Video & Animation

The operator's standard workflow for making Minara marketing / launch / demo videos. Built on the hyperframes skill, with a stricter process - confirm style → high-level storyboard → copy → build each scene separately → confirm a per-second script → iterate on a local test page → render only once the user is satisfied.

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.

Create Minara product guide/tutorial videos from screen recordings supplied by the operator. Use for /minara-guide-video, “给录屏加字幕配音”, retiming a new recording cut, building a timeline approval player, or correcting caption/product meaning. Drives caption-recording from source analysis through pre-render approval, final verification, and local delivery. Not for synthetic HyperFrames marketing videos.

Plan, edit, composite, caption, and finish videos with a coordinated Remotion and FFmpeg workflow. Use for talking-head, podcast, educational, social, product, documentary, and narrated video editing, including synchronized B-roll and motion graphics.

Generate background music and synced sound effects for Minara videos from the per-second script. Translates each scene's mood and on-screen actions (taps, typing, success, notifications, transitions) into a cue sheet, then renders BGM + SFX locally (numpy synth, no external service), with optional ducking under the voiceover. Use for /minara-music, "配乐", "生成音乐", "BGM", "按脚本生成音乐", "加音效", or when a minara-video / minara-guide-video project needs a music bed.

Biblioteca modular e auto-contida (Subtitles Studio) para geração de Legendas Dinâmicas estilo "Lockup" (palavra gigante no centro + palavras de apoio encaixadas geometricamente nos vãos + efeito de luz/shine + anti-pisca) para vídeos verticais (Reels/TikTok/Shorts) no HyperFrames. Use quando o usuário pedir para gerar legendas dinâmicas, criar um novo estilo de legenda, mudar cores de destaque ou embutir legendas em composições HyperFrames.

The user's scriptwriting and storytelling skill for short-form video (Instagram Reels first, also TikTok, Shorts, YouTube hooks). Writes hooks, full spoken scripts, on-screen text scripts, captions, B-roll shot lists and CTAs, using proven viral format families (Advice, Opinion, Relatable, Storytelling, Easy-to-use), story structures (normal → explosion → new normal; stakes → big question → head fake → re-hook; the Kallaway hook → lock-in → body/re-hook format; but/therefore) and current platform research. Use whenever the user asks to write, plan, brainstorm, fix or rate a script, hook, caption, video idea, content calendar, story, or "what should I post", or says script, hook, idea, format, storytelling, caption, CTA, or "make this go viral".

The user's one video editing skill. Edit any video by conversation, from raw footage to a finished file. Instagram Reels / vertical 9:16 (the default; the user mainly posts on Instagram), YouTube 16:9 hooks and long form, and anything else (montage, travel, interview, tutorial, screen recordings). Transcribes with ElevenLabs Scribe, cuts retakes and dead air, colour grades with cinematic LUTs (including Apple Log), adds word-timed captions (one word at a time in red/white or red/black by default; yellow and documentary styles too), split-explainer layouts learned from reference Reels, HyperFrames motion graphics, punch-ins, B-roll, free sound effects, music beds, and renders locally. Use whenever the user drops footage, names a video file, or says edit, cut, Reel, Instagram, YouTube, captions, subtitles, grade, LUT, Apple Log, animate, overlay, highlight this screen recording, or render.

>- Klipper og designer videoer fra et råopptak. Skriver ned alt som blir sagt (lokalt, på norsk), foreslår kutt på en side i nettleseren som brukeren godkjenner, klipper bort pauser og feilstarter, og legger på motion graphics i merkevaren: store ord bak personen, ord som fyller skjermen, zoom, skjermopptak, teksting ord for ord, lydeffekter og rolig musikk. Designet vises i nettleseren før den ferdige MP4-en lages. Bruk når brukeren vil redigere et opptak, lage en reel, klippe en video eller skrive /video-edit.

The user's CINEMATIC editing skill, separate from the everyday video-editor. Use only when the user wants something more cinematic, artistic, filmic, moody or "like Gawx" — a personal documentary, a travel film, a making-of, a story Reel with a film look — or says creative, cinematic, film look, moody, aesthetic, short film, Gawx. Low-key red/teal film grade (Cine_Gawx LUT), grain, halation, vignette, framed 4:5 / 1.37:1 / 2.39:1 pictures, small red chapter cards, music-led montage bursts, object close-ups, bookend storytelling with the meaning said out loud. For normal talking-head Reels with one-word captions use video-editor instead.

Sachverhalte, Anleitungen, Schulung und Wissensvermittlung.

Analiza un archivo de video (short, reel, tiktok o video largo) para saber qué se ve, qué se dice, de qué trata, cuál es el mejor momento para el hook y la portada, qué capítulos tiene y qué problemas o faltantes hay antes de publicarlo. Extrae fotogramas y transcripción en local con ffmpeg y Whisper. Úsala cuando alguien comparta la ruta de un video y pida revisarlo, resumirlo o prepararlo para publicar, o como paso de análisis dentro de la skill publicar.

Analyze a TikTok video from its URL (or a downloaded video file) by extracting metadata, a timestamped audio transcript and sampled frames, then reading them to pull out facts, claims, steps, products, quotes and on-screen text. Use this whenever the user pastes a tiktok.com / vm.tiktok.com / vt.tiktok.com link, asks what a TikTok says or shows, wants a TikTok summarized, fact-checked, or turned into notes, or says Claude can't open a TikTok link. Also use for any short-form vertical video file (mp4/mov) the user wants analyzed. Claude cannot play video or fetch TikTok pages itself, so do not guess from the URL, run this skill first.

Create motion-design videos from a prompt - explainers, promos, kinetic typography, logo reveals, title sequences, data/infographic animations, 3D scenes, photo montages, social ads - entirely designed in HTML/CSS/SVG/canvas/WebGL with GSAP, Lottie and three.js, rendered frame-accurately to video and edited with music, beat sync and transitions. Also for designed scenes mixed with real footage. Use whenever the user asks for motion graphics, an animated video, an animation, an intro/outro, a lyric or kinetic text video, an animated infographic, or a "motion design" piece.

Make piano-score videos with original two-hand leaps, glow trails and camera. Review notation revisions, reuse sampled audio tools, and stop for score approval then actual keyframe/audition approval before exporting.

Edit videos from a natural-language prompt. Cuts footage to a story, color grades dynamically along the story arc (the look evolves with the narrative), adds animated text, music with ducking and beat-synced cuts, sound design, and motion-graphics transitions/titles designed in HTML/CSS/JS. Use whenever the user wants to cut, edit, trim, assemble, grade, color, title, caption, add music to, or add transitions/motion graphics to video, or make a trailer, teaser, reel, recap, promo or short film from footage.

Produce a narrated, subtitled screen-recorded demo video of a web app feature — scripted browser interactions recorded with agent-browser, a local Kokoro voiceover, and subtitles burned in with ffmpeg. Use whenever the user wants a demo video, walkthrough, screencast, feature tour, explainer or recorded demo of something in their app, wants to "show" a feature to a team, customers or stakeholders as a video, or asks for a voiceover or subtitles on a recording — even if they don't say "skill" or name any of these tools.

Turn a public media URL (YouTube, Apple Podcasts, 小宇宙, Spotify public episodes, Bilibili) into a timestamped transcript with the user's own VideoScript account, using the `videoscript` CLI. Use when the user asks for a transcript, subtitles, SRT/VTT, or "what was said in this video/podcast", when captions are missing or need cleaning, or when a long recording must be turned into text for notes, summaries or further AI work.

End-to-end video editing pipeline tailored for Moroccan Darija footage in FilmCraft: mobile MTP ingestion with chunk resumption, featherweight zero-lag editing proxies (960x720, -bf 0), faster-whisper Arabic/Darija word-level transcription with stdin-streamed timestamp injection, chronological master timeline assembly, and YouTube long-form to Shorts spin-off funnels.

Turn one prompt into a finished motion-graphics video (MP4) - script, storyboard, animated scenes (kinetic type, 3D, transitions, camera moves), voiceover, synced captions, sound effects, music ducking and a loudness-normalized mix. Use when the user asks for a promo, explainer, launch teaser, TikTok/Reels/Shorts, product demo or any animated video.

Generate professional voiceovers using ElevenLabs AI. Use when the user needs to create voiceovers for videos, audio narration, or text-to-speech content. Supports multiple voices with character presets (narrator, salesperson, expert) for natural delivery. Includes single scene regeneration for fine-tuning.

Continuous-scene Remotion camera system with mandatory intake, approved-reference review, anti-neon rules, and first-pass QA.

>- QA and finishing gate for ANY rendered video (generated or not): probe specs against expectations (duration, fps, resolution, codec, pix_fmt, audio), measure integrated loudness and normalize to −14 LUFS with an audio-only re-encode, generate tiled contact sheets and consecutive-frame strips for visual review, run a structured critic pass with a scoring rubric, and verify render determinism. Use whenever the user asks to check, verify, audit, review, diagnose, normalize, or prep a video for publishing — MP4/WebM/MOV — or when a video "looks off" and they want to know what's wrong before re-rendering.

>- Generate voiceovers with WORD-LEVEL timings: free multilingual TTS via edge-tts (default), local Kokoro/Piper, or ElevenLabs — and turn any audio into word timestamps with faster-whisper. Produces vo.mp3, a words/captions JSON (caption pages of 1–3 words with per-word time windows), and an embeddable vo.js for canvas/web video pipelines. Use whenever the user needs narration, voiceover, TTS, a dubbed track, text-to-speech in any language, or word-synced captions/subtitles (TikTok/Hormozi style) for a video or animation.