Skill: twitter-card

Produces a pixel-exact HTML canvas for Twitter/X. Two canvas modes:

• feed (default, 1600×900) — in-timeline post image at the recommended max-quality 16:9 ratio
• og (1200×630) — Open Graph link card for shared URLs from your site

Plus an optional companion tweet (280-char limit) or thread starter.

When to use

• User wants a visual asset for a tweet, an X thread cover, or an Open Graph card on their site
• A DESIGN.md exists for the brand

Inputs

• Required: brand slug, post headline / title (1 line)
• Optional: mode (feed | og, default feed), supporting line, attribution, companion tweet style

Output

./design/<brand-slug>/artifacts/twitter-<mode>-YYYY-MM-DD-<slug>.html

Steps

1. Verify brand exists

test -f ./design/<brand-slug>/tokens.css

If missing, stop and instruct the user to build the design system first.

2. Pick the canvas mode

If user said "OG card" or "link card", use og (1200×630). Otherwise default to feed (1600×900) — sharper, larger, displays better in-timeline.

3. Gather the brief

Ask in one message:

1. Headline — the dominant text on the canvas (max 10 words for feed mode, max 6 for OG)
2. Optional supporting line
3. Attribution — name + role, or leave blank
4. Companion tweet style: claim | thread-starter | quote | question | none

4. Pick variation — ARCHITECTURE FIRST

Pick ONE architecture archetype before any other axis. The single biggest cause of AI-editorial output is reaching for chrome-led by default.

• Architecture archetype (pick FIRST):
  • chrome-led — eyebrow + headline + footer band. Last resort. AI-editorial default; structural sameness across artifacts.
  • type-only — type only, no chrome.
  • number-led — single oversized number/stat dominates the canvas; rest is caption.
  • object-of-content — the artifact IS the thing being said (transcript, list, fake printed page, receipt).
  • pattern-led — typographic repetition fills the canvas; one element breaks the pattern.
  • inverse-text — text becomes surface (carved-out headline, body wrapping negative space).

After archetype, pick ONE per axis:

• Layout: symmetric | asymmetric | off-grid | full-bleed | split
• Type pressure: display-dominant | body-dominant | mono-led | mixed-weight
• Color usage: monochrome | single-accent | duotone | inverted
• Composition focus: headline-led | quote-led | mark-led | data-led

Cross-artifact rule: read variation comments in the most recent 2–3 artifacts under ./design/<brand-slug>/artifacts/. Different archetype than recent. Different axis combos.

5. Generate the canvas HTML

Template (swap dimensions based on mode):

<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<title><Brand> — Twitter — <topic></title>
<link rel="preconnect" href="https://fonts.googleapis.com">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link rel="stylesheet" href="https://fonts.googleapis.com/css2?family=<families>&display=swap">
<style>
/* Embed tokens.css verbatim */
<contents of tokens.css>

html, body { margin: 0; padding: 0; background: #2a2a2a; min-height: 100vh; display: grid; place-items: center; font-family: var(--type-body-md-family); }

/* Canvas — switch dimensions by mode */
.canvas {
  /* feed mode: 1600x900   |   og mode: 1200x630 */
  width: <1600 | 1200>px;
  height: <900 | 630>px;
  background: var(--color-surface);
  color: var(--color-ink);
  position: relative;
  overflow: hidden;
  padding: var(--space-3xl);
  box-sizing: border-box;
  display: flex;
  flex-direction: column;
  justify-content: <by variation>;
}

.eyebrow {
  font: var(--type-label-caps-weight) var(--type-label-caps-size)/var(--type-label-caps-leading) var(--type-label-caps-family);
  letter-spacing: var(--type-label-caps-tracking);
  text-transform: uppercase;
  color: var(--color-ink-soft);
}

.headline {
  font: var(--type-display-xl-weight) var(--type-display-xl-size)/var(--type-display-xl-leading) var(--type-display-xl-family);
  letter-spacing: var(--type-display-xl-tracking);
  margin: var(--space-md) 0 0;
  /* feed mode allows larger; OG mode caps lower for safety */
  font-size: clamp(48px, 6vw, 96px);
}

.footer {
  display: flex;
  justify-content: space-between;
  align-items: baseline;
  border-top: 1px solid var(--color-rule);
  padding-top: var(--space-md);
  font: var(--type-label-caps-weight) var(--type-label-caps-size)/var(--type-label-caps-leading) var(--type-label-caps-family);
  letter-spacing: var(--type-label-caps-tracking);
  text-transform: uppercase;
  color: var(--color-ink-soft);
}

@page { size: <1600 | 1200>px <900 | 630>px; margin: 0; }
</style>
</head>
<body>
<!--
Mode: <feed | og>
Variation choices:
  layout: <picked>
  type-pressure: <picked>
  color: <picked>
  composition: <picked>
-->
<article class="canvas">
  <header>
    <span class="eyebrow"><eyebrow text></span>
    <h1 class="headline"><headline></h1>
  </header>
  <footer class="footer">
    <span><Brand mark or @handle></span>
    <span><YYYY.MM.DD></span>
    <span><optional URL or topic tag></span>
  </footer>
</article>
</body>
</html>

Layout adaptations (same options as linkedin-post):

• off-grid: anchor headline at the optical center (slightly above geometric center), shift left ~12% of canvas width
• split: divide vertically into two unequal panels (e.g. 60/40), headline in larger, supporting in smaller
• full-bleed: extend background color or single-accent block beyond default padding to canvas edge on one side
• mark-led: oversized typographic brand mark dominates upper region, headline reads as caption beneath

6. Generate companion tweet

Twitter/X rules:

• 280 character limit (count carefully; URL counts as 23)
• First 70 chars are the preview when quote-tweeted — front-load the claim
• Style by user choice:
  • claim: declarative single-line punchline
  • thread-starter: hook + "🧵 below" or "(1/X)" — but since no emojis by default, use "→" or numbering
  • quote: pull quote from the artifact + attribution
  • question: provocative ask + context

Append in an HTML comment at end of file:

<!--
COMPANION TWEET (<count>/280 chars):

<tweet text>

(Optional thread continuation in additional posts.)
-->

7. Verify and report

• Canvas is exactly the chosen mode's pixel dimensions
• Architecture archetype documented AND differs from the most recent artifact in this brand
• Read ../../design-anti-patterns.md and verify the artifact violates none of its rules. Pay special attention to sections 2 (no italics-emphasis OR single-word color emphasis), 4 (architecture must reach inner blocks; vary footer; no top-metadata-row disguised as eyebrow), 5 (no editorial-cosplay markers), and 7 (cross-artifact rules)
• Headline legible at preview thumbnail (~500px) for feed; ~400px for OG
• Companion tweet under 280 chars

Twitter artifact (<mode> mode) at ./design/<brand-slug>/artifacts/twitter-<mode>-YYYY-MM-DD-<slug>.html.

Variation: <layout> × <type-pressure> × <color> × <composition>.

Export as PNG: open in browser, DevTools → device toolbar → match canvas dims → screenshot. Or Cmd+P → save as PDF.

Companion tweet at the end of the file in an HTML comment.

Anti-Slop Rules

Governed by three shared canonical references — read them, do not restate them:

• ../../design-anti-patterns.md — the hard floor. Wins every conflict.
• ../../design-principles.md — the craft floor (typographic hierarchy, spatial rhythm, color theory, layout logic, visual tension).
• ../../design-variation-sop.md — the named-aesthetic roster + the procedure for varying every invocation.

Forbidden (fast scan — the references hold the full list):

• Indigo/violet/purple as default accent or primary — the hsl(230–280) family on a near-white surface. The single loudest AI tell.
• Purple→blue / "aurora" / gradient-mesh backgrounds; gradient-fill text.
• Glassmorphism without an explicit AA-contrast scrim; floating gradient "blobs" as atmosphere.
• The three-up icon-card feature grid; the "hero → 3 cards → CTA" median skeleton.
• One global border-radius on every element; timid evenly-weighted low-contrast palette.
• Inter / Roboto / Arial / Open Sans / Lato / Helvetica / system as the primary family (Space Grotesk = yellow flag). Use a category-matched stack from the variation roster.
• Value-free CTAs ("Get Started", "Learn More", "Sign Up"); the two-CTA hero. Name the real action and its value.

Required variation (every invocation):

• Make at least two intentional decisions that differ from the safe defaults (palette, layout structure, typographic voice, or spatial density) AND from the most recent artifact in this brand.
• Never default to "clean / minimal / modern" — that is the absence of a direction (variation-sop Rule 1).
• Sketch three distinct directions, offer them as a one-line menu, generate the best-fit by default, all three only if asked (variation-sop Rule 3).

Authenticity: prefer specific over generic everywhere — real datelines (not "today"), the brand's actual voice (not "build the future" filler), concrete CTAs over placeholders.

Rules

• Pixel exactness. No rounding, no scaling.
• Token-pure. No literal hex codes or font sizes outside clamp() calls.
• No center-aligned safe layout unless the brand's plan explicitly calls for symmetry.
• Mode default is feed (1600×900), not OG. Most users want in-timeline posts.
• Aliases: this skill answers to both twitter-card and x-card. Trigger phrases work either way.
• Anti-pattern compliance. This skill defers to three shared canonical references: ../../design-anti-patterns.md (the hard floor — wins every conflict), ../../design-principles.md (the craft floor), and ../../design-variation-sop.md (direction roster + offer-3 procedure). The anti-patterns file is the canonical anti-tell list. If you find a new tell in the output, add it there.