Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Cursor rules for Cursor AI development with React, TypeScript, and shadcn/ui integration.
.claude/skills/amariahak-cursor-rules-for-cursor-ai-development-with-react-typescript-and-shadcn-ui-int/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 58% | 0% |
Cursor rules for Cursor AI development with React, TypeScript, and shadcn/ui integration.
Synced from https://github.com/PatrickJS/awesome-cursorrules/tree/main/rules/cursor-ai-react-typescript-shadcn-ui-cursorrules-p.mdc.
You are an expert AI programming assistant that primarily focuses on producing clear, readable React and TypeScript code.
You always use the latest stable version of TypeScript, JavaScript, React, Node.js, Next.js App Router, Shadcn UI, Tailwind CSS and you are familiar with the latest features and best practices.
You carefully provide accurate, factual, thoughtful answers, and are a genius at reasoning AI to chat, to generate code.
Style and Structure
Naming Conventions
TypeScript Usage
UI and Styling
Performance Optimization
Other Rules need to follow:
Don't be lazy, write all the code to implement features I ask for.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 26,756 | 25,465 | -5% | 1 | 1 | 0% | 6,206 | 6,429 | +4% | 0 | 0 | — |
case-02 | fail→fail | 26,902 | 20,574 | -24% | 1 | 1 | 0% | 6,195 | 4,996 | -19% | 0 | 0 | — |
case-03 | fail→fail | 25,183 | 26,103 | +4% | 1 | 1 | 0% | 6,194 | 6,417 | +4% | 0 | 0 | — |
case-04 | pass→pass | 11,863 | 17,364 | +46% | 1 | 1 | 0% | 2,306 | 2,337 | +1% | 0 | 0 | — |
case-05 | pass→pass | 13,374 | 10,740 | -20% | 1 | 1 | 0% | 2,446 | 2,276 | -7% | 0 | 0 | — |
case-06 | pass→pass | 8,303 | 7,921 | -5% | 1 | 1 | 0% | 1,644 | 1,765 | +7% | 0 | 0 | — |
case-07 | pass→pass | 12,094 | 16,478 | +36% | 1 | 1 | 0% | 2,276 | 3,792 | +67% | 0 | 0 | — |
case-08 | fail→pass | 12,281 | 14,686 | +20% | 1 | 1 | 0% | 2,610 | 3,722 | +43% | 0 | 0 | — |
case-09 | fail→fail | 26,600 | 10,814 | -59% | 1 | 1 | 0% | 4,028 | 2,715 | -33% | 0 | 0 | — |
case-10 | fail→pass | 11,922 | 12,089 | +1% | 1 | 1 | 0% | 2,559 | 2,916 | +14% | 0 | 0 | — |
case-11 | fail→fail | 5,281 | 5,172 | -2% | 1 | 1 | 0% | 929 | 1,267 | +36% | 0 | 0 | — |
case-12 | fail→pass | 3,523 | 6,957 | +97% | 1 | 1 | 0% | 630 | 1,039 | +65% | 0 | 0 | — |
case-13 | fail→fail | 15,813 | 16,993 | +7% | 1 | 1 | 0% | 3,569 | 4,010 | +12% | 0 | 0 | — |
case-14 | fail→pass | 16,495 | 13,733 | -17% | 1 | 1 | 0% | 3,850 | 3,451 | -10% | 0 | 0 | — |
case-15 | fail→pass | 12,014 | 18,407 | +53% | 1 | 1 | 0% | 1,700 | 2,692 | +58% | 0 | 0 | — |
case-16 | fail→fail | 12,383 | 11,129 | -10% | 1 | 1 | 0% | 2,679 | 2,471 | -8% | 0 | 0 | — |
case-17 | fail→pass | 11,192 | 10,678 | -5% | 1 | 1 | 0% | 2,401 | 2,577 | +7% | 0 | 0 | — |
case-18 | fail→pass | 13,154 | 9,194 | -30% | 1 | 1 | 0% | 1,627 | 1,921 | +18% | 0 | 0 | — |
case-19 | fail→pass | 15,609 | 9,611 | -38% | 1 | 1 | 0% | 3,156 | 2,334 | -26% | 0 | 0 | — |
case-20 | fail→fail | 11,610 | 12,711 | +9% | 1 | 1 | 0% | 2,434 | 3,087 | +27% | 0 | 0 | — |
case-21 | fail→fail | 11,822 | 14,264 | +21% | 1 | 1 | 0% | 2,507 | 3,834 | +53% | 0 | 0 | — |
case-22 | fail→pass | 18,228 | 17,427 | -4% | 1 | 1 | 0% | 4,031 | 4,245 | +5% | 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. 22 cases were attempted. The headline lift of +41 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.