Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A pop-culture retrospective on how 1980s design language shaped today's apps — the scenes, the turning point, and the takeaway. Built as a decision-grade story deck for talk audience, design community.
.claude/skills/nexu-io-hps-memphis-pop/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 67% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 48% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 42% | 0% |
A locked single-theme deck plugin: the memphis-pop skin from the upstream MIT-licensed lewislulu/html-ppt-skill (36 themes × 31 layouts), shipped as one self-contained seed — example.html — that carries the full token sheet, the Memphis decor devices, the slide scaffold, and the keyboard runtime.
Start from example.html. Replace content only. Never rewrite the design system or the runtime script. Never introduce colors or fonts outside this spec.
Positioning vs. sibling Memphis-adjacent plugins: this is the full-bleed scattered confetti-dot pattern Memphis. daisy-days is hand-drawn daisy childlike (flower illustration devices), block-frame is big-color-block neobrutalist (no pattern). The three decor systems are mutually exclusive — do not mix.
Every variable below is defined in the seed's :root. Slides reference tokens only — never hard-code a color in slide markup.
css:root{ --bg:#fef6e8; /* warm cream canvas */ --bg-soft:#fdebc7; --surface:#ffffff; /* card face */ --surface-2:#fff1d1; --border:#111111; /* everything outlines in near-black */ --border-strong:#111111; --text-1:#111111; --text-2:#333333; --text-3:#666666; --accent:#ff3d8b; /* hot pink — the lead */ --accent-2:#37c2d7; /* teal */ --accent-3:#ffcc00; /* yellow */ --good:#6ac04c; --warn:#ffcc00; --bad:#ff3d8b; --grad:linear-gradient(135deg,#ff3d8b,#ffcc00 50%,#37c2d7); --grad-soft:linear-gradient(135deg,#fdebc7,#fff1d1); --radius:10px; --radius-sm:6px; --radius-lg:18px; --shadow:5px 5px 0 #111; /* hard offset, never blurred */ --shadow-lg:9px 9px 0 #111; --font-sans:'Space Grotesk','Inter','Noto Sans SC',sans-serif; --font-serif:'Space Grotesk','Noto Sans SC',sans-serif; --font-mono:'JetBrains Mono',SFMono-Regular,Menlo,monospace; --font-display:'Archivo Black','Space Grotesk',sans-serif; --letter-tight:-.01em; --letter-normal:0; --ease:cubic-bezier(.4,0,.2,1); }
Fonts come from Google Fonts @import only (Archivo Black, Space Grotesk, Noto Sans SC, JetBrains Mono). No other external resource of any kind.
body background tiles three offsetradial-gradient layers of 3px dots (pink / teal / yellow at 200 / 220 / 260px tile sizes) over --bg. It is full-bleed on every slide; slides stay transparent so the pattern always shows through. Never paint an opaque slide background.
.card is 2.5px solid #111 with--shadow (hard offset, zero blur). Hero cards may use --shadow-lg.
circles, plus signs and half-tone arcs, each stroked in #111 and filled with one accent, scattered absolutely behind/around content (.memphis-shape). 2–4 per slide maximum.
.pill carries a 2px black border, an accentfill and a slight rotate(±2deg); used as tags and badges.
.divider-accent is a hot-pink bar witha black outline; section dividers pair it with a giant outlined numeral.
.gradient-text (the tri-color--grad) is reserved for big numbers and 1–2 highlight words.
The upstream catalog is shared across the html-ppt family; compose pages from these master categories:
| group | layouts | |---|---| | Openers & transitions | cover · toc · section-divider | | Text-centric | bullets · two-column · three-column · big-quote | | Numbers & data | stat-highlight · kpi-grid · table · chart-bar/line/pie/radar | | Code & terminal | code · diff · terminal | | Diagrams & flows | flow-diagram · arch-diagram · process-steps · mindmap | | Plans & comparisons | timeline · roadmap · gantt · comparison · pros-cons · todo-checklist | | Visuals | image-hero · image-grid | | Closers | cta · thanks |
Default sequence: cover → toc → (section-divider → 2–4 content pages) × N → thanks. Charts are always pure CSS or inline SVG — never Chart.js or any external library. The seed demonstrates: cover, toc, section-divider, two-column, kpi-grid, stat-highlight, chart-bar, process-steps, big-quote, thanks.
<section class="slide" data-title="..."> inside<div class="deck" id="deck"> — a horizontal scroll-snap strip, each slide exactly 100vw × 100vh (one screen, no internal scrolling), 16:9 / 1280×720 baseline with clamp() type scales, padding 72px 96px.
.deck-header (deck title + theme chip), .deck-footer(attribution + N / total counter), .progress-bar.
← → Space PageUp PageDown Home End navigate.#/N (1-based) deep-links a slide; navigation syncs thehash via history.replaceState wrapped in try/catch (srcdoc-safe).
Event identity and auto-focuses <body> so keys work without a click — these solve real iframe-host bugs; do not "simplify" them away.
<div class="notes">…</div> per slide.example.html; keep all <style> blocks and the <script>verbatim.
content, real numbers — no lorem ipsum, no placeholder images.
2–4 geometric SVG shapes per slide.
near-black #111 outlines, white card faces — nothing else.
.notes.#/5 deep-links, no slide overflowsvertically, every shadow is a hard offset (no blur anywhere).
Visual system, token vocabulary, theme palette, and layout taxonomy come from the upstream MIT-licensed lewislulu/html-ppt-skill (© lewis <sudolewis@gmail.com>), theme memphis-pop. The LICENSE file ships alongside this skill — keep it in place when redistributing.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-08 | fail→pass | 28,976 | 51,521 | +78% | 1 | 1 | 0% | 6,104 | 10,192 | +67% | 0 | 0 | — |
case-01 | fail→fail | 52,551 | 56,996 | +8% | 1 | 1 | 0% | 8,262 | 10,212 | +24% | 0 | 0 | — |
case-02 | fail→pass | 43,702 | 41,198 | -6% | 1 | 1 | 0% | 8,254 | 10,204 | +24% | 0 | 0 | — |
case-03 | fail→pass | 53,188 | 36,305 | -32% | 1 | 1 | 0% | 8,250 | 10,200 | +24% | 0 | 0 | — |
case-04 | pass→fail | 40,400 | 36,492 | -10% | 1 | 1 | 0% | 8,215 | 10,165 | +24% | 0 | 0 | — |
case-05 | pass→fail | 54,461 | 38,409 | -29% | 1 | 1 | 0% | 8,218 | 10,168 | +24% | 0 | 0 | — |
case-06 | pass→fail | 52,249 | 42,119 | -19% | 1 | 1 | 0% | 8,218 | 10,168 | +24% | 0 | 0 | — |
case-07 | fail→pass | 40,324 | 36,188 | -10% | 1 | 1 | 0% | 6,892 | 10,178 | +48% | 0 | 0 | — |
case-09 | fail→fail | 39,229 | 38,848 | -1% | 1 | 1 | 0% | 8,228 | 10,178 | +24% | 0 | 0 | — |
case-10 | fail→fail | 39,052 | 50,586 | +30% | 1 | 1 | 0% | 8,222 | 10,172 | +24% | 0 | 0 | — |
case-11 | fail→fail | 40,381 | 48,521 | +20% | 1 | 1 | 0% | 6,168 | 9,834 | +59% | 0 | 0 | — |
case-12 | pass→fail | 53,653 | 42,849 | -20% | 1 | 1 | 0% | 8,225 | 8,980 | +9% | 0 | 0 | — |
case-13 | fail→fail | 54,120 | 37,675 | -30% | 1 | 1 | 0% | 8,219 | 10,169 | +24% | 0 | 0 | — |
case-14 | fail→pass | 33,156 | 48,911 | +48% | 1 | 1 | 0% | 7,190 | 10,179 | +42% | 0 | 0 | — |
case-15 | fail→pass | 29,243 | 35,404 | +21% | 1 | 1 | 0% | 5,981 | 9,666 | +62% | 0 | 0 | — |
case-16 | fail→fail | 41,475 | 37,153 | -10% | 1 | 1 | 0% | 8,213 | 10,163 | +24% | 0 | 0 | — |
case-17 | fail→fail | 39,535 | 83,497 | +111% | 1 | 1 | 0% | 8,216 | 10,166 | +24% | 0 | 0 | — |
case-18 | fail→pass | 27,642 | 51,350 | +86% | 1 | 1 | 0% | 5,766 | 10,162 | +76% | 0 | 0 | — |
case-19 | fail→fail | 40,643 | 31,916 | -21% | 1 | 1 | 0% | 6,945 | 8,789 | +27% | 0 | 0 | — |
case-20 | fail→pass | 45,604 | 42,934 | -6% | 1 | 1 | 0% | 8,208 | 8,898 | +8% | 0 | 0 | — |
case-21 | fail→fail | 40,215 | 51,417 | +28% | 1 | 1 | 0% | 8,201 | 10,151 | +24% | 0 | 0 | — |
case-22 | fail→pass | 40,087 | 35,112 | -12% | 1 | 1 | 0% | 8,202 | 10,162 | +24% | 0 | 0 | — |
case-23 | fail→fail | 22,188 | 7,751 | -65% | 1 | 1 | 0% | 3,216 | 3,279 | +2% | 0 | 0 | — |
case-24 | pass→pass | 17,513 | 35,471 | +103% | 1 | 1 | 0% | 2,531 | 10,154 | +301% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted. The headline lift of +21 percentage points is the difference between those two pass rates over the 24 comparable cases. 4 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.