.claude/skills/visual-verify/SKILL.md
Visual UI verification with agent-browser. Use after implementing UI components to take screenshots, verify interactions, and self-check your work. FASTER than E2E tests.
npx skillsauth add lucidlabs-hq/agent-kit visual-verifyInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Verwende diesen Skill nach UI-Implementierungen um deine Arbeit visuell zu verifizieren.
┌─────────────────────────────────────────────────────────────────┐
│ │
│ DEVELOPMENT (90% der Zeit) PRODUCTION (10%) │
│ ───────────────────────── ────────────────── │
│ /visual-verify E2E Tests (Playwright) │
│ │
│ ✅ Schnell (Sekunden) ✅ Automatisiert │
│ ✅ Interaktiv ✅ CI/CD Integration │
│ ✅ Flexibel ✅ Regression Testing │
│ ✅ AI Self-Verification ✅ Kritische Flows │
│ │
│ Für: Jede UI-Änderung Für: Login, Checkout, │
│ Layout prüfen Payment, Signup │
│ Komponenten testen │
│ Responsive checken │
│ │
└─────────────────────────────────────────────────────────────────┘
Regel: Agent-Browser für Development-Verification. E2E nur für kritische Production-Flows.
# Installation (einmalig)
cd frontend && pnpm add -D agent-browser
pnpm exec agent-browser install # Chromium downloaden
cd frontend && pnpm run dev
# Seite öffnen
agent-browser open http://localhost:3000/[path]
# Accessibility Snapshot (zeigt interaktive Elemente mit Refs)
agent-browser snapshot -i
# Screenshot der aktuellen Ansicht
agent-browser screenshot
# Screenshot mit spezifischem Pfad
agent-browser screenshot ./screenshots/feature-name.png
# Full-Page Screenshot
agent-browser screenshot --full ./screenshots/full-page.png
# Nach snapshot -i siehst du Refs wie @e1, @e2, etc.
# Element klicken
agent-browser click @e1
# Text eingeben
agent-browser fill @e2 "Test Input"
# Hover
agent-browser hover @e3
# Nach Interaktion: neuer Snapshot + Screenshot
agent-browser snapshot -i
agent-browser screenshot
# Mobile Viewport
agent-browser viewport 375 812
# Tablet Viewport
agent-browser viewport 768 1024
# Desktop Viewport
agent-browser viewport 1920 1080
# Screenshot nach Viewport-Änderung
agent-browser screenshot ./screenshots/mobile.png
Bei UI-Implementierungen diesen Loop durchführen:
┌─────────────────────────────────────────────────────────┐
│ SELF-VERIFICATION LOOP │
│ │
│ 1. UI implementieren │
│ 2. agent-browser open http://localhost:3000/page │
│ 3. agent-browser snapshot -i (Struktur prüfen) │
│ 4. agent-browser screenshot (visuell prüfen) │
│ 5. Interaktionen testen (click, fill, hover) │
│ 6. Responsive testen (viewport ändern) │
│ 7. Bei Problemen: Code anpassen, Loop wiederholen │
│ │
└─────────────────────────────────────────────────────────┘
# Öffnen und Screenshot
agent-browser open http://localhost:3000/dashboard
agent-browser screenshot ./screenshots/dashboard-layout.png
# Viewport-Varianten
agent-browser viewport 375 812
agent-browser screenshot ./screenshots/dashboard-mobile.png
agent-browser viewport 1920 1080
agent-browser screenshot ./screenshots/dashboard-desktop.png
# Snapshot zeigt alle interaktiven Elemente
agent-browser snapshot -i
# Prüfen ob Button klickbar
agent-browser click @e1
agent-browser snapshot -i # Zustand nach Klick
# Prüfen ob Form funktioniert
agent-browser fill @e2 "[email protected]"
agent-browser click @e3 # Submit
agent-browser snapshot -i # Ergebnis prüfen
# Hover über Element
agent-browser hover @e1
agent-browser screenshot ./screenshots/button-hover.png
# Dialog öffnen
agent-browser click @e1 # Trigger Button
agent-browser wait visible "[role='dialog']"
agent-browser screenshot ./screenshots/modal-open.png
# Dialog schließen
agent-browser press Escape
agent-browser screenshot ./screenshots/modal-closed.png
# Session für spezifischen Test
agent-browser --session auth-test open http://localhost:3000/login
agent-browser --session auth-test fill @e1 "[email protected]"
agent-browser --session auth-test fill @e2 "password"
agent-browser --session auth-test click @e3
agent-browser --session auth-test screenshot ./screenshots/after-login.png
# Separate Session für anderen Test
agent-browser --session dashboard-test open http://localhost:3000/dashboard
# Nach Role suchen
agent-browser find role button click --name "Submit"
# Nach Text suchen
agent-browser find text "Sign In" click
# Nach Label suchen
agent-browser find label "Email" fill "[email protected]"
Nach jeder UI-Implementation:
./screenshots/feature-state.pngdevelopment
Deploy invoice-accounting-assistant to HQ server. Runs tests first (TDD), then builds and deploys. Use when ready to push changes to staging/production.
documentation
Update README with current project status and features. Use after completing features.
tools
--- name: time-report description: Cross-project time report. Aggregates all session data from ~/.claude-time/sessions/. Use to see how much time was spent across all projects. disable-model-invocation: true allowed-tools: Bash, Read argument-hint: [all | this-week | this-month | last-month | {project-name}] --- # Time Report: Cross-Project Session Overview ## Objective Read ALL session files from `~/.claude-time/sessions/*.json` and produce an aggregated time report. Supports filtering by pe
testing
Run tests with coverage report and gap analysis. MANDATORY before /commit and PR creation. Use after implementation or during validation.