Video & Animation

Caveman form for internal agent traffic (briefs, agent returns, research fan-outs, workflow returns, tool-call notes). Fragments, one fact per line, code, paths, errors verbatim, nothing a reader pays for twice. The `internal` text register; forge text uses social-text, docs and human replies use full prose.

把已有中文口播、访谈、课程或产品素材剪成短视频并添加字幕；从素材检查、转录选段到本地 MP4 导出和质量检查。用户说剪短、剪口播、访谈切片、删停顿、加字幕、一句话出片时使用。

Hỏi MKT chọn kiểu hình cho một video của redsun-reels — "minh hoạ" (nhân vật và bối cảnh vẽ, bản hiện tại), "người thật quay sẵn" (clip Pexels/Pixabay) hay "người thật do AI tạo" (luôn có nhãn nội dung AI) — ghi vào brief và giữ các luật cứng không được vượt. Dùng ngay sau khi tạo brief (skill tao-reel bước 1, cả luồng tự động), hoặc khi MKT nói "dùng người thật", "người thật AI", "clip Pexels", "video có người thật", "đổi sang minh hoạ", "đổi kiểu hình".

Clay 3D style in motion - soft rounded plasticine forms, matte subsurface clay materials, pastel or playful palettes, fingerprints and handmade imperfection, soft miniature-set lighting and depth of field - plus motion rules for both smooth "claymorphism" (squash and stretch, jiggle, inflate pops, blob morphs) and stop-motion claymation (animating on twos, boiling surfaces, holds, micro-jitter). Includes Blender (via the blender MCP), After Effects compositing and Higgsfield prompt recipes. Use when the user asks for clay 3D, claymation, claymorphism, plasticine, play-doh look, soft 3D, cute 3D characters, Aardman-style, stop-motion 3D, or tactile handmade 3D motion.

Pro-Motion Editor — Claude Code skill for promo videos, reels, YouTube videos, ads and product motion graphics (9:16 and 16:9) the glyph-motions way — onetake's deterministic HTML composition + shutter-blur render pipeline, plus a hard-won rule set: camera that travels to every click, 3D perspective that changes per beat, glass panels that FORM, text written word-by-word on the spoken word, a hook that is the densest edit, layout variety, Instagram safe zone, real UIs. Use for any animation / motion graphic / ad / reel for a talking-head or product, Hinglish voice-driven cuts, and when the user says 'one take', 'glyph', 'animation', 'ad', 'reel graphics'.

Create platform-ready thumbnails from real video frames. Use when the user wants a thumbnail, cover image, YouTube cover, Shorts cover, Bilibili cover, Xiaohongshu cover, or other video poster image.

Flat 2D illustration and animation - geometric construction, solid fills, limited 60-30-10 palettes, flat cel shading, silhouettes, rounded consistent forms, and the motion rules that bring them alive (eased transforms, overlap and follow-through, shape morphs, stroke draw-ons, trim-path bursts, idle loops, parallax and infinite zooms, Duik character rigs). Covers strict flat, flat 2.0 / Kurzgesagt-style semi-flat, corporate explainer and mid-century variants, with After Effects and HyperFrames/GSAP recipes. Use when the user asks for flat design, flat 2D, vector animation, explainer animation, Kurzgesagt style, corporate explainer, character animation with shape layers, icon animation, or clean illustrated motion graphics.

Cut one long podcast, interview, course, livestream, or other source video into social-ready shorts, reels, highlights, or clip timelines from existing project media. Use when the user asks to cut a long video into Shorts, Reels, TikToks, Xiaohongshu posts, best moments, highlights, or multiple clips.

Write forge-facing text (issues, pull-request bodies, review comments, commit bodies, changelog titles): scannable human prose derived from adhd-format; keeps PR template, receipt fence, conventional commits, Keep-a-Changelog intact.

Turn multiple product shots, event footage, travel clips, gameplay moments, UGC/product footage, B-roll, or mixed media into social-ready reels, highlights, recaps, or montage-style short videos from existing project media.

AI video generation via Fal.ai, Seedance 2.0, Kling, MiniMax Hailuo, xAI Grok Imagine, and OFox. Use when the user wants to generate a video clip — text-to-video, image-to-video, first/last-frame transitions, reference-guided generation, multi-shot, or generatively editing / extending an existing clip.

Add motion graphics at the right moments without blocking the story. Use when the user wants to enhance talking-head, lecture, tutorial, interview, podcast, or creator videos with motion graphics.

Format complex technical reports, architectural reviews, and diagnostic outputs for high cognitive focus, executive clarity, and ADHD readability using visual hierarchy, bionic bolding, chunked lists, and alert callouts.

Plan AI short films with story, shots, prompts, and continuity.

用 AI Agent 端到端生产 45–65 秒 9:16 竖屏知识类短视频（视频号 / 抖音 / 小红书 / YouTube Shorts）：选题 → 口播脚本 → 分镜 → Gemini TTS 配音 → HyperFrames(HTML+GSAP) 代码动效 → 程序化音效 + Lyria BGM 混音 → 封面烧进首帧 → 自动质检 → 发布文案 → 一键发布到 Cloud Run 手机交付页（手机点开即可存相册、复制文案）。内含 3 道人工确认关卡和 3 道自动质检门禁（motion_qa 静止帧检测、impeccable 反 AI 味设计检测、只看 MP4 的独立审片 Agent）。Use when the user wants to make a short explainer / commentary video, turn a news topic or article into a vertical video, or batch-produce talking-point videos with code-driven motion graphics. Don't use for editing real camera footage (talking-head cuts) or for long-form >3 min videos.

按人的修改意见改一条已经生成好的视频（HyperFrames 项目），只改提到的地方，重新渲染成新版本并登记回工作台的内容卡片。工作台在内容详情里点「让 Claude 修改」时在后台调用，参数里带内容 id 和修改意见。全程无人值守，不能向任何人提问。

把流水线里一条内容的分镜脚本做成成片 MP4（HyperFrames 解说视频），登记回工作台的内容卡片。工作台点「生成视频」时在后台调用，参数里带内容 id。全程无人值守，不能向任何人提问。

把一批视频素材（DJI 无人机、相机、手机拍摄的旅拍/活动/探店 B-roll）剪辑成可发布的短视频成片——裁切画幅、生成中文字幕、配 BGM、做封面、校准码率与响度，输出竖屏 9:16 与横屏 16:9 双版本。当用户提供多个视频文件并要求「剪一个短视频」「做成视频号/抖音能发的片子」「加字幕配 BGM」「把这段素材剪出来」时应使用本技能。

Create original short videos using the 空中群集｜近表面转向到稳定落地 Creative DNA for MiniMax H3 or Seedance 2.0. Use when the user wants this causal, camera, motion, rhythm, or payoff structure with new subjects and surfaces.

>- Apple Human Interface Guidelines interaction and UX patterns. Use this skill when the user asks about "onboarding flow", "user onboarding", "app launch", "loading state", "drag and drop", "search pattern", "settings design", "notifications", "modality", "multitasking", "feedback pattern", "haptics", "undo redo", "file management", data entry, sharing, collaboration, full screen, audio, video, haptic feedback, ratings, printing, help, or account management in Apple apps. Also use when the user says "how should onboarding work", "my app takes too long to load", "should I use a modal here", "how do I handle errors", "when should I ask for permissions", "how to show progress", or "what's the right way to confirm a delete". Cross-references: hig-foundations for underlying principles, hig-platforms for platform specifics, hig-components-layout for navigation, hig-components-content for data display.

Personal brand and creator video: vlogs, tutorials, opinion pieces, day-in-the-life, and channel content. Authentic pacing, jump cuts, talking-head craft, and a mandatory asset-request protocol. Use for creators, coaches, founders, and solo brands.

Upgrade Remotion, and related packages

>- Turn a script or an idea into a production brief you can shoot from - the shot list, the b-roll, the on-screen moments, the gear, and a runtime budget that keeps the edit from sprawling. Use for "shot list", "production brief", "how do I film this", "what b-roll do I need", or planning a shoot before recording.

>- Draft YouTube chapters from a timestamped transcript and check basic timestamp constraints before the creator uses them. Use for "add chapters", "timestamps", "break this video into sections".