Video & Animation

Create polished Chinese knowledge-explainer videos with Remotion from a topic or source material. Use when the user asks for a Vibe 知识大赏-style science, technology, history, business, psychology, or general knowledge video; wants a Remotion project, animated explainer, coded video, narration script, storyboard, subtitles, preview clip, or final MP4; or wants to turn documents and verified research into visual video scenes.

Read a codebase and unfold it into an interactive, zoomable map of what the system does, so a newcomer reaches a developer's understanding. The overview is an architecture diagram with inputs and outputs marked; clicking a part zooms into what fits its code (process flowchart, sequence, state machine, data model, class diagram, C4, deployment, behaviour tree, timeline) or opens screenshots and videos. Boxes cite their source files, hover traces data flow, guided tours and a presentation mode teach it, and a built-in editor lets anyone refine it. Exports PNG, SVG, PDF, Mermaid, draw.io and Excalidraw. Use whenever someone wants to understand, explain, present, document, review or onboard people onto a codebase, repo, service, pipeline, ROS 2 stack, firmware or app, e.g. "diagram this repo", "how does this code work", "architecture overview", "walk me through this project" or "I'm bad at presenting what I built", even if they never say diagram.

Use for any CapCut / 剪映 work from Claude through the CapCut MCP — uploading local media, building or editing CapCut cloud drafts (video clips, music, sound effects, text, captions, filters, transitions, text templates), material/template search, automatic captions, AI rough cuts (AI Lab), AI image/video generation, and fixing the CapCut MCP connection or login. Adapts the official CapCut Codex plugin to Claude; read before the first CapCut tool call.

Seedance 2.5 skill for Dola AI 15-second — specialized Seedance 2.5 video generation with hard-locked single-segment 15-second output, bulk TXT/Notepad prompt queueing, resilient skip-and-continue failure handling, end-of-run retry analysis, and an opt-in 15-minute Dola AI 15-Second Mode watchdog. Normal video generation runs on autopilot without confirmation; the cron/watchdog stays OFF until the user manually says “dola ai 15 second mode activate”. Always generates each video as one complete call and preserves the original Seedance 2.5 generation rules.

Seedance 2.5 skill for Dola AI 30-second — specialized Seedance 2.5 video generation with hard-locked single-segment 30-second output, bulk TXT/Notepad prompt queueing, resilient skip-and-continue failure handling, end-of-run retry analysis, and an opt-in 15-minute Dola AI 30-Second Mode watchdog. Normal video generation runs on autopilot without confirmation; the cron/watchdog stays OFF until the user manually says “dola ai 30 second mode activate”. Always generates each video as one complete call and preserves the original Seedance 2.5 generation rules.

Use when planning and producing visual explainer videos in either short-form (under 60 seconds) or long-form (8–12 minutes), including scripts, researched story structure, AI visual prompts, motion-graphics directions, narration, editing handoffs, and quality audits.

XBEAR MEDIA'nın krem kanvas + editoryal kart stilinde (Seçkin Ayazlar Reels/YouTube) Remotion ile animasyonlu video kurgusu. Konuşan kafa videosuna motion graphics, altyazı, sayaç, kart, geçiş ve ses efekti eklerken; sahne planlarken; tasarım sayfası hazırlarken kullan.

将 Vibe 已选视觉机制导演成事件发生的过程，比较观看路径供用户选择，输出事件分镜、无字幕设计测试与连续验证计划。用于分镜策划和局部改镜；真实动画与声音验证交制作阶段。

用户要剪视频、拼片段、加片头剧照、抽帧/抽音轨、导出 mp4 时使用（Android 宿主 weizhi:video）。桌面无 MediaCodec 不适用。

Rough-cut a Cinch skills video in Descript in the house style. Use when the coach says "cut the new cornering video", "clean up this footage", or "add captions to the skills video". Needs the Descript connector.

根据 SRT 字幕制作可导入剪映的透明章节进度条。先确认章节与原话范围，再打开 Playground 让用户按名称选样，支持配色和人物替换，交付窄条透明 MOV。不用于文案短动效或完整视频剪辑。

制作 GitHub 热榜/排行榜类中文口播讲解视频的完整产线（暗色科技风 Remotion 成片）。当用户要求「做一期榜单讲解视频」「GitHub 趋势榜视频」「排行榜 Top10 视频」或要求与 koubo-gh 同风格的口播成片时使用。覆盖：快照抓取、口播稿、edge-tts 配音+字级时间戳、词锚动效、名次独立页、内容镜三件套、怎么用三步配图镜、全屏插图背景、机器闸（beat_lint/motion_check/sfx_check/contract_lint）、loudnorm 交付、竖屏封面。

Edit a video by selecting transcript words or sentences in the conversation, including filler words, then render the requested cuts to a new video. Use when someone wants transcript-based video editing in chat.

Read and summarize a single video URL or local video using metadata, chapters and captions first, then sampled frames with OCR and speech transcription when needed.

将 B站、YouTube、抖音、小红书等视频及多源网页文章转化为高信息密度、图文并茂 (含 Mermaid 架构图与精准截帧)、学术级引用的高质量技术博客，并自动推送到 Notion (直接对接 next-notion -> Hexo)。

Runs a fresh, targeted read of one video's actual frames and audio, for a question the video's fixed pre-processed context can't answer — a specific visual or on-screen detail, a storyboard or pacing walkthrough, or any video that has no pre-processed record yet: freshly uploaded, never tagged, or a video the user just attached in chat. Use only when the fixed context doesn't already cover what's asked; check it first, don't re-read a video for something already there (an attached video has no fixed context, so read it directly).

Turn a user-supplied English VLOG into this series' vertical learning video with bilingual captions, replay-and-explanation cards, supplied narration, and a clean blind-listening replay. Use for new episodes or revisions to this established style.

Use when adding, reviewing or fixing animation and transitions in a web UI - hover and press feedback, modals, dropdowns, toasts, page and scroll reveals, springs, easing and durations. Decides whether something should animate at all, then how, with reduced-motion support.

Lavori sui video - tagliare, unire, comprimere, convertire formati, estrarre audio, sottotitoli, cambiare risoluzione o fps, video per YouTube/TikTok/Instagram/Discord, registrazioni di gameplay (OBS), montaggio (DaVinci Resolve, Premiere, CapCut). Comandi ffmpeg pronti e spiegati. Usala quando si parla di video, clip, registrazioni, codec o dimensioni dei file video.

制作或修改中文 Vibe 知识大赏科普短视频，统筹事实核对、短字幕分镜、音乐卡点、本地渲染、双尺寸封面和发布文案；适用于从主题与音乐制作整套内容，或继续修改已有 Vibe 视频项目。

Creare animazioni di motion design (reel, intro di logo, video di brand) generandole in codice con Node, canvas e ffmpeg, da consegnare in clip pronte per il montaggio. Usala quando l'utente dice "fammi un'animazione", "un reel animato", "motion graphics per il brand", "animazione del logo", "video senza After Effects", o quando un'animazione ha scatti, momenti fermi o forme che spariscono.

Write an Instagram caption that earns the tap on "more": a hook in the first 125 characters, the keyword CTA above the fold, searchable keywords, up to 5 sized hashtags, and a pinned first comment, linted by a local checker that shows exactly what the feed displays. Use when the creator asks for "a caption", "caption this", "write the description", "hashtags for this", "fix my caption", or has a finished reel, carousel or photo and needs the words.

Generate short video clips with whatever the user has — Higgsfield CLI (Seedance, Kling, Veo), Google Gemini API (Gemini Omni Flash), Google Flow or the Gemini web app via computer use — and fall back to animated stills (Ken Burns) when no video provider exists. Covers text-to-video, image-to-video, first/last-frame transitions, reference-guided clips, reframing to 9:16/16:9, and burned captions. Use when the user asks to make a video or clip, animate an image, image-to-video, B-roll, a Reel/Short/TikTok clip, a transition, or to reframe or caption a clip. Not for narrated multi-scene explainers (creator-explainer) or ads (creator-ad).

Run an evidence-based system audit of Brand OS. Verifies aspect ratio compliance (strict 3:4 for carousels, 9:16 for video), 96px safe margins, proof bank substantiation, zero broken wiki-links, anti-slop compliance, and placeholder elimination. Saves a dated report in audits/.