Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Design — URL To Code · dkeken | DecimalAI

frontendby dkeken

Design — URL To Code

dkeken/design-url-to-code

Source dkeken/codex-skills-alternative ↗License MIT

Turn a live website into a working local frontend you can edit. Fetches the page, extracts its design system and structure, and rebuilds it locally in your stack, then screenshot-compares to the original.

SkillScore64Maintained

Automatic

Frontend / UI pack

Last verified 8/6/2026

What it changes

Passed 12 of 19 test cases with the skill, and 10 of 19 without it.

Without skill55%

With skill64%

Checks it passed that the model alone missed

Selects React, Vite, Tailwind CSS, and TypeScript when no stack is specified
Rebuilds clean custom components instead of copy-pasting scraped HTML DOM dumps
Updates font size tokens in theme configuration when screenshot comparison reveals a discrepancy

3 fixed11 passed either way7 still failing1 broke

Pass-rate lift: +9 pp
Tokens: +2%
Agent turns: 0%
Response time: -36%

Model gemini-3.6-flashMeasured 8/6/2026v114 of 22 cases passed with skill · 19 comparable

Tokens, turns and time describe resource use, not a dollar price.

Use this skill

Save it into your project's skills folder. Claude Code picks it up in its next session.

curl -fsSL --create-dirs -o .claude/skills/dkeken-design-url-to-code/SKILL.md https://app.decimal.ai/s/dkeken-design-url-to-code/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add dkeken-design-url-to-code, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"

1 · One-time — install the CLI (docs ↗)

pip install "decimalai>=0.15"

2 · Get the skill — writes .agents/skills/dkeken-design-url-to-code/ and pins it in .decimal/skills.lock

decimalai skills add dkeken-design-url-to-code

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/dkeken-design-url-to-code/SKILL.md https://app.decimal.ai/s/dkeken-design-url-to-code/SKILL.md

Raw file: /s/dkeken-design-url-to-code/SKILL.md ↗ · machine-readable summary at /s/dkeken-design-url-to-code.json ↗

Safety checks: 3 of 3 passed

Code scan CleanIntent PassedContent Clean

See it in action

One of 3 test cases the skill fixed. Same model, same prompt, with and without the skill.

1 of 3

Prompt

In our screenshot diff between our local server and https://render.com, the main title font size is 48px locally but measured 56px on the live site. What action is required?

What was checked, without and with the skill
Checked	Without skill	With skill
Updates font size tokens in theme configuration when screenshot comparison reveals a discrepancy	Missed	Passed

case-19 · gemini-3.6-flash · 8/6/2026 · answers are capped and redacted

What's inside

The instructions your model reads when the skill loads.

SKILL.md61 linessingle file

Design — URL To Code

Recreate a live site/page as an editable local frontend — faithful to the original and verified by comparison. Unlike image-to-code, you have the real HTML/CSS to extract from, so fidelity can be higher; the work is rebuilding it clean in your stack rather than copy-pasting markup.

When to use

"Clone/rebuild this live site locally", "make it look like {url}".
You want a real, editable starting point matching an existing site's design.
Capturing a design language from a site you admire (legitimately — see rules).

When NOT to use

The source is a static image/mockup, not a live URL → design-image-to-code.
You need interactivity/flow wiring after → design-prototype.
The goal is critique, not recreation → design-audit.

Inputs

Source URL(s).
Target stack: ask; default match repo, else React + Vite + Tailwind + TS.
Scope: full clone vs "capture the design language only".

Workflow

Capture the page: open it in a browser-automation tool (DOM snapshot + full-page screenshot) and/or pull HTML/CSS via a web-fetch tool. The browser gives you computed styles; the screenshot is your comparison target.
Extract the design system into ./design/url-dna.md: colors, type scale, spacing rhythm, layout structure, components, motion. Capture tokens, not pixel-by-pixel copies.
Rebuild locally in the target stack with your own clean components — not pasted markup. Map extracted tokens → real config (e.g. Tailwind theme extension).
Verify: run locally, screenshot at the source's viewport, compare to the original; iterate to close gaps (layout → type → spacing → color).
Assets: download/extract where permitted; generate missing raster with an image tool; placeholder otherwise.

Worked example

source: https://example.com/landing   stack: React+Vite+Tailwind+TS
capture: browser snapshot (computed CSS) + full-page screenshot
url-dna.md:
  palette: bg #0B0B0F, text #E7E7EA, accent #6366F1
  type: Inter; scale 14/16/20/32/56; tight tracking on display
  spacing: 4px base, section pad 96px
  layout: sticky nav, hero (h1 + sub + 2 CTAs), 3-col feature grid, footer
  motion: fade-up on scroll
rebuild: Tailwind theme = those tokens; components Nav/Hero/FeatureGrid/Footer
verify: screenshot @1440 vs original → hero h1 was 48 not 56, fixed; gap close
done: build clean, tokens in tailwind.config, runs at :5173

Track record

Benchmark

12 / 19

test cases passed, vs 10 / 19 without

Adoption

Users: 0
Times used: 0

Safety and source details

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/2/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

dkeken / codex-skills-alternative

skills/design-url-to-code/SKILL.md

View on GitHub ↗

License MITImported 8/2/2026Version v1

More skills from the same source and nearby in the registry, with their headline benchmark lift.

Ready to supercharge your AI agents?

See it in action

What's inside

Design — URL To Code

When to use

When NOT to use

Inputs

Workflow

Worked example

Track record

Benchmark

Adoption

SkillSafety

Source

Related skills