是算出来的;示意的地方写明"示意"。 4. **先例子,后道理。** 名字、公式、结论,排在观众看过例子之后。开头十秒让观众看见一件事或者自己试一下,然后才是片名。 5. **该画的画,该停的停。** 提问、命名、结论、引文、出处,不配新画面——参考片里大约每五句有一句是这样。 6. **一个世界、一把尺子、几样道具用到底。** 一个场景撑两句以上,每句只改一处;结尾回到开头的画面。 7. **没有人声,外观统一,画面现写。** 文案就是字幕,每句一眼读完;一条片一套色、字、画法,每种颜色全片一个意思。画面和代码是为这份文案现写的,不套模板,不搬别的片子里的东西。 8. **每个事实带出处。** 年份、人名、数字上屏;不加水印,不加署名。