Video & Animation

Turn a student's own course material (lecture transcripts, slides, notes, problem sets) into narrated, animated study videos with graphs, worked examples, quizzes, captions, sound effects and optional Blender 3D scenes. Every number in the video must come from the course material; a checker enforces it. Use when someone asks to make study or review videos from their lectures or class materials.

Make a narrated how-to video of an app flow: real screens captured in Chrome, a Claude Motion film with one step per scene, English narration by ElevenLabs (Matilda) under each caption, mixed into one MP4. Use for how-to, tutorial, docs or QA videos of an app.

按需把一部成片拆成可复用的制作参考：测镜头节奏与响度，标注段落与音轨分工，把原片事实与可迁移方法分开， 导出不含原片人名台词的 production_reference.json 供下次制作参考。不在默认生产路径上。 触发词：拆片、拆解成片、制作参考、参考模板、production reference。

Production pipeline for mathematical and technical animations using Manim Community Edition. Creates 3Blue1Brown-style explainer videos, algorithm visualizations, equation derivations, architecture diagrams, and data stories. Use when users request: animated explanations, math animations, concept visualizations, algorithm walkthroughs, technical explainers, 3Blue1Brown style videos, or any programmatic animation with geometric/mathematical content.

Turn a topic into a finished short video (YouTube Shorts, Reels, TikTok vertical, landscape or square) without a video editor, by writing a YAML spec and rendering it locally with the open-source yamlreel CLI. Use when the user asks to make, script or render a short video, explainer, code tip, system-design video, or project/product launch video, or mentions yamlreel. Outputs an MP4, poster, contact sheet and ready-to-paste share copy. Not for editing existing footage.

Design and build the visual layer of a long-form talking-head video as one persistent world the camera moves through, instead of a sequence of separate graphics. Use when asked to edit a video, add motion graphics, design the visuals for a script, plan an edit, or fix an edit judged confusing, text-heavy, hard to follow, or "clearly made by an AI". Covers the persistent world, camera grammar, the spotlight rule, spatial open loops, ornament and token discipline, transcript timing, and the verification gates.

Đạo diễn và tạo video 3D dọc theo sát lời thoại/phụ đề, với hình ảnh đẹp, phong phú, có ý nghĩa và chuyển động đồng bộ; phù hợp nội dung tài chính, kiến thức, gia đình, tình cảm, phát triển bản thân và các chủ đề khác.

ランディングページにプロダクトデモを組み込むスキル。 CSSアニメーションによるプロダクトUIモック、ヒーロー動画、カード内動画など、 Webサイトへのデモ演出パターンを実装する。 トリガー例: "動画を入れて", "ヒーロー動画", "デモ動画を埋め込み", "hero video", "product video", "LP動画", "ランディングページに動画", "プロダクトデモ", "アニメーションで見せたい", "UIモック"

Make a vertical cartoon story short (about 40-90 s) in the "2D cartoon characters in a painted 3D world" style, with voice narration, free CC0 sound effects and scene-based music, rendered with three.js inside Remotion at 1080x1920. Use when the user asks for a video "like the bridge video", "in that cartoon 3D style", a history/science story short with cartoon characters in 3D, or says "use the cartoon short skill".

Turns any raw screen recording video (desktop app, website, web portal, SaaS, ERP) into a narrated, HUD-annotated MP4 user guide with synchronized SRT captions. The user ONLY supplies a video file path (plus optional editing flow notes or voice/tone preferences) — NEVER ask the user for a recipe JSON. Extracts 3 FPS frames via FFmpeg (e.g., ~630 frames for a 210s video), analyzes ALL frames first so the agent fully understands the workflow before editing, validates timing via --plan_only, and renders the final MP4 + SRT.

Creates a 1080p60 dark-stage product launch film from two inputs — an X post (URL or text) and a website URL or brand profile — using a self-contained Remotion template. Honeycomb hex mosaic open, brand lockup, headline-over-product-capture feature beats with ecosystem hex tiles, giant statement card, CTA card with pointer click, honeycomb burst, typed domain outro, and an automatic beat-matched soundtrack. Use when asked to make a launch video from a post and a site, turn a feature thread into a product launch spot, or render a brand launch video.

Automatic light and colour correction for videos (auto colour grade, fix exposure, white balance, brightness, make clips match; 自动调色/调光影/统一颜色/白平衡/曝光). Fixes dark, yellow, blue, over-exposed, grey or flat clips, decodes LOG footage (DJI D-Log, Sony S-Log3, Panasonic V-Log, Apple Log) and iPhone HDR, grades each shot of a finished edit separately, applies a look preset (natural, real-estate, warm, vivid, cool-modern, cinematic) and reframes to 9:16 / 16:9 / 1:1 / 4:5 without stretching. One file or a whole folder; originals are never changed. Use for any request to colour-correct, grade, brighten, unify colours or change the aspect ratio of videos. Not for cutting/editing the story.

Director-grade AI video production for any provider — cloud APIs (ListenHub, fal, MuAPI, others), local GPU (ComfyUI) or copy-paste prompts. Plans shots, runs pre-flight checks to prevent wasted renders, writes model-ready prompts with premium aesthetics, diagnoses failed takes, and edits clips into a finished video with ffmpeg. Use when the user wants to make, plan, fix, or edit an AI video, short video, ad, short drama, product or property video, image-to-video, voiceover or soundtrack, or asks to set up a video generation provider or local video model.

把一段口播文案直接做成"小人国效果"AI 视频的完整技能：自动准备多视角人物图与场地鸟瞰空镜，按组级策划拆出四套差异化视觉方案，用景别跳切/机位差45°/主体四隔离/微距爆点做电影感镜头组，再转 3D 白模运镜预演防翻车，最后自动调 即梦 Seedance 2.5 出 480P 成片。触发场景：小人国/微缩世界/巨物对比/微观视角视频、人物在巨型键盘桌面食物上的视频、需要"人小物大"尺度反差的 AI 视频、要求把文案做成电影感分镜并直接出片。

用代码做动画和视频：片头/动效/字幕动画/数据可视化动效/UI 演示/讲解片段/循环动图，输出 mp4/webm/gif/webp，并有机器验收回环保证「画面真的在动、没有丢中间态」。用户说「做个动画/动效/片头/开场/蒙太奇/动图/懒加载 banner/视频片段/演示动画/用代码生成视频/HTML 转视频/remotion 风格」，或要求把一段内容做成可以播放的片子时使用。不用于文生视频（云视频模型）、真人素材剪辑、配音配乐。

Music beds (BGM) and sound effects for a video, as WAV files with full provenance, handed to audio-production-skill MIX - generated locally (ACE-Step 1.5 for BGM, EzAudio for SFX, both optional installs, never downloaded) or chosen, trimmed and looped from the user's own library with a declared licence (no engine); seamless loop beds and transition stingers timed to a cut. Use when asked to make background music or a music bed of a given length and mood, a sound effect (a door, footsteps, a whoosh), a loopable bed or a stinger that lands on a cut, to build or search a folder of licensed BGM / SFX, or to fit a library track to a scene. EzAudio runs only after the user has been told about its training data. All ops are EXPERIMENTAL; nothing is mixed here.

从输入视频一气呵成生成中文解说成片或原声剧情短片。用户提供 .mp4 / .mov / .mkv / .webm，并要求剪辑、添加旁白、配音、总结或制作短剧/电视剧/电影/纪录片/科普解说时使用。在同一个自包含技能内完成视频理解、导演立意与选镜剪辑、解说撰稿与校验、TTS 配音及音画混音合成。触发词：视频解说、视频旁白、生成解说、视频 recap、video recap、voiceover、narration、recap。

Three.js animation - keyframe animation, skeletal animation, morph targets, animation mixing. Use when animating objects, playing GLTF animations, creating procedural motion, or blending animations.

Production-ready animation patterns for React / Next.js — button, modal, toast, stagger, page transitions, exit animations, scroll, and layout — built on motion-foundations tokens and springs. Use when animating a specific UI element in React or Next.js — button, modal, toast, stagger, page transition, or scroll.

Video and audio processing with FFmpeg. Use for format conversion, resizing, compression, audio extraction, and preparing assets for Remotion. Triggers include converting GIF to MP4, resizing video, extracting audio, compressing files, or any media transformation task.

This skill should be used when the user asks to "fix janky CSS animation", "make animation 60fps", "stop layout thrashing", "animate width/height/top/left smoothly", "convert animation to transform", "animate box-shadow performantly", "animate height auto", "FLIP animation", or "why is my scroll/hover animation choppy".

Plans, scripts, edits and checks short-form social media marketing videos (TikTok, Instagram Reels, YouTube Shorts, LinkedIn, Facebook) in the user's own editor (DaVinci Resolve, Final Cut Pro, CapCut, iMovie). Specialized for real estate - broker personal-branding videos, developer and project promos, on-location shoots at project sites and offices, and high-volume templated series aligned with sales and marketing teams. Covers creative briefs, hooks, scripts and voice-over, shot lists, vertical framing, captions and graphics, music and sound, platform sizes and pacing, timecoded revisions and export checks, while keeping claims, prices and disclosures approved and the project editable. Use when the user wants to make, plan, shoot, edit, review or publish a marketing or promotional social video, Reel, TikTok or Short, or mentions a broker, listing, property tour, developer project or video template.

Делает короткие ролики о фичах любого продукта пользователя — обучающие, промо, для инвесторов или лендинга — в стиле Linear/Stripe: интерфейс продукта в кадре, курсор нажимает, текст печатается, камера ведёт взгляд, голос называет происходящее. HTML-сцены с покадровым рендером через Playwright и ffmpeg, стиль из бренд-профиля проекта (токены, шрифты, логотипы из его кода), экраны и надписи — из кода или прототипа текущего проекта, 16:9 и 9:16 с субтитрами и без, на любом языке. Готовит текст и промт для озвучки в Gemini TTS, раскладывает присланную озвучку по словам, ставит черновую музыку или трек пользователя, сводит и проверяет файлы. Используй всегда, когда просят «сделай / запиши / собери видео, ролик, моушн, туториал, промо или анимацию на N секунд про фичу, экран или раздел», «вертикальный кат для сторис», «подготовь озвучку для ролика», «наложи озвучку или музыку», «сделай ролик динамичнее», — даже если слова «моушн» нет.

AI 视频制作 SOP，三种模式：广告模式（15–60 秒信息流广告）、TVC 模式（自家产品/品牌出镜的电视广告和品牌片，用真实产品照片做参考）、短片模式（1–10 分钟叙事短片、微电影、短剧，先出总纲和分场表，再一场一个文件）。把任何故事、剧本或一句想法，做成可直接执行的 AI 视频制作包（markdown）——需求收集 → 故事（3–5 句大白话 + 要传达的）→ 镜头表脚本 + 品牌风险清单 → 一致性锚点 → 产出物清单 → 人物三视图提示词 → 车辆/道具/场景资产提示词（Nano Banana Pro 文生图）→ Seedance 2.0 全能参考多镜头视频提示词 → 物料位（App 截图）/ 配音清单 → 剪映操作清单 → 上线前对照清单。后期最小化：牌子数字等画面文字直接生成，包装大字、结尾画面、角标用 Nano Banana 出成品图，剩下的只留剪映一步操作。不做关键帧、不做多宫格，视频提示词直接引用资产图。交付分三处：本地制作包是完整版；提示词逐条贴在聊天里；同步到 Notion 等共享位置的是不含提示词的团队版。Use when the user wants to make an AI video, ad film, commercial, TVC, product film, brand film, short film, micro film, short drama, storyboard script, video prompts, Seedance prompts, or turn a story into a video. 触发：做 AI 视频、做视频广告、广告片、TVC、电视广告、产品广告、产品片、品牌片、短片、微电影、短剧、5 分钟短片、分镜脚本、镜头表、视频提示词、Seedance 提示词、全能参考提示词、把故事做成视频、出一套制作包、Facebook 视频广告脚本。写视频提示词时优先用本 skill，不要用 tvc-director 的多宫格和 Phase 格式。