By Rabusek
Mobile & Web QA toolkit dla Claude Code — skills, subagents, hooks, MCP. PII redaction, MR review, flake triage, cross-device validation.
Orchestrates parallel cross-device validation for mobile QA specs. Use when asked to validate a spec across multiple devices or to find cross-device flake. Spawns one parallel sub-task per target device (default Pixel 8, Samsung Galaxy S24, iPhone 15), aggregates pass/fail into a matrix, and highlights scenarios that fail on a subset of devices.
Proposes new locator candidates when a mobile or web QA test selector breaks after a release. Use proactively when a test fails with NoSuchElementException, ElementNotFound, or selector timeout. Reads a DOM dump or page source, prefers accessibility-id, and ranks candidates by confidence with justification.
Reviews mobile/web QA test MRs for team conventions. Use proactively when reviewing pull/merge requests touching WDIO, Appium, or Playwright specs. Checks ticket prefix in branch and describe blocks, flags XPath selectors, hardcoded waits (browser.pause, driver.sleep, setTimeout), real PII in fixtures, and missing accessibility-id on new elements.
Groups overnight QA test failures by root cause for the morning digest. Use proactively at the start of a workday or when handed a test-results JSON file. Reads test-results, clusters failures by signature, ranks by impact, suggests owners via git blame, and drafts Jira ticket text.
Classifies a flaky WDIO/Appium spec as intermittent, deterministic-flake, or locator-broke by analyzing the last 10 BrowserStack runs. Use when the user names a spec and asks "is this flaky", "why is X failing on CI", "should we skip this test", or "triage this flake".
Generuje polskie podsumowanie MR/PR (5–8 punktów, format Slack) z diffa lub gita. Wykrywa numery ticketów, zmienione selektory, nowe spec-y, usunięte it.skip, zmiany w package.json. Użyj gdy użytkownik prosi o "podsumowanie MR", "PR po polsku", "streszczenie diffa", lub poda referencję MR.
Parses a BrowserStack/SauceLabs/LambdaTest test failure (pasted JSON or session ID) into a clean 5-line markdown summary plus likely root-cause hypotheses. Use when the user shares a cloud-grid failure, a session UUID, asks "why did this run fail", or pastes WDIO/Appium grid output.
Redacts PII (emails, phone numbers, credit cards, PESEL, NIP, IBAN, parcel/tracking IDs, BrowserStack session UUIDs, bearer tokens) from logs, fixtures, or pasted text before sending to external APIs. Use when the user says "redact", "anonymize", "scrub", "sanitize logs", or "clean fixture before paste".
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Requires secrets
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Needs API keys or credentials to function
Needs API keys or credentials to function
Uses power tools
Uses Bash, Write, or Edit tools
Uses power tools
Uses Bash, Write, or Edit tools
Pluginowy zestaw narzędzi Claude Code dla zespołów Mobile/Web QA — skills, subagenty, hooki i MCP w jednym pakiecie.
| Skill | Opis |
|---|---|
redact-pii | Maskuje dane osobowe (PESEL, email, numery telefonów, adresy) w logach przed wklejeniem do ticketu lub Slacka |
parse-grid-failure | Parsuje fail z BrowserStack/Sauce Labs/Selenium Grid i wyciąga: stack trace, kroki repro, urządzenie, sesję |
flake-triage | Klasyfikuje czy fail jest flakiem (timing, sieć, env) czy realnym bugiem; sugeruje retry vs investigate |
mr-summary-pl | Generuje polskie podsumowanie MR/PR pod review: zakres zmian, ryzyka, sugerowane testy regresyjne |
| Agent | Opis |
|---|---|
sentinel | Code review pod kątem QA — sprawdza testowalność, accessibility-id, brak XPath, retry policy |
healer | Auto-naprawa flaków — analizuje historię, proponuje konkretne poprawki w kodzie testów |
triage | Pierwsza linia segregacji failów: bug vs flake vs env, przypisanie ownerstwa |
device-matrix | Walidacja cross-device — sprawdza czy testy obejmują wymagane urządzenia/wersje OS |
| Hook | Trigger | Opis |
|---|---|---|
block-force-push.sh | PreToolUse na git push --force | Blokuje force-push do gałęzi chronionych (main, release/*) |
pii-scan.sh | UserPromptSubmit | Skanuje prompt na PESEL/email/telefon zanim trafi do modelu |
audit-edit.sh | PostToolUse na Edit|Write|MultiEdit | Loguje wszystkie edycje plików do .audit/edits.jsonl |
notify-on-failed-test.sh | Stop | Wysyła powiadomienie do Slacka jeśli ostatni test failed |
Szczegóły konfiguracji i wymagania — patrz hooks/README.md.
Pakiet konfiguruje cztery serwery MCP dla zespołu QA:
~/qa-logs/Szczegóły, klucze i format .env — patrz mcp/README.md.
| Command | Opis |
|---|---|
/new-spec <TICKET-ID> | Generuje szablon WDIO/Appium spec z happy path + edge case + timeout, accessibility-id only |
# w sesji Claude Code
/plugin marketplace add Rabusek/qa-mobile-pack
/plugin install qa-mobile-pack@Rabusek
git clone https://github.com/Rabusek/qa-mobile-pack ~/qa-mobile-pack
claude --plugin-dir ~/qa-mobile-pack
Albo na stałe — dodaj wpis do ~/.claude/settings.json:
{
"enabledPlugins": {
"qa-mobile-pack@local": true
},
"extraKnownMarketplaces": {
"local": {
"source": { "source": "local", "path": "/Users/<ty>/qa-mobile-pack" }
}
}
}
Dla użytkowników na starszych wersjach Claude Code lub w środowiskach gdzie nie można korzystać z /plugin install:
git clone https://github.com/Rabusek/qa-mobile-pack
cd qa-mobile-pack
./install.sh
Skrypt linkuje skills/*, agents/*, commands/* do ~/.claude/ i dopisuje hooki do ~/.claude/settings.json. Idempotentny — można odpalać wielokrotnie.
/plugin install)install.shmcp/.env.example):
SLACK_WEBHOOK_URL — dla notify-on-failed-test.shBROWSERSTACK_USERNAME, BROWSERSTACK_ACCESS_KEY — dla MCP BrowserStackJIRA_TOKEN lub LINEAR_API_KEY — dla MCP ticketówGITLAB_TOKEN lub GITHUB_TOKEN — dla MCP review/redact-pii
[wklej log z prawdziwym PESEL i emailem]
→ Skill zwróci log z <PESEL_REDACTED>, <EMAIL_REDACTED>. Bezpieczne do wklejenia.
/parse-grid-failure
[wklej output z https://automate.browserstack.com/dashboard/...]
→ Dostajesz strukturyzowany raport: stack trace, urządzenie, kroki repro, link do video sesji.
@sentinel zreview ten MR: https://gitlab.com/<org>/<repo>/-/merge_requests/123
→ Subagent sentinel używa MCP GitLab, czyta diff, raportuje problemy QA i sugeruje testy regresyjne.
npx claudepluginhub rabusek/qa-mobile-packComprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.
Production-grade engineering skills for AI coding agents — covering the full software development lifecycle from spec to ship.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Complete developer toolkit for Claude Code