Video & Animation

Generate structured JSON world prompts for Tencent Hy-Nomas video world model in ComfyUI. Use this skill whenever the user wants to create video prompts for Hy-Nomas, needs help formatting world descriptions as JSON, mentions "Hy-Nomas", "world model", "segment prompts", or wants to generate videos with timed actions, camera control, or chained segments. Also use when converting plain text video ideas into the structured Hy-Nomas JSON format, or when debugging why a Hy-Nomas prompt isn't working. This covers all generation modes: text-to-world, image-to-world, camera-controlled, interactive drive, and multi-segment chained worlds.

Turn a raster photo into a V6 GemScript collage with a small tightly cropped high-contrast vintage black-and-white film print, generous negative space, a high-lightness low-saturation Morandi paper background, multicolor flat-back rhinestone calligraphy using exact custom text, controlled mixed-shape accents, organic stroke extensions, nonlinear loose-gem clusters, a source-responsive layout, and an optional automatic 3:4 composite that stacks the effect above the untouched original. Use for GemScript, rhinestone-script, Y2K jewel-sticker, vintage-film collage, custom English word or name, effect-only, and effect-plus-original image requests.

Make a demo video, highlight reel or GIF of a web app, a terminal or a 3D scene, in code, smooth and frame-exact. Use when asked for a demo video, a highlight reel, a product GIF, or when a screen recording looks janky.

Full start-to-finish video editing that runs every installed video skill as one pipeline. Cuts silences and retakes, plans the edit from the transcript, then adds captions, motion graphics, face-tracked effects, transitions, camera moves, b-roll, sound effects and music, checks its own work and exports the final MP4. Routes each beat to the right tool (awa-video-reel, the 45-effect engine, Brand Nat codes, HyperFrames, Remotion, the Video Editor plugin, MultiCam shot) and follows the saved style (my-video-style). Also makes fully animated videos from a script with an AI voiceover (ElevenLabs, Google Gemini, HeyGen or free Kokoro). Use when someone types /vaibhav-video-editor, says "edit this video start to finish", "do the full edit", "make this a finished reel", "edit it in Nick style", "make an animated video with a voiceover", or asks for an edit combining cuts, motion graphics, effects, transitions and sound. For a single effect only, use /effects instead.

哲学内核驱动的「无口播知识情绪片」生产：2-4 分钟、16:9 横版、无人声，中文主字幕+英文译文同屏， 全片由代码逐帧画出（Canvas/MS 路线 + seek(t) 确定性渲染 + Playwright 截帧 + ffmpeg）。 与同类 skill 的差异在内容层：五问选题门（写不出「献给每一个___的人」不立项）、 九列运动脚本级分镜表（不许一句话意图）、对标帧逆向先于写码、契约基准截图分段并行、 程序化密度自查与 impact 帧级对位。经过 90s 首跑与 135s 正片验证。 用户说「做一支 vibe 知识片 / 无口播知识短片」、给出一个概念要拍成情绪片、 或要求「哲学向知识动画」时使用。有人声口播的讲解片、剪辑已有素材、写实镜头为主的片子不适用。

用代码逐帧做视频（Coding Video）的风格库与方法：上百种纯代码可实现的视觉风格（C 类，标题材和技术：Canvas 2D、WebGL、SVG、Three.js、p5.js、Manim、Remotion 等），以及用真实素材（T 类：录像、录屏、照片、数据）和 AI 生成素材（G 类：生图、生视频、语音合成）由代码剪辑合成的做法，按内容类型分（科普、口播、带货、宣传片、短剧、MV、新闻、纪录片、教程、旅行、体育、儿童绘本）；每种一句可复制的风格提示词；选风格的索引；提示词写法、素材的来源和处理、配乐与旁白指南；以及从“选风格”到“出样片”再到“沉淀成风格包”的三级流程。给视频定风格、写风格提示词、让 AI 写渲染代码出片、用素材做视频、问“有哪些风格 / 哪种适合 X”、配声音时使用。

Create, revise, or review short product videos from a real app. Use for app demos, storyboards, or existing showtime videos. Supports iOS Simulator capture and local composition.

Edit a raw talking-head video into a finished vertical short in the user's reference style (STYLE_GUIDE.md) - transcribe, cut silences and mistakes lip-sync safe, storyboard, word-by-word captions with highlighted keywords, motion graphics that illustrate the speech, SFX on transitions, render 1080x1920 MP4. Use whenever the user sends/uploads a raw video to edit.

Universal, tool-aware AI video production skill. Use when a user wants to plan, create, edit, render, review, or publish videos from an idea, brief, article, product, script, or source assets across niches, languages, and platforms.

Create silent motion-graphics explainers from a prompt through ImageGen references, actual Three.js reconstruction, causal motion and deterministic MP4 rendering. Use for process explainers, cutaways, reassembly, product films, kinetic typography and data visualization; not for footage-only editing.

Specialized skill for creating technical, educational, and explainer faceless motion graphics videos using Remotion (React & TypeScript). Handles vertical 9:16 (TikTok, Reels, Shorts) and horizontal 16:9 (YouTube explainer) formats with 9 selectable visual archetypes, mandatory character-driven storytelling (<CharacterFigure />), Stabilo highlighter wipes (<StabiloHighlight />), atmospheric dust & aurora glow (<DustParticles />, <AuroraMesh />), automatic audio sync & BGM ducking, and the HPMR Scriptwriting Blueprint.

成片工坊总入口：画风视频、音乐 MV、科普动画、三维微缩场景都从这里进。先追问到每个细节（场景、人物、情绪、结构、声音、禁忌），不知道选什么就给样片墙和内容点子；画风写成数值卡让用户确认，分镜联系表确认后才生产；做完交给独立子 agent 评委打分迭代，最后归档到 ~/video-forge。用户说「做个视频」「做个动画」「做 MV」「画风视频」「做个微缩场景」「用 XX 方案做」「不知道做什么画风」时用。

Create original cinematic 3D/stylized vertical videos from an idea or a script + MP3 in 20 Audira styles: picks the style, previews, then renders 1080x1920 60fps with Remotion. By Agus Dwi Rianto (AUDIRA MDA).

Build, render, and verify deterministic code-driven videos (narrated explainers, vertical 9:16 shorts, lyric music videos, data stories, and audio visualizers) using Canvas 2D, procedural Web Audio, Playwright, and FFmpeg. Includes 86 visual style specifications, audio timing analyzers, and automated verification gates for seek determinism, safe-area text bounds, cut harshness, colour fidelity, and loudness. Use when asked to create a video by code, programmatic animation, faceless explainer, vertical short, or lyric visualizer.

Direct, render and verify motion graphics with a deterministic HTML-to-video pipeline. Each piece borrows its visual grammar from a real "source world" artifact, records every choice in a decision ledger, plans timing in a frame-numbered beat sheet, renders frame-exact MP4, ProRes (with alpha) or WebM through a bundled parallel renderer, and is measured by a bundled checker for holds, loudness, flashes, colour tags and platform safe zones. Use for kinetic type, explainers, narrated video essays and paper-collage documentaries (the look people call "Vox style"), cinematic showreels and brand films, title sequences, animated charts, maps, globes and flight routes, product promos, lower thirds, counters, social verticals (Reels, Shorts, TikTok) and editor overlays, for Remotion or HyperFrames projects, and whenever someone wants an idea, script or dataset turned into a short video, even when the words "motion graphics" never appear.

Dérusher une vidéo parlée (face caméra, vocal, screen rec) : couper les silences, les euh/hum/mmh, les bégaiements, les faux départs et les phrases reprises. Sortie = un MP4 monté propre, un export prêt à poster et le rapport des coupes, dans $DERUSH_OUT_ROOT/<IMG>/. FR/EN. Triggers: derush, dérushe, coupe les silences, enlève les euh, jump cut, nettoie ma vidéo, monte ma vidéo.

Keep the conversation's title showing where a piece of work stands (rendering, needs review, done, failed), so someone scanning their chat list sees which job is still running and which is ready without opening each one. Use in any job that runs long or waits on a person (a video render, an image set, an upload, a build, a review round), whenever its state changes, and when someone asks to see the state in the chat name, sidebar or title. Applies only where you can rename the conversation.

Make on-brand image ads, video ads, landing pages and websites with the Glorya connector. Use it whenever the user asks for an ad, a video ad, a landing page or a website for their brand, or mentions Glorya.

Turns a long local video (podcast, talk, interview, stream) into short vertical 9:16 clips with face-tracked framing, animated captions, titles and hashtags. Claude picks the most viral moments from a timestamped transcript; a local ffmpeg pipeline does everything else. No API keys. Use when the user asks to make shorts, reels, TikToks or clips from a video or podcast, to find viral or quotable moments, or to re-cut, retitle or re-render clips made earlier.

Edita reel falado (Instagram, TikTok, Shorts) direto no computador. Transcreve localmente, corta pausas e erros com aprovação, junta vários vídeos, alterna planos, aplica transições, coloca texto na tela (resumo ou legenda da fala), prints e efeitos sonoros, segue referências de estilo e exporta em 9:16. Use quando pedirem "edita esse vídeo", "faz o reel", "junta esses vídeos", "corta as pausas", "coloca legenda", "coloca o print", "põe um efeito sonoro", "usa esse estilo" ou chamarem /editar-reel.

Use when the user wants any machine or mechanism (real or imagined) turned into an interactive, offline, single-page 3D simulator they can take apart — a cutaway, exploded view, "how does X work" explainer, drivable model, or a recorded film of one with a scored soundtrack. Runs the teardown Claude Code skill.

Make videos that look drawn by hand (pen, pencil, charcoal, marker, crayon, watercolour, ink) with p5.brush inside HyperFrames, in any style. Ships use cases to start from - a hand-drawn explainer, an animated doodle story, and redrawing a real video of a person - and grows by adding use-case folders. Use when someone asks for a "hand-drawn", "drawn", "sketched", "doodle", "sketchnote", "whiteboard", "watercolour" or "rotoscope" video or animation, "draw my video", or "turn me into a drawing". Runs on Windows and macOS.

Direct Sticky Character Studio — build, inspect and export animated silent emotional short films (loneliness, social anxiety, burnout, discipline) from local photos (e.g. Google Flow) and music using the sticky-studio MCP tools or `node cli.js`. Use when the user asks to make, edit, preview or render a Sticky hoodie-character video, storyboard, or silent film.

Build a short vertical (9:16) talking-head explainer with animated graphics, captions, and a theme you can swap. Use when someone asks for a short explainer video, a vertical teaching clip, a themeable motion graphic over a presenter, or a portable HTML canvas render pipeline without a heavy framework.