Video & Animation

Produce and publish a polished vertical app promo video from Android emulator captures, end to end and reproducibly, for any app (app-agnostic, spec-driven). Captures clean UI stills, composes 1080x1920 motion-graphic scenes, generates English and Urdu neural voice-over, synthesizes a continuous music bed, auto-times and assembles the cut with FFmpeg, builds a custom 9:16 and 16:9 thumbnail, then prepares uploads to YouTube, your Facebook Page, Instagram and TikTok. Use when the user wants an app promo video, marketing video, app ad, app trailer, Play Store or Google Play promo video, app demo, feature launch video, or a vertical social clip, in English or Urdu, shipped to the channels. Keywords and triggers - app promo video, app marketing video, promo reel, YouTube Short, YouTube reel, TikTok video, Instagram Reel, Facebook Reel, app ad, app trailer, app demo video, feature launch, bilingual Urdu and English voiceover, emulator screen-capture video, vertical 9x16

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。

Make a short vertical video (TikTok, Reels, Shorts) about a finding from real material — data, photos, satellite images, menus, documents — to start a debate and send people to the user's page. The material fills the frame; the website shows only as the source. Use it when the point of the video is what the user found (a price gap, a city that grew, a number that surprises), not how to use the site. For showing how an app works (a chatbot answering), use demo-video instead.

Design and build interactive web animation that feels crafted, not templated. Use when the user asks for motion on a website or web app - page-load or hero sequences, scroll-driven or scroll-triggered effects, parallax, pinned or horizontal scroll sections, smooth scroll, text reveals and split-text, hover and press micro-interactions, page or route transitions (View Transitions API), modal/popover/dropdown enter-exit, layout and shared-element animation, SVG line drawing and morphing, animated charts, Lottie or Rive playback, canvas or WebGL/Three.js scenes, or "make it feel smoother / more premium / award-winning". Covers GSAP (ScrollTrigger, SplitText, Flip), Motion (formerly Framer Motion and Motion One), Anime.js v4, CSS transitions/@keyframes/@starting-style/scroll-driven animations, View Transitions, Lottie, Rive, D3, Flubber, Konva and Three.js/React Three Fiber. Not for video rendering (use Remotion).

Makes a 6 s character scene for a reel in one of nine visual styles: hologram, cyberpunk neon, paper cutout stop motion, technical blueprint, thermal camera, Japanese TV anime, ASCII terminal, watercolour or claymation. A worker walks across an office carrying a stack of papers and the top sheet slips. Use when the user runs /looks, asks to show the same scene in different art styles, or names one of these styles for a short character animation.

Makes a 2D pixel-art explainer reel (9:16) in code. Chunky pixel characters with big heads, outlined sprites, a side-view pixel world (sky, clouds, mountains, grass, houses, trees, signs), speech bubbles, a pixel-dissolve transition and 1 to 2 word captions. Use when the user runs /pixel-world or asks for a 2D pixel, Growtopia-like, or retro block-world animation.

Make 9:16 first-person story-narration shorts in a chunky 3D block-toon style (rounded-box characters with painted faces, ink outlines, always-moving camera, word-by-word colour captions) from a script or voiceover, rendered in code.

Builds cinematic motion graphics in the browser (promo videos, openers, intros, bumpers, kinetic typography, illustrated explainer videos in 16:9 or 9:16) with HTML + CSS + GSAP (+ Three.js when needed), so the result moves like real video, not like presentation slides. Use this skill whenever the user asks for a "promo video", "opener", "animated intro", "motion graphic", "bumper", "kinetic typography", "cinematic text animation", "explainer", "explainer video" (visual journalism, educational cartoon, collage), or shows a reference promo/explainer video and wants a web version of it, even if they never say "motion graphic". Also use it when the user complains that a web animation looks "like a PPT/presentation" and wants it more cinematic, wants to render a web animation to an MP4 file, or wants a cartoon/animated explainer built DIRECTLY in After Effects through the Higgsfield bridge/MCP.

Plan and build a product launch video in code. Define the project, pick the right kind of video for it, write the story and beat map, build the animation as an HTML page driven by seek(t), render it to MP4 with Playwright and ffmpeg, cut it to music and sound effects, and check every frame. Use this skill whenever someone wants a launch video, product video, promo, teaser, announcement clip, demo video, hackathon demo, feature drop or motion graphics for a product, even if they only say "make a video for my app", "something like this launch video", or share a link to a video they like.

Entiende, identifica y diseña la iluminación de una toma, luz clave, relleno, contraste, ratio, lado inteligente / lado tonto, motivación y luces prácticas, y la traduce a un prompt en inglés para video con IA (Higgsfield, Seedance, Kling, Veo), a luces de Blender, o a un plano de luces. Úsala SIEMPRE que el usuario quiera analizar la luz de un frame, clip o película; decidir cómo iluminar una escena o un personaje; escribir la parte de luz de un prompt de video o imagen con IA; iluminar una escena en Blender; o cuando aparezcan palabras como luz, iluminación, sombras, contraste, key light, fill, relleno, contraluz, rim, chiaroscuro, "look cinematográfico", "se ve plano", "más dramático" o "más cálido", aunque no diga "iluminación".

用代码从零生成中文视频（科普片/纪录片/宣传片/动画短片）：写剧本→edge-tts 旁白→实测时长排时间轴→PIL 或 Canvas 逐帧渲染→ffmpeg 编码→numpy 采样级混音→自动质检。当用户要"做一个视频 / 生成宣传片 / 把报告变成视频 / 做动画 / 配音配字幕 / 出成片 / 修视频音画不同步 / 视频画质糊"，或提到 ffmpeg、edge-tts、逐帧渲染、字幕、分镜、配乐、混音时使用。 | Code-generated video production (Chinese narration, deterministic frame rendering, ffmpeg, edge-tts, QA gates). Use when asked to make a video, turn a report into a video, animate a story, fix A/V sync or blurry text in a generated video.

Decide y controla a dónde mira el espectador dentro del cuadro y en qué orden, encuadre (centro, tercios, proporción áurea), bloqueo y líneas de mirada, foco y profundidad de campo, movimiento de cámara, color y vestuario, líneas guía, cuadro dentro del cuadro, luz y sonido, y lo traduce a un prompt en inglés para video con IA, a cámara y foco en Blender, o a notas de dirección. Úsala SIEMPRE que el usuario quiera que algo "se note", que el ojo vaya a un producto, un personaje o un detalle; analizar por qué un plano funciona o se ve confuso; componer un frame, un anuncio, un storyboard o una escena de Blender; o cuando aparezcan palabras como composición, encuadre, punto focal, regla de tercios, foco, desenfoque, rack focus, "que resalte", "no se ve qué es lo importante" o "a dónde mira la gente", aunque no diga "dirigir la mirada".

El director de orquesta de todas las skills de cine y publicidad. Lleva un anuncio o pieza de video de principio a fin, tarea de fondo, nivel de conciencia y sofisticación del mercado (Schwartz), gran idea (Ogilvy, Sullivan), historia (Truby, Cron, StoryBrand), formato de comercial, guion por segundos, diseño de cada toma (capas, mirada, luz, ángulo), shot list con prompts en inglés para video con IA y revisión final, llamando a `publicidad`, `alfabetizacion-visual`, `dirigir-la-mirada`, `iluminacion-cinematografica` y `angulos-de-camara` en el momento justo. Úsala SIEMPRE que el usuario pida hacer, escribir, dirigir, storyboardear o revisar un anuncio, comercial, reel, video para Meta o TikTok, guion de video, concepto creativo, campaña con varias piezas o shot list, o diga "hazme un anuncio", "qué video hacemos", "esto no pega" o "dale una vuelta", aunque no nombre ningún libro.

Entiende, identifica y elige ángulos de cámara (picado, contrapicado, cenital, nivel de ojos, holandés, ras de suelo) y los traduce a lo que se necesita después, prompt en inglés para un generador de video con IA (Higgsfield, Seedance, Kling, Veo), parámetros de cámara en Blender, o un shot list. Úsala SIEMPRE que el usuario quiera analizar un plano, un frame, un clip o una película; planear tomas o un storyboard; escribir el prompt de cámara de un video con IA; colocar la cámara de una escena de Blender; o cuando aparezcan palabras como ángulo, picado, contrapicado, cenital, high angle, low angle, bird's eye, dutch angle, encuadre, "desde arriba", "desde abajo" o "cómo se ve más poderoso / más vulnerable", aunque no diga "ángulo de cámara".

Frames a talking-head cut in DaVinci Resolve Studio like a professional editor. Crops each camera setup to a clean, centred frame that removes unchosen clutter at the edges (appliances, magnets, spare chairs, lamps) and keeps chosen set dressing (plants, flowers), never cuts off the head or a gesture, adds punch-ins on visible jump cuts with the eyes kept in place, adds slow push-ins on key lines of plain setups, and stays inside the zoom budget of the source. Produces a new timeline version through resolve-bridge. Runs on its own and leaves a report to review afterwards. Use when the user says the frame looks cluttered, loose or amateur, jump cuts are too visible, the speaker is off centre, or asks for punch-ins, push-ins or a declutter crop.

Organises raw footage for a scripted video. Transcribes every recording locally, finds each take and matches it to a storyboard scene, scores completeness, broken words and audio, reads on-set verdicts, suggests the best take or a patch from two takes, catalogues b-roll against the storyboard's b-roll notes, writes TAKES.md and offers a rename with film-style names (S02A_T03_original). Use when new footage arrives, when the user wants takes organised, compared or picked, wants to know which scenes have no usable take, or asks to rename clips.

Website and web-app speed audit and fix - Core Web Vitals (LCP, INP, CLS, FCP, TBT, TTFB) and 69 checks: server location vs. visitors' country, rendering, JavaScript, third parties, images, video, fonts, caching, CDN, navigation, interaction, animation and monitoring. For sites built with AI builders (Lovable, Claude, Base44, Bolt, v0, Replit, Cursor), React/Next.js/Vue apps, Supabase/SQL and API latency, and WordPress. Ranks fixes by seconds saved into Critical / Should improve / Optional with a ready prompt per fix. Two modes: audit (report only, no changes) and fix (after the owner picks items or asks to speed up, implement safely and re-measure before/after). Use when asked to check, audit, speed up or optimize a site or app, to explain a PageSpeed result, why a site is slow, or before a launch. Triggers: בדיקת מהירות, האתר איטי, לשפר מהירות, לפני השקה.

根据用户提供的视频时长、主题和参考素材，自动判断文生视频、图生视频或参考生视频模式，并生成结构清晰、具有导演调度意识的 Muse Video 提示词；当用户要制作超过 10 秒的视频时，自动按最长 10 秒切割（如 25 秒切为 10+10+5），并逐段输出带"视频一（10秒）"标注的完整提示词。最终输出格式以本 skill 的输出格式规范为准。

Make a vertical animated cartoon short. Offers 10 ideas, writes the script, then turns the user's voiceover into a finished MP4 with music and SFX at half the voice level. Use for any cartoon short or animated story video request.

Design the sound of a 3D Cartoon Animation film - an original music bed, a quiet foley sound on every visible action, accents on story beats, measured levels against the voice, stop-downs before the biggest hits - with the engine's 24 synthesised effects, and master it to -14 LUFS. Use after the picture is locked.

Plan a 3D Cartoon Animation film shot by shot - one shot per narration clause, with the words, place, subject, action and its landing word, face swaps, camera move, number device, caption lockup and sound cues - and review it as rendered stills. Use after the voiceover (or with draft timing) and before building shots.

Build original 3D cartoon characters for any topic (ball head, bean body, stick limbs, mitten hands) with skin tones, brand-colour outfits, hair, hats, glasses and beards, and act them with painted faces that swap through a blink. Use when casting a film.

Turn any topic into a 3D Cartoon Animation film idea - the angle, the hook, the cast, the places, the numbers that become objects, and the one big literal picture. Use first, before writing the script.

Make a vertical (9:16, 1080x1920, 30 fps) cartoon story reel for Reels / TikTok / Shorts with the toon-reel-engine - the script method, the episode files, the Skia toon look, three-layer lighting, camera direction, acting, the mandatory frame checks, sound, mix and the final export. Use whenever someone asks for a new reel, episode, short story video, legend retelling or animated short with this engine, or asks to light, stage, check, render or finish one.