Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Configure visual regression testing with Percy, Chromatic, or custom screenshot comparison
.claude/skills/a5c-ai-visual-regression-setup/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -68% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 13% | 0% |
Configure visual regression testing for desktop applications to catch unintended UI changes.
json{ "type": "object", "properties": { "projectPath": { "type": "string" }, "provider": { "enum": ["percy", "chromatic", "reg-suit", "custom"] }, "framework": { "enum": ["playwright", "cypress", "puppeteer"] } }, "required": ["projectPath"] }
javascript// playwright.config.js import { defineConfig } from '@playwright/test'; export default defineConfig({ use: { screenshot: 'only-on-failure' } }); // test.spec.js import { test } from '@playwright/test'; import percySnapshot from '@percy/playwright'; test('visual test', async ({ page }) => { await page.goto('/'); await percySnapshot(page, 'Home page'); });
javascriptconst { toMatchImageSnapshot } = require('jest-image-snapshot'); expect.extend({ toMatchImageSnapshot }); test('matches screenshot', async () => { const screenshot = await page.screenshot(); expect(screenshot).toMatchImageSnapshot({ failureThreshold: 0.01, failureThresholdType: 'percent' }); });
playwright-electron-configdesktop-ui-testing process| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 18,214 | 10,128 | -44% | 1 | 1 | 0% | 4,023 | 2,500 | -38% | 0 | 0 | — |
case-02 | pass→pass | 14,053 | 7,660 | -45% | 1 | 1 | 0% | 2,620 | 1,529 | -42% | 0 | 0 | — |
case-03 | fail→pass | 8,661 | 5,895 | -32% | 1 | 1 | 0% | 1,372 | 1,516 | +10% | 0 | 0 | — |
case-04 | fail→pass | 8,079 | 3,584 | -56% | 1 | 1 | 0% | 1,604 | 674 | -58% | 0 | 0 | — |
case-05 | pass→pass | 7,713 | 6,546 | -15% | 1 | 1 | 0% | 1,474 | 1,372 | -7% | 0 | 0 | — |
case-06 | pass→pass | 2,580 | 2,344 | -9% | 1 | 1 | 0% | 425 | 773 | +82% | 0 | 0 | — |
case-07 | pass→pass | 11,712 | 5,559 | -53% | 1 | 1 | 0% | 2,122 | 1,515 | -29% | 0 | 0 | — |
case-08 | pass→pass | 13,194 | 3,134 | -76% | 1 | 1 | 0% | 2,512 | 984 | -61% | 0 | 0 | — |
case-09 | fail→pass | 10,480 | 1,905 | -82% | 1 | 1 | 0% | 2,162 | 701 | -68% | 0 | 0 | — |
case-10 | fail→pass | 16,169 | 5,908 | -63% | 1 | 1 | 0% | 1,315 | 1,483 | +13% | 0 | 0 | — |
case-11 | fail→pass | 6,960 | 1,911 | -73% | 1 | 1 | 0% | 1,246 | 736 | -41% | 0 | 0 | — |
case-12 | pass→pass | 5,593 | 5,778 | +3% | 1 | 1 | 0% | 1,009 | 1,407 | +39% | 0 | 0 | — |
case-13 | pass→pass | 4,840 | 4,226 | -13% | 1 | 1 | 0% | 699 | 1,243 | +78% | 0 | 0 | — |
case-14 | pass→pass | 4,486 | 3,331 | -26% | 1 | 1 | 0% | 889 | 998 | +12% | 0 | 0 | — |
case-15 | fail→pass | 10,701 | 1,678 | -84% | 1 | 1 | 0% | 2,241 | 677 | -70% | 0 | 0 | — |
case-16 | pass→pass | 2,603 | 2,060 | -21% | 1 | 1 | 0% | 457 | 770 | +68% | 0 | 0 | — |
case-17 | fail→pass | 9,064 | 6,497 | -28% | 1 | 1 | 0% | 1,933 | 1,691 | -13% | 0 | 0 | — |
case-18 | pass→fail | 4,974 | 7,480 | +50% | 1 | 1 | 0% | 1,085 | 1,800 | +66% | 0 | 0 | — |
case-19 | pass→pass | 5,142 | 4,945 | -4% | 1 | 1 | 0% | 1,083 | 1,415 | +31% | 0 | 0 | — |
case-20 | pass→pass | 6,234 | 6,649 | +7% | 1 | 1 | 0% | 1,263 | 1,747 | +38% | 0 | 0 | — |
case-21 | pass→pass | 11,778 | 8,691 | -26% | 1 | 1 | 0% | 2,143 | 2,026 | -5% | 0 | 0 | — |
case-22 | pass→pass | 4,390 | 2,907 | -34% | 1 | 1 | 0% | 850 | 974 | +15% | 0 | 0 | — |
case-23 | pass→pass | 5,628 | 2,575 | -54% | 1 | 1 | 0% | 1,071 | 846 | -21% | 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. The headline lift of +30 percentage points is the difference between those two pass rates over the 23 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.