Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Redesign an existing site or app without losing what already works — audits the current surface first, classifies what to preserve (brand, IA, SEO, content), picks a refresh/reskin/rebuild scope, then modernizes deliberately. Use when the user says "redesign", "modernize this site", "refresh the UI", "make this look current", or points at an existing page/URL they want improved rather than rebuilt from scratch. Invoke when the user asks for redesign on their UI, or mentions 'redesign' alongside
.claude/skills/educlopez-redesign/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 5748% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 16% | 0% |
<!-- HARNESS MIRROR — do not edit here. Canonical source: commands/redesign.md. After editing source, copy into cli/assets/<harness>/ and repo-root harness mirrors. -->
Context: this sub-skill is one lens of the broader ui-craft skill. If the ui-craft skill is also installed, read its SKILL.md first for Discovery + Anti-Slop + Craft Test, then apply the specific lens below.
Redesign $ARGUMENTS. Load the ui-craft skill and read references/craft-intent.md. A redesign is not a greenfield build wearing the old content — most of its value is in what you deliberately keep.
Step 1 — Audit what exists (before any opinion):
npx ui-craft-detect <path>) or live page (npx ui-craft-detect <url>) and note findings..ui-craft/brief.md if present — prior constraints still bind.Step 2 — Classify what to preserve. Wrong-to-change unless the user explicitly asks:
Step 3 — Pick the scope and say it out loud:
| Scope | What changes | When | |---|---|---| | Refresh | Tokens only: type scale, spacing rhythm, radii, shadows, color tuning | Site is structurally fine, looks dated | | Reskin | Visual layer + component styling; layout skeleton and section order stay | Bones are good, skin is generic or old | | Rebuild | Layout and composition too; content, IA, and brand carry over | Layout itself is the problem |
Default to the smallest scope that fixes the stated complaint. For rebuild, print a short before/after section map and get a nod before writing code.
Step 4 — Declare the Craft Read (surface kind, audience, theme, DESIGN_VARIANCE, signature bet) per references/craft-intent.md. Redesigns default to variance 4–5: the brand is already committed, so spend boldness on typography and composition, not on a new identity.
Step 5 — Modernize with levers, in this order: typography (scale + pairing per references/typography.md) → spacing rhythm → color tokens (references/color.md) → imagery treatment → motion (entrance/hover only where focal). Fix every detector finding from Step 1 along the way.
Never: invent a new brand palette over a committed one, drop routes or pages, flatten heading semantics for looks, rewrite copy voice unasked, or ship the redesign with more detector findings than the original had.
Output: edit code directly (refresh/reskin) or section map first (rebuild). Print the Review Format table, the detector before/after counts, and end with the Craft Read line.
Next step: /critique — check the modernized surface against hierarchy and anti-slop before shipping (rung 1).
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-08 | fail→pass | 6,046 | 2,584 | -57% | 1 | 1 | 0% | 1,016 | 1,259 | +24% | 0 | 0 | — |
case-01 | fail→fail | 28,623 | 6,710 | -77% | 1 | 1 | 0% | 6,215 | 1,233 | -80% | 0 | 0 | — |
case-02 | fail→fail | 3,064 | 4,424 | +44% | 1 | 1 | 0% | 145 | 1,049 | +623% | 0 | 0 | — |
case-03 | fail→fail | 3,591 | 5,039 | +40% | 1 | 1 | 0% | 179 | 1,151 | +543% | 0 | 0 | — |
case-09 | fail→pass | 12,031 | 10,272 | -15% | 1 | 1 | 0% | 1,874 | 2,407 | +28% | 0 | 0 | — |
case-04 | fail→pass | 2,826 | 37,055 | +1211% | 1 | 1 | 0% | 120 | 7,018 | +5748% | 0 | 0 | — |
case-05 | pass→fail | 11,319 | 3,634 | -68% | 1 | 1 | 0% | 1,700 | 1,454 | -14% | 0 | 0 | — |
case-06 | pass→pass | 12,917 | 14,226 | +10% | 1 | 1 | 0% | 2,585 | 3,517 | +36% | 0 | 0 | — |
case-07 | fail→pass | 8,833 | 2,565 | -71% | 1 | 1 | 0% | 1,483 | 1,268 | -14% | 0 | 0 | — |
case-10 | pass→pass | 9,778 | 5,423 | -45% | 1 | 1 | 0% | 1,760 | 1,737 | -1% | 0 | 0 | — |
case-11 | pass→pass | 8,898 | 3,333 | -63% | 1 | 1 | 0% | 1,452 | 1,393 | -4% | 0 | 0 | — |
case-12 | fail→pass | 6,296 | 5,189 | -18% | 1 | 1 | 0% | 1,442 | 1,666 | +16% | 0 | 0 | — |
case-13 | fail→pass | 8,053 | 2,920 | -64% | 1 | 1 | 0% | 1,322 | 1,314 | -1% | 0 | 0 | — |
case-14 | fail→pass | 4,100 | 2,396 | -42% | 1 | 1 | 0% | 628 | 1,255 | +100% | 0 | 0 | — |
case-15 | fail→pass | 5,538 | 2,670 | -52% | 1 | 1 | 0% | 920 | 1,402 | +52% | 0 | 0 | — |
case-16 | fail→pass | 8,756 | 1,806 | -79% | 1 | 1 | 0% | 1,539 | 1,138 | -26% | 0 | 0 | — |
case-17 | fail→pass | 8,095 | 2,948 | -64% | 1 | 1 | 0% | 1,454 | 1,387 | -5% | 0 | 0 | — |
case-18 | fail→pass | 9,455 | 2,719 | -71% | 1 | 1 | 0% | 1,705 | 1,245 | -27% | 0 | 0 | — |
case-19 | fail→pass | 9,815 | 2,307 | -76% | 1 | 1 | 0% | 709 | 1,254 | +77% | 0 | 0 | — |
case-20 | pass→pass | 6,350 | 2,164 | -66% | 1 | 1 | 0% | 1,015 | 1,225 | +21% | 0 | 0 | — |
case-21 | fail→pass | 6,322 | 1,490 | -76% | 1 | 1 | 0% | 1,155 | 1,089 | -6% | 0 | 0 | — |
case-22 | pass→pass | 10,430 | 6,784 | -35% | 1 | 1 | 0% | 1,675 | 1,889 | +13% | 0 | 0 | — |
case-23 | fail→pass | 7,700 | 1,169 | -85% | 1 | 1 | 0% | 1,433 | 1,063 | -26% | 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. 23 cases were attempted, and 18 counted toward the lift figure. The other 5 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +57 percentage points is the difference between those two pass rates over the 18 comparable cases. 1 case got worse with the skill loaded, and it is 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.