Files
open-design/skills/open-design-landing/README.md
T
Zakaria a46764fb1b
ci / Validate workspace (push) Has been cancelled
landing-page-ci / Validate landing page (push) Has been cancelled
landing-page-deploy / Deploy landing page (push) Has been cancelled
github-metrics / Generate repository metrics SVG (push) Has been cancelled
refresh-contributors-wall / Refresh contributors wall cache bust (push) Waiting to run
first-commit
2026-05-04 14:58:14 -04:00

4.9 KiB
Raw Blame History

open-design-landing

Reusable skill that produces a world-class editorial landing page in the Atelier Zero design language — the warm-paper, italic-serif, collage-on-grid aesthetic shared by Monocle, Apartamento, and Études.

The skill is parameterized: fill one typed inputs.json, run one script, get a self-contained HTML file. Optionally generate 16 surreal collage assets with gpt-image-2, or fall back to paper-textured SVG placeholders so the layout still feels intentional with zero image budget.

Read first — the agent contract, inputs schema, and self-check live in SKILL.md. This README is the human quick-start.

30-second tour

# 1. Paper-textured placeholders so the layout renders immediately.
npx tsx scripts/placeholder.ts ./out/assets/

# 2. Compose the standalone HTML from the worked example.
npx tsx scripts/compose.ts inputs.example.json ./out/index.html

# 3. Open it.
open ./out/index.html

That's it. Three commands, full editorial landing page, no API keys.

To brand it for yourself, copy inputs.example.json to inputs.json, edit the fields (the schema is self-documenting — see schema.ts), and re-run step 2.

The three image strategies

Strategy Cost Latency When
placeholder $0 <1s First pass, demo, internal review.
generate ~$0.40 ~6 min Final delivery; original collage plates.
bring-your-own $0 0s You have art direction PNGs ready to drop in.

Set inputs.imagery.strategy to one of the three.

# generate mode (requires FAL_KEY in env)
FAL_KEY=fal-... npx tsx scripts/imagegen.ts inputs.json --out=./out/assets/

Without FAL_KEY, the imagegen script prints the prompts so you can route them through the /gpt-image-fal slash-command skill manually.

Layout at a glance

8 numbered Roman-numeral sections, all responsive at 1280 / 1080 / 880 / 560:

I.   Hero          — display headline + 3 stat rings + 4-step index + collage plate
II.  About         — manifesto + studio stamp + tilted side-note
III. Capabilities  — 4 cards (skills / systems / adapters / BYOK) + ribbon
IV.  Labs          — 5 portrait cards + filter pills + progress bar
V.   Method        — 4 numbered steps with thumbnails on hairline timeline
VI.  Selected work — dark slab, 2 tilted cards (one rotated -1.2°, one +2.4°)
VII. Testimonial   — pull quote + 5 partner glyphs
VIII. CTA          — closing pitch + ribbon + email pill
     Footer       — 4 link columns + huge italic-serif kicker word

Every section has scroll-reveal motion (IntersectionObserver, respects prefers-reduced-motion).

Files

skills/open-design-landing/
├── SKILL.md                 # ← agent contract (read this first)
├── README.md                # ← you are here
├── schema.ts                # typed inputs (single source of truth)
├── styles.css               # Atelier Zero stylesheet (single source of truth)
├── inputs.example.json      # Open Design as the worked example
├── example.html             # canonical rendering, regenerable from inputs.example.json
├── scripts/
│   ├── compose.ts           # inputs.json + styles.css → index.html
│   ├── imagegen.ts          # gpt-image-2 wrapper (fal.ai backend)
│   └── placeholder.ts       # SVG paper-textured frames
└── assets/
    ├── *.png                # 16 collage plates (Open Design instance)
    ├── image-manifest.json  # slot → file / dimensions / prompt mapping
    └── imagegen-prompts.md  # human-readable prompt pack

Regenerate the canonical example

After editing styles.css, schema.ts, or inputs.example.json:

npx tsx scripts/compose.ts inputs.example.json example.html

The example.html in this folder is the pre-rendered known-good demo — useful as a visual reference and for QA against the live composer output.

Migrating from editorial-collage

This skill replaces the older editorial-collage folder:

See also