[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7B34YRMMZX89A2DYC22MKRF":3,"skill-same:01M7B34YRMMZX89A2DYC22MKRF":49},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":32,"ownerDescription":33,"updatedAt":34,"install":35,"resources":48},"01M7B34YRMMZX89A2DYC22MKRF","reference-driven-graphics",null,"Build custom branded animated graphics (three.js\u002Fcanvas\u002Fparticle effects, hero animations, generative visuals) from visual references or Figma sketches — the reference → prompt → exact-geometry → screenshot-iterate → parameterize → embed workflow. Use when the user wants an animation \\\"like X's site\\\" (Stripe, Linear, etc.), wants to turn a Figma sketch or SVG into an animated graphic, needs to embed generated animation code into Webflow\u002FReact\u002FVue, or wants to scale one effect across many cards\u002Fsections.","gethamster\u002Fskills","gethamster","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F215697643?v=4","01M7B2R6PGB3SDZZ71E0WE2S74","Hamster","01M7B2R6Q21HWJRXZZWB78BNDG",4,"MIT","https:\u002F\u002Fgithub.com\u002Fgethamster\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002F.experimental\u002Freference-driven-graphics\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: reference-driven-graphics\ndescription: \"Build custom branded animated graphics (three.js\u002Fcanvas\u002Fparticle effects, hero animations, generative visuals) from visual references or Figma sketches — the reference → prompt → exact-geometry → screenshot-iterate → parameterize → embed workflow. Use when the user wants an animation \\\\\\\"like X's site\\\\\\\" (Stripe, Linear, etc.), wants to turn a Figma sketch or SVG into an animated graphic, needs to embed generated animation code into Webflow\u002FReact\u002FVue, or wants to scale one effect across many cards\u002Fsections.\"\nversion: \"1.0.0\"\ncategory: \"Marketing\"\n---\n\nTurn visual references and hand-drawn sketches into production animated graphics through short, verifiable iterations. Based on relate_studio's workflow (Syntetika\u002FStripe case) plus lessons from the mermaider pipeline.\n\n## When to use\n\n- \"I want an effect like [site]'s hero\" — particles, morphing data-viz, generative motion\n- A Figma sketch\u002FSVG exists and must become the *exact* geometry of an animation\n- Generated animation code needs embedding in Webflow, React, Vue, or plain HTML\n- One effect must scale across many cards\u002Fsections with different content\n\n## Core rules\n\n1. **Always start from 1–2 concrete references.** Never design from adjectives. Get screenshots or a URL of the target effect. Identify the tech with Wappalyzer or by reading the page source (usually three.js, canvas 2D, or CSS). State the reference explicitly in the first prompt.\n2. **First prompt is plain words, zero detail.** Describe what the effect *does*, not how to build it (\"particles form illustration A, bloop into a transition state, reform as illustration B\"). Take the first working iteration, then refine. Never front-load every requirement.\n3. **Strip AI cruft immediately.** First-pass output always includes floating text, gradients, watermark-ish labels, extra UI. Remove them as the second prompt, before any refinement.\n4. **One change per prompt.** Every refinement is its own short message. Never batch five adjustments — you can't tell which one broke it.\n5. **Exact geometry comes from SVG, never from description.** When the shape matters, draw it (Figma pen tool → export SVG) and hand the file over: \"rebuild the animation to match this geometry.\" Describing a shape in words is the anti-pattern; the SVG *is* the spec.\n6. **Debug with screenshots, not adjectives.** When something looks wrong (gaps, overflow, jitter), screenshot it, circle\u002Fpoint at the problem, describe what bothers you. 2–3 screenshot rounds beat 10 text rounds.\n7. **Name the animation stages to fix jerkiness.** Stalls and snapping mean the model has no state model. Spell it out: `static → transition → static`, ask for eased, overlapping transitions between named states. For loops: every animated property must be a pure function of loop phase t∈[0,1), state at t=0 equals t=1, no accumulation (`+=` in a tick function is a bug).\n8. **Parameterize before shipping.** All tunables live in one config block at the top (particle count, durations, colors, parallax strength) — and colors come from the project's brand tokens, not the model's taste. Ask \"what do I adjust and where\" and make the answer be \"the config block.\"\n9. **Decouple content from canvas via attributes.** Text\u002Flabels never live inside the animation. Ask for a `data-*` attribute contract so markup (any framework) triggers\u002Ffeeds the effect, and one shared script serves every instance. This is what makes the effect scale across cards.\n10. **Embedding is a question, not a guess.** Ask \"how do I get this onto [my stack]?\" — for Webflow: an Embed block for markup + the engine in Project Settings before `\u003C\u002Fbody>`. For React\u002FVue: a component wrapping the canvas with props mapped to the config block.\n\n## Worked example (the Syntetika case)\n\n1. Reference: Stripe's data-viz particle morph. Wappalyzer says three.js.\n2. Prompt: *\"I'm trying to figure out how this effect on Stripe's site is created — the data visualisation changes per selected feature and particles transition through a bloop state (screenshots attached). Looks like three.js.\"*\n3. Strip the floating text it added by default.\n4. The graph shape was wrong → drew it in Figma, exported SVG, sent it: Claude parsed the path and rebuilt particle targets to match.\n5. Gaps and overflowing lines → screenshots + \"lines spill past their boundary here\" → fixed in two rounds.\n6. *\"Can I keep the text outside the animation and connect via attributes?\"* → engine split from markup; text freely styled in Webflow.\n7. Tuned config: 4000 particles instead of 2000, longer transition, parallax on.\n8. Scaled: each card carries `data-*` attributes; one script in the body tag drives all of them.\n\n## Anti-patterns\n\n- **Describing geometry in prose** → model invents a different shape every round. Draw it, export SVG. (Failure mode: infinite \"no, more curved\" loops.)\n- **Batched change requests** → regressions you can't bisect. One change per prompt.\n- **Text baked into the canvas** → can't restyle, can't localize, breaks responsive layout. Attributes contract instead.\n- **\"Make it smoother\"** without naming stages → the model reshuffles easings blindly. Name the state machine.\n- **Accepting magic numbers scattered through the code** → next week nobody can tune it. Config block at top, brand colors injected.\n- **Skipping the reference** → brand-less generic output (\"AI slop gradients\"). Reference first, always.\n\n## Verification checklist before shipping\n\n- Loops perfectly (record 2 cycles, frame 1 == last frame)\n- No layout shift; canvas sized by container, devicePixelRatio handled\n- Reduced-motion fallback (`prefers-reduced-motion` → static first frame)\n- Config block documented; brand hexes only\n- Works with the real markup\u002Fattribute contract, not just the demo page\n","Turn visual references and hand-drawn sketches into production animated graphics through short, verifiable iterations. Based on relate_studio's workflow (Syntetika\u002FStripe case) plus lessons from the mermaider pipeline.\n\n## When to use\n\n- \"I want an effect like [site]'s hero\" — particles, morphing data-viz, generative motion\n- A Figma sketch\u002FSVG exists and must become the *exact* geometry of an animation\n- Generated animation code needs embedding in Webflow, React, Vue, or plain HTML\n- One effect must scale across many cards\u002Fsections with different content\n\n## Core rules\n\n1. **Always start from 1–2 concrete references.** Never design from adjectives. Get screenshots or a URL of the target effect. Identify the tech with Wappalyzer or by reading the page source (usually three.js, canvas 2D, or CSS). State the reference explicitly in the first prompt.\n2. **First prompt is plain words, zero detail.** Describe what the effect *does*, not how to build it (\"particles form illustration A, bloop into a transition state, reform as illustration B\"). Take the first working iteration, then refine. Never front-load every requirement.\n3. **Strip AI cruft immediately.** First-pass output always includes floating text, gradients, watermark-ish labels, extra UI. Remove them as the second prompt, before any refinement.\n4. **One change per prompt.** Every refinement is its own short message. Never batch five adjustments — you can't tell which one broke it.\n5. **Exact geometry comes from SVG, never from description.** When the shape matters, draw it (Figma pen tool → export SVG) and hand the file over: \"rebuild the animation to match this geometry.\" Describing a shape in words is the anti-pattern; the SVG *is* the spec.\n6. **Debug with screenshots, not adjectives.** When something looks wrong (gaps, overflow, jitter), screenshot it, circle\u002Fpoint at the problem, describe what bothers you. 2–3 screenshot rounds beat 10 text rounds.\n7. **Name the animation stages to fix jerkiness.** Stalls and snapping mean the model has no state model. Spell it out: `static → transition → static`, ask for eased, overlapping transitions between named states. For loops: every animated property must be a pure function of loop phase t∈[0,1), state at t=0 equals t=1, no accumulation (`+=` in a tick function is a bug).\n8. **Parameterize before shipping.** All tunables live in one config block at the top (particle count, durations, colors, parallax strength) — and colors come from the project's brand tokens, not the model's taste. Ask \"what do I adjust and where\" and make the answer be \"the config block.\"\n9. **Decouple content from canvas via attributes.** Text\u002Flabels never live inside the animation. Ask for a `data-*` attribute contract so markup (any framework) triggers\u002Ffeeds the effect, and one shared script serves every instance. This is what makes the effect scale across cards.\n10. **Embedding is a question, not a guess.** Ask \"how do I get this onto [my stack]?\" — for Webflow: an Embed block for markup + the engine in Project Settings before `\u003C\u002Fbody>`. For React\u002FVue: a component wrapping the canvas with props mapped to the config block.\n\n## Worked example (the Syntetika case)\n\n1. Reference: Stripe's data-viz particle morph. Wappalyzer says three.js.\n2. Prompt: *\"I'm trying to figure out how this effect on Stripe's site is created — the data visualisation changes per selected feature and particles transition through a bloop state (screenshots attached). Looks like three.js.\"*\n3. Strip the floating text it added by default.\n4. The graph shape was wrong → drew it in Figma, exported SVG, sent it: Claude parsed the path and rebuilt particle targets to match.\n5. Gaps and overflowing lines → screenshots + \"lines spill past their boundary here\" → fixed in two rounds.\n6. *\"Can I keep the text outside the animation and connect via attributes?\"* → engine split from markup; text freely styled in Webflow.\n7. Tuned config: 4000 particles instead of 2000, longer transition, parallax on.\n8. Scaled: each card carries `data-*` attributes; one script in the body tag drives all of them.\n\n## Anti-patterns\n\n- **Describing geometry in prose** → model invents a different shape every round. Draw it, export SVG. (Failure mode: infinite \"no, more curved\" loops.)\n- **Batched change requests** → regressions you can't bisect. One change per prompt.\n- **Text baked into the canvas** → can't restyle, can't localize, breaks responsive layout. Attributes contract instead.\n- **\"Make it smoother\"** without naming stages → the model reshuffles easings blindly. Name the state machine.\n- **Accepting magic numbers scattered through the code** → next week nobody can tune it. Config block at top, brand colors injected.\n- **Skipping the reference** → brand-less generic output (\"AI slop gradients\"). Reference first, always.\n\n## Verification checklist before shipping\n\n- Loops perfectly (record 2 cycles, frame 1 == last frame)\n- No layout shift; canvas sized by container, devicePixelRatio handled\n- Reduced-motion fallback (`prefers-reduced-motion` → static first frame)\n- Config block documented; brand hexes only\n- Works with the real markup\u002Fattribute contract, not just the demo page\n",{"name":5,"version":24,"category":25,"description":7},"1.0.0","Marketing","main","fe3bb1476c3ec23d8a366eac0fa8cd9f0852360a","skills\u002F.experimental\u002Freference-driven-graphics",0,"https:\u002F\u002Fgithub.com\u002Fgethamster\u002Fskills","https:\u002F\u002Fgithub.com\u002Fgethamster","https:\u002F\u002Ftryhamster.com","We're building the AI-native future of work, unlocking the promised acceleration for your whole team, from idea to impact.","2026-11-05 11:33:33",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":47},[37,40,43],{"tool":38,"command":39},"npx","npx skills add gethamster\u002Fskills --skill reference-driven-graphics",{"tool":41,"command":42},"bunx","bunx skills add gethamster\u002Fskills --skill reference-driven-graphics",{"tool":44,"command":45},"pnpm","pnpm dlx skills add gethamster\u002Fskills --skill reference-driven-graphics","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgethamster\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002F.experimental\u002Freference-driven-graphics\u002FSKILL.md",true,[],[50,57,64,71,78,85],{"skillNo":51,"name":52,"nameZh":6,"description":53,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":19},"01M7B5MWYW0KE251276S56KQDZ","writing-waterfall-requirements-documents","Write a waterfall requirements document: elicit, structure and review testable, traceable requirements, then baseline the specification before design.","https:\u002F\u002Fgithub.com\u002Fgethamster\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-waterfall-requirements-documents\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M7B5MRCXWWPKREPB1CHQ7H90","writing-vision-and-values-statements","Write a V2MOM vision a reader can test and a ranked list of values that settles real tradeoffs, so the rest of the document has a clear anchor.","https:\u002F\u002Fgithub.com\u002Fgethamster\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-vision-and-values-statements\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M7B5MKH2CH22YZAHQW3080VQ","writing-system-prompts-for-claude","Write Claude prompts and CLAUDE.md files that supply exact inputs, scope the task, and define done so the first attempt can be checked.","https:\u002F\u002Fgithub.com\u002Fgethamster\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-system-prompts-for-claude\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7B5ME1Y3X9RNRKP5HKHKACZ","writing-page-specific-website-copy","Website copywriting by page type: write homepage, landing page, pricing page and feature page copy that fits each page's visitor and single job.","https:\u002F\u002Fgithub.com\u002Fgethamster\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-page-specific-website-copy\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7B5M8MQNM8MN8PYH6BVWJ2B","writing-internal-press-releases","How to write an internal press release: the one-page working backwards press release that opens a PR\u002FFAQ document before anything is built.","https:\u002F\u002Fgithub.com\u002Fgethamster\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-internal-press-releases\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":19},"01M7B5M3P3YH1XF37R8PWXXSBK","writing-email-sequences-that-sell","Email sequence copywriting for nurture and sales: give each email one job, write clear subject lines, and close with one specific call to action.","https:\u002F\u002Fgithub.com\u002Fgethamster\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-email-sequences-that-sell\u002FSKILL.md",[],[]]