Dashi Animation Style

Before you start: read this whole install prompt, tell me in a few lines what this skill does and what installing it involves, and wait for my go-ahead before you do anything else. Prerequisites: none for the install itself: no account, no API key, no paid service, and nothing else to install first; installing is only copying files. For using the skill afterwards, the machine must be macOS (the skill was only tested on Apple silicon; Windows is untested, and fonts and Chrome paths differ there) with Node 22 or later, ffmpeg (with ffprobe) and Google Chrome, all installed by me, not by you: the skill's own rules say software is for the user to install, and you must never download or install software for me. Optional for use: whisper-cpp for automatic transcription (without it, a written script plus silence detection is the fallback), and a few macOS on-demand fonts (Lantinghei, Yuanti, Shouzha, Pianpian and friends) that I download myself in Font Book if I want them; without them the affected styles fall back to PingFang. Allow real time per piece: one measured run took about 30 minutes for a 23-second voiceover in one ready-made style, and a brand-new style built from a reference image took 1 hour to 1 hour 45 minutes per style in the author's tests. Install "dashi-animationstyle" for me. It turns a 15-60 second Chinese voiceover into an animated explainer video in a chosen visual style: a 1920x1080, 30 fps mp4 with the voiceover and subtitles, ready to drop into Jianying. Every frame is drawn by the agent with HTML/Canvas code and captured frame by frame with headless Chrome; no image-generation model is used. It ships 7 ready-made styles (flat flowchart, hand-drawn notes, vintage newspaper, vintage paper-cut, surreal collage, 3D cartoon, black-background tech), can build a new style from a reference image or video I supply, or propose 2-3 style directions from a plain-language description for me to pick from, with a style dictionary, general explainer-animation rules and a local render pipeline. Repository: https://github.com/chuspeeism/dashi-animationstyle (the skill is the whole repository at the pinned commit: the SKILL.md at the root defines the skill, and it works from the pipeline/, knowledge/, styles/ and examples/ folders beside it, so the whole repository without .git is installed) Pinned version: install the code at commit aa620f3b859aa5d191b783b2eb0fb59263731c76 exactly. Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Clone https://github.com/chuspeeism/dashi-animationstyle into a temporary folder, check out commit aa620f3b859aa5d191b783b2eb0fb59263731c76, and copy the whole repository without its .git folder from that checkout. Then compare every file with the fingerprints recorded for this listing on 2026-10-19: [FINGERPRINTS] - .gitignore: SHA-256 e8d0fc52141865308edffa2aa6b7168328b82d7e44868da3e2cbfdc22573ef42 - LICENSE: SHA-256 ed01b8353d3bca3d8fe56bb30028a689159070aa2d3da4118a9256e2afb646f6 - README.md: SHA-256 391df20828f8c2b5acf9408a8fe61ad915f08fb9a7aaedeb54c347279eb08f84 - SKILL.md: SHA-256 1659d6734a53dd2589819166fa49fc59dd02c146fb277853215c6d1655e4f200 - docs/1_立体卡通.gif: SHA-256 8c4b5b5d7737085d0119abc2833b5488f65b13be0a68980a8caf181a2aaeba14 - docs/2_黑底科技.gif: SHA-256 0efedfd8e61e5d4cd66c632fa620be2916c3bdfc6ed28a5d0c7e72fc1f7719c5 - docs/3_超现实拼贴.gif: SHA-256 3973e7ece4bb4ddcbd141d04259b09728ef9144d82e4eb5b11da63920dab4bd8 - docs/4_复古剪纸.gif: SHA-256 78c31b197fcdc8ff10256fba1ad08824c92c3e3e77ed26fa9e739cb63c3fb7aa - docs/5_复古报刊.gif: SHA-256 bee92b9db6b07c4f60ff85cc719c75d98b270ebe38b0c8625673a671f0fa28a5 - docs/6_扁平流程图.gif: SHA-256 77d886b1a85a230f8aa07ecffc3e3eca44d1157e404528df4d07b479d611299e - docs/7_手绘笔记.gif: SHA-256 b4abcba0ac5f7ba5622de65c968684f0b44c24afea10dd9e9d2f08764fc532db - examples/token/timeline.js: SHA-256 9a45be0f78bc96bc91108a5928f6409e4b19eca1acc328f10652a3598f4080a8 - examples/token/作品.json: SHA-256 1c42aff65d4c541a92eb57372a1fde4bde59c830ee6c0d864472ff26a8b2e9a6 - examples/token/分镜.json: SHA-256 5d4df76384dc2a07f1e7d39e7d54fdb6aec07065858330750a929707843f06b1 - examples/token/分镜.md: SHA-256 01610966ee98a86f4b737e6028b0493baabc50c65d87cca100e6ca9a44e1a22f - examples/token/口播.wav: SHA-256 e77ae349f352aedcd4e2592d6c15e9fb8ef85f100c87fe855d3af631a683c842 - examples/token/字幕.json: SHA-256 065bc284bf5602221c270b22b90801d54997146638b9254bc08d8abfe7c1af82 - examples/token/字幕.srt: SHA-256 bc4d6a9bcb515025da960d4608a75d6a96e15a6d225611dd06157648dfb01461 - examples/token/稿子.txt: SHA-256 ea3a92688275c14b81df580d693b8700c9604557e468b78852bd8a368ed10f23 - knowledge/从11套里学到的.md: SHA-256 5bec10ee025e2bc1b7ceb2cc5101697631d989ac1ce4f1d327d683a8e4e2fb39 - knowledge/动效和画面变化.md: SHA-256 682654de68a7fb12ab30707b2f2e460bdcfbc70e929b97545471047b25441ed1 - knowledge/讲解动画通用规则.md: SHA-256 8e98e6309a96cf238d407169998cee8873029976af439b012a3cd25d1247aebf - knowledge/风格卡模板.md: SHA-256 974ae3a5a924ed7de1c0bba4b016bb4f6a20debb9d026401848f40859a8cf810 - knowledge/风格词典.md: SHA-256 545e246af58df4d22010eaf0c08e9f823a91156ab6953cbd5a2744059081ab96 - pipeline/render.mjs: SHA-256 2b3a1abaf5635e6dc3fa4131cbec6edf1aa03145df84427e8bde4c8ea828fc62 - pipeline/丰富度检查.mjs: SHA-256 eb73e747e7aede5faaac005813370c151ec8b8d9ebaadd181b0abc3e0d4b2e35 - pipeline/动效部件.js: SHA-256 c487f55345805564dc8bef1dea168b01c865baa0d46e16676573d38b33501996 - pipeline/取色.mjs: SHA-256 1035b28278864a38ff5e23a74481e7f6447ff227cf6e45addf538b7664afef60 - pipeline/字幕.mjs: SHA-256 fa147dc1bd99ba6007b207104313d4eb7f3edcea41a67863d1bfcb6bf2f80140 - pipeline/新建作品.mjs: SHA-256 127563fc4ba3244439ac1399fc3e17324c32e147f5bebbe2c5d02b0f9c856744 - pipeline/生成timeline.mjs: SHA-256 cfdfcaaa74b30044aac29c6519fa88eb6095a925a48d94901135a705694fc8f4 - pipeline/自检.mjs: SHA-256 7eb2590426967c09b6cbc4771e1d322c33814702f68884e9b0072fde1530c651 - styles/复古剪纸/模板.html: SHA-256 2d6961805607995320a18065dac03b3aa8ae0d415309182542d73fb20b9ef01b - styles/复古剪纸/示例_Token.html: SHA-256 ceba34d932a40210ee919b53d1cd3549b9009c39903ad5f000e6e845e7fcb746 - styles/复古剪纸/静帧/复古剪纸_13.70s.jpg: SHA-256 8adb8c4a26b6da027bb7e35de6456956f3e2101b2e3a1cd3f4d72f015397a0fb - styles/复古剪纸/静帧/复古剪纸_18.40s.jpg: SHA-256 280ebc6d221cd39ddec249dc20b03c4c2a8d371233d090e2b6a70e7d378cefcb - styles/复古剪纸/静帧/复古剪纸_7.58s.jpg: SHA-256 c1449a8fdfffd6e3d3f0eb03cfbd2804b1d2e6a8703b7091e375506bfce049b4 - styles/复古剪纸/风格卡.md: SHA-256 4183949355b6fe7d6380645af06ca867e288b664794ad7c7465204951e1afa9d - styles/复古报刊/模板.html: SHA-256 678005dc2ede0ee25b799347d51655381ca5503423e3390220aa735fd96d8a8c - styles/复古报刊/示例_Token.html: SHA-256 4e5588ad3d9002e8c4e0ff8f14acfc33a4cf4ec2a5acc25e7100a409b0752c52 - styles/复古报刊/静帧/复古报刊_1.90s.jpg: SHA-256 a75e5a98f3d3d12fc10de27962e42a0639f8d00365e20891008b29eb410ac118 - styles/复古报刊/静帧/复古报刊_13.68s.jpg: SHA-256 805cd18cdfd4ae495473325914de20b7c3f6bd8bac5ea574577573649f3f89c3 - styles/复古报刊/静帧/复古报刊_20.90s.jpg: SHA-256 1089f186c7bc143da51562897e5cc023e2a85a3ba07aa34944b8bedc9851d1de - styles/复古报刊/风格卡.md: SHA-256 64f76c1278288787720f7d14bf8b8d2818f2142dddbb188338bf52b0e859fc96 - styles/扁平流程图/模板.html: SHA-256 1b0bc47ad9e8f3e635b501a1dd350576a092672e06d5f1eaacf861bd3733c70b - styles/扁平流程图/示例_Token.html: SHA-256 b997916223594cd9babbb84338ac45ae0aafef99a6d25e4d92cab943ce64a2d0 - styles/扁平流程图/静帧/扁平流程图_12.40s.jpg: SHA-256 8e350230180734f16a5028c96ffbf506dde9cf899a95ec4472abca3f280608fd - styles/扁平流程图/静帧/扁平流程图_20.80s.jpg: SHA-256 48331a29e37662ebb1f90ee152179ecfb1987728108fd023f98c7b5f75ad2e03 - styles/扁平流程图/静帧/扁平流程图_7.60s.jpg: SHA-256 bb117aa8abe3aa5e6453446cfb113faf779550f39441f5c58500209e35fac412 - styles/扁平流程图/风格卡.md: SHA-256 19a8c89c4c48588998ed2565045e0ebd75970f61f95b9e6a8b64c6f847ee6fd8 - styles/手绘笔记/模板.html: SHA-256 5240f117cedead4a34fbff452923713c5c995228033f3fdf938030a9e8e37618 - styles/手绘笔记/示例_Token.html: SHA-256 fab673069ef7dbbe392d78d04690778f8573ec11569f6a13275c5c5e459e3c9c - styles/手绘笔记/静帧/手绘笔记_12.85s.jpg: SHA-256 647fcc6c0513c010e9fc60bdd13157a251463b2fa43608238ca18f3ad751168e - styles/手绘笔记/静帧/手绘笔记_20.55s.jpg: SHA-256 4c158b798bf7f7228e6dfa5fcd4428f425953ace18c782f87d75ca243cf09772 - styles/手绘笔记/静帧/手绘笔记_7.60s.jpg: SHA-256 08bba6863f5d58b9cd3265709c06b837b73f2da4b5e9415b1c1b7433a9b5b7ab - styles/手绘笔记/风格卡.md: SHA-256 1ede857e8ca9d048c8577b3415ff05418c97aa083d31291b2da68ed926a88b2b - styles/立体卡通/模板.html: SHA-256 ba61a840089f55037b90a4a7022ab709a037e333c16b54a455c376e57148a9ca - styles/立体卡通/示例_Token.html: SHA-256 9dc00304f0148a99d86026595e9e9ae57fcea7f07c9ef6c61335917cc4cc22f0 - styles/立体卡通/静帧/立体卡通_13.68s.jpg: SHA-256 601863980f28de7d44740cd503f5a56aaaf00ac6f81e159a4e7a1e1368cc9185 - styles/立体卡通/静帧/立体卡通_20.90s.jpg: SHA-256 f4ec26113413f83eddb6af1db3c0a09a31f65403ac1322c1451208fa0c71abe3 - styles/立体卡通/静帧/立体卡通_7.58s.jpg: SHA-256 3eeb39c4ebebf220a790ff14dbc1894551551b5284453783ad5f985799292e9a - styles/立体卡通/风格卡.md: SHA-256 3a1878f0d8096b636b487c2fe6e13bc2e7b62b3508d77aa80ec0cc5b23601d3b - styles/超现实拼贴/模板.html: SHA-256 ae00a9c21f8a30cc24a09c67d2614b489d23189cb3e5ed6f9cb711908808a8b1 - styles/超现实拼贴/示例_Token.html: SHA-256 26a0a161e265d821f62c6cad5c370243edc3293c9edfa820eeabc7aee2543edf - styles/超现实拼贴/静帧/超现实拼贴_12.55s.jpg: SHA-256 1ef2c576cf8ef44bf44c82a83e5521eed4db9de5d55a15407d423c3f7d3ba4b3 - styles/超现实拼贴/静帧/超现实拼贴_20.90s.jpg: SHA-256 0ffee2e8ed1524f5231554eaf6ca3aa7db070a2241100d2f11bd90330d6c122d - styles/超现实拼贴/静帧/超现实拼贴_7.60s.jpg: SHA-256 76745c151b355c45c200ba5d08be5932c70b2cc8803106a3d103035016d5736e - styles/超现实拼贴/风格卡.md: SHA-256 bb3ef2567d3bdd87fb7fad71757f177b3bd0c2ae9affe12fea70ac3e9e0c30f4 - styles/霓虹赛博/实测_Token.html: SHA-256 c4ba5a7b968b11becfa8814fe678c5efdc0b9f312880eb1a8aca7a43827df95c - styles/霓虹赛博/静帧/霓虹赛博_11.50s.jpg: SHA-256 9e39942b92ec8286ac52e2af40a174f49848a859a5a4683085f39c48c23706dd - styles/霓虹赛博/静帧/霓虹赛博_3.40s.jpg: SHA-256 a3f4b6762922b84570517c735d42bb090248bbee5fe7292ca2ee50276a02c043 - styles/霓虹赛博/静帧/霓虹赛博_7.45s.jpg: SHA-256 265120ac5b63a10f93956839bcac67a0d0cb599cfe459fd0f0b5a5f24ead758b - styles/霓虹赛博/风格卡.md: SHA-256 d9944c1b84b2d188af2e4fb552dc1cae494e9c027933e4552519f3089fc657c6 - styles/黑底科技/模板.html: SHA-256 db60b25014e72b30c2f4628a0d151600ac4886c51210e498de110b28909eaa53 - styles/黑底科技/示例_Token.html: SHA-256 062f3c8cf610ddccb726dccab1b426383cd90e7828f18f743754eb9c576e1ce5 - styles/黑底科技/静帧/黑底科技_13.75s.jpg: SHA-256 06e0183ac94591a54bca407c8613adfb7270327651fdb3e1b4c834bcc73ff3db - styles/黑底科技/静帧/黑底科技_20.90s.jpg: SHA-256 d18a73009134cbbfbdc55eaaf9c6395ca9bb911ace2819712fd1ef8713d0bede - styles/黑底科技/静帧/黑底科技_7.60s.jpg: SHA-256 24cf01843411c2be46849b4ceff5cc0777489650f802b99d1a1619467e416d06 - styles/黑底科技/风格卡.md: SHA-256 5e5ea4edc229f42539c649094731bab4a5acfbee6e0613ded8b0c4739f6d2911 If a file does not match, it changed since the listing was prepared: do not install it yet, tell me what you found and let me decide. 2. Read SKILL.md in full and tell me in plain words: what the skill does when used, which services it connects to, whether it reads or sends any personal information or keys, and anything that seems unusual. Stop on any red flag and tell me before going further. Confirm these points from the files: - The installed package is the whole repository: 79 files, 31,980,437 bytes in total, most of it media (7 demo GIFs in docs/, 24 style stills as JPGs in styles/, and the example voice recording). The working parts are SKILL.md (37,294 bytes, the full playbook in Chinese), README.md, LICENSE, 5 knowledge files in knowledge/, 8 plain-Node pipeline files in pipeline/ (an environment self-check, a work scaffolder, a subtitle builder, a timeline generator, the renderer, a colour picker, a richness checker and a shared effects library; Node built-ins only, no npm dependencies and nothing to npm-install), 8 style folders in styles/ (7 listed styles, each with a style card, a template page and a Token example page, plus an extra tested-once neon-cyber folder that is not in the style list), and the bundled Token example in examples/token/ (audio, script, subtitles, storyboard and timeline for a 21-second piece). - Nothing is sent anywhere during installation itself; the only download at install time is the repository from GitHub at the pinned commit. The files name no account, no API key, no telemetry and no remote service. - At use time, everything runs locally on files I provide, and nothing is sent to any external service: the pipeline scripts call node, ffmpeg/ffprobe and, only if I have installed it myself, the local whisper-cli for transcription; the renderer drives my local headless Chrome through the Chrome DevTools Protocol on ws://127.0.0.1 on a random local port (localhost only) and assembles the mp4 with ffmpeg. SKILL.md forbids external resources outright: no CDN, no web fonts, no web images, and no image-generation model. The only environment variables the code reads are CHROME_PATH (where my Chrome is), WHISPER_MODEL (where my local whisper model is) and FONT_CHECK_SIMULATE_NEW_MAC (a font-check test flag): file paths and a flag, never keys or secrets. My voiceover, script and any reference image stay in my own project folder; output works are never written inside the skill folder, and styles/ and examples/ are read-only (copied into a work folder before any change). - Do not install any software during this install or to make the skill work: if Node 22+, ffmpeg or Chrome is missing, or whisper-cpp or an on-demand font would help, tell me what to install or download myself (for example brew install node ffmpeg, Google Chrome, or optionally brew install whisper-cpp plus its model) and stop there. That is the skill's own rule, not just mine. - One file is not under the MIT licence: examples/token/ (the example voice file) holds the author's own voice recording, which the LICENSE file expressly excludes from the MIT grant; it is provided only for trying the skill and must not be reused in other works. The skill ships no third-party images or fonts; a reference image I supply for a new style stays in my own work folder, its copyright is my responsibility, and it never appears in the finished video. - The scope is one 15-60 second piece of Chinese voiceover at a time (longer videos: pick the few segments that most need pictures). Music, sound effects, transitions and mixing with live footage are done afterwards in an editor such as Jianying; the output is a plain mp4 with the voiceover. It was tested only on macOS on Apple silicon and only with Claude Opus 5.5; other models and Windows are untested, and whisper transcription can mishear proper nouns and English words, with my written script as the authority. 3. When I say "yes, install", install it as a skill: copy the whole repository without its .git folder into the agent's skills directory, in a folder named "dashi-animationstyle", keeping all 79 files. The upstream route in the README, git clone https://github.com/chuspeeism/dashi-animationstyle.git into ~/.claude/skills/dashi-animationstyle, installs the same content; either route is fine as long as every file matches its fingerprint and no .git folder is kept as part of the skill. 4. Verify with no network calls: frontmatter valid, all 79 files in place, fingerprints still matching. 5. Report what was installed, where, and what I still need to do myself (for example: install Node 22+, ffmpeg and Chrome myself if any is missing, optionally install whisper-cpp and download any on-demand fonts in Font Book myself, and provide my own voiceover file and script when I want a video; using the skill is a separate step that only starts when I ask for it, and the local use-time processing disclosed above happens only then, on files I name). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.