Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
geekjourneyx avatar

Broadsheet Card

  • 1 installs
  • 1 repo stars
  • Updated June 28, 2026
  • geekjourneyx/broadsheet-card

Creates validated six-page vintage-broadsheet editorial image cards from a topic or notes, exporting HTML and PNGs for social carousels and article images.

About

A skill that treats card creation as publishing, editing content into a CardPlan and ColorPlan then typesetting six broadsheet-style pages and rendering validated PNGs. A developer or creator uses it for WeChat article images, Xiaohongshu/Rednote carousels, or social graphics.

  • CardPlan, three cover candidates, and ColorPlan required before HTML
  • Validation scripts for card plan, layout, and contrast plus PNG rendering

Broadsheet Card by the numbers

  • 1 all-time installs (skills.sh)
  • Ranked #1,200 of 1,335 Generative Media skills by installs in the Skillselion catalog
  • Data as of Jul 24, 2026 (Skillselion catalog sync)
npx skills add https://github.com/geekjourneyx/broadsheet-card --skill broadsheet-card

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs1
repo stars1
Last updatedJune 28, 2026
Repositorygeekjourneyx/broadsheet-card

What it does

Creates validated six-page vintage-broadsheet editorial image cards from a topic or notes, exporting HTML and PNGs for social carousels and article images.

Files

SKILL.mdMarkdownGitHub ↗

Broadsheet Card

Create a six-page Vintage Broadsheet editorial card set. Treat the job as publishing, not decoration: edit the content first, then typeset it, render PNGs, and validate the output. The cards can be used as WeChat article images, Xiaohongshu/Rednote carousels, social graphics, or standalone publication-style visuals.

Default Output

  • Canvas: 1080 x 1440 px
  • Ratio: 3:4
  • Pages: 6 unless the user explicitly asks otherwise
  • Format: HTML plus PNG exports
  • Files:
  • index.html
  • manifest.json
  • output/page-01-cover.png through output/page-06-checklist.png

Required Workflow

1. Parse the user input into a ContentBrief. 2. Read references/copywriting.md and create a CardPlan. 3. Read references/cover-system.md; generate three cover candidates and select one. 4. Read references/color.md; create a ColorPlan. 5. 🔴 CHECKPOINT: Do not write HTML until CardPlan, cover selection, and ColorPlan are coherent. 6. Read references/vintage-broadsheet-design.md. 7. Read references/layout-recipes.md. 8. Read references/html-contract.md. 9. Generate HTML using templates/base.html and templates/components.css as the structural model. 10. Run scripts/validate-card-plan.js. 11. Run scripts/validate-layout.js. 12. Run scripts/validate-contrast.js. 13. Fix failures by rewriting content or layout. Do not shrink text below the limits. 14. Run scripts/render.js to export PNGs. 15. Run scripts/make-thumbnail-preview.js when image tooling is available. 16. Review against references/quality-rubric.md. 17. Return the generated file list and a short note on the editorial structure.

Hard Rules

  • Never generate HTML directly from the topic. Create a CardPlan first.
  • Generate three cover candidates before choosing the final cover.
  • Create a ColorPlan and derive CSS tokens from it.
  • Each page must carry one main conclusion.
  • Cover headline: 8-18 Chinese characters preferred.
  • Body text: 28px or larger.
  • Critical text must stay inside the 72px safe area.
  • Red accent appears at most twice per page.
  • Important text must not sit on noisy texture or image backgrounds.
  • If content is too long, summarize it. Never shrink the font to fit.
  • Avoid external fonts, CSS, scripts, or images unless the user provides trusted assets.
  • If the user asks for multiple visual styles in one card set, narrow to Vintage Broadsheet or ask them to choose.

Default Page Sequence

1. Cover: strong hook 2. Problem: the mistake or tension 3. Framework: the core decision model 4. Detail: the most important mechanism 5. Comparison: wrong way vs right way 6. Checklist: save-worthy summary

When to Refuse or Ask for More Input

Ask for more content if the user provides only a vague theme and expects factual claims. Refuse or redirect requests that require sacrificing readability for heavier vintage effects, mixing many visual styles, or cramming a long article into one image.

Failure Handling

TriggerFirst repairIf still failing
Source content is vague or factual claims are unsupportedAsk for source text, audience, and intended useProduce only a planning brief, not final cards
Topic is medical, legal, financial, or otherwise high-stakesAsk for sources and narrow to educational framingRefuse specific advice; offer a neutral explainer structure
CardPlan does not fit six pagesRewrite page claims and shorten body copyAsk user to split the topic into multiple card sets
Layout or contrast validation failsRewrite HTML/CSS using the same design tokensReduce content density; never reduce body text below 28px
Playwright render failsReport the exact command and errorDeliver validated HTML plus next render command

Anti-Patterns

Do not create:

  • platform-locked output when the user asked for general article images or social cards
  • vintage decoration without CardPlan, ColorPlan, and validation
  • yellow parchment, scrapbook, wanted-poster, or generic newspaper-template visuals
  • long article screenshots, tiny body text, or dense filler columns
  • financial, medical, or legal recommendation cards without sources and careful framing

Related skills

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.