40 skills · Open source · MIT

Stop fixing your AI agent's tests.

You ask it to write a Playwright test. It gives you CSS selectors, waitForTimeout, and login boilerplate copy-pasted into every file. You spend more time fixing the output than you saved. These 40 skills help your agent generate tests that follow modern QA practice instead of generic LLM patterns.

Works with Claude Code, Codex, Cursor, Gemini CLI, VS Code, Windsurf
Built by Petr Kindlmann — QA automation across 6+ production sites, creator of breakit, Complicer, and KindLM.
Terminal
$ npx skills add petrkindlmann/qa-skills
Installed 40 skills from petrkindlmann/qa-skills
10 categories · 14 reference files
$ claude
> Write a Playwright test for login
Using skill: playwright-automation
getByRole selectors
storageState fixture
No waitForTimeout
Page Object pattern
SelectorsAvoids waitForTimeout and brittle CSS selectors
PlanningPlans tests before generating code
ReliabilityClassifies flaky failures by root cause
PatternsFollows reusable fixtures and POM patterns

You've seen these in PR reviews

The tests technically pass. But you know they'll break next week, and nobody will know why.

Selectors

div.container > ul > li:nth-child(3) — breaks the moment someone touches the layout. No data-testid, no getByRole. Just fragile paths through the DOM.

Architecture

Every test logs in from scratch. No storageState, no fixtures, no page objects. Tests that take 40 seconds when they should take 4.

Strategy

Same depth everywhere. The settings page gets the same coverage as checkout. No risk assessment, no pyramid — just tests for the sake of tests.

After installing, your agent starts following QA patterns you'd normally enforce in review

playwright
Uses getByRole instead of CSS selectors. Creates fixtures instead of beforeEach hooks. Follows your POM. Has an explicit list of 10 things it will not do — like waitForTimeout or force:true without justification.
test-reliability
When a test is flaky, it classifies the root cause (timing? data? environment?), quarantines it in CI, and applies the right fix — instead of just retrying three times and hoping.
ai-test-generation
Before writing any code, it builds a coverage matrix, identifies risks, and generates assertions separately from scenarios. You review a structured plan, not a wall of generated code.
ci-cd
Gives you ready-to-use GitHub Actions workflows. Unit tests on push, E2E on PR, full suite on merge to main. Sharding, artifact storage, flaky quarantine — all configured.
test-strategy
Produces an actual strategy doc your engineering manager can read. Test pyramid analysis, risk matrix, quality gates, realistic coverage targets — not a shelf document.
qa-metrics
Tracks flakiness rate, defect escape rate, and MTTR with formulas you can plug into Grafana. Includes targets by company stage — what's realistic for a 20-person startup vs. a 500-person org.
Other tools by Petr
breakit

25 AI personas test your website in real Playwright browsers. Findings deduplicated, scored, delivered as HTML report.

Complicer

GDPR, accessibility, and security compliance monitoring. Scheduled audits, visual review flows, evidence reports.

KindLM

Behavioral regression testing for AI agents and LLM workflows. YAML-defined tests, CI-native, multi-model.

N8Calls

Voice agents, chatbots, lead qualification, and workflow automation. 15 live builds, fixed pricing.

40 skills across 10 categories

Coverage across the full QA lifecycle — from strategy through automation to production monitoring.

Foundation
1 skill
qa-project-context
Strategy
4 skills
test-strategy, test-planning, risk-based-testing, exploratory-testing
Automation
7 skills
playwright-automation, cypress-automation, api-testing, unit-testing, mobile-testing, visual-testing, performance-testing
Specialized
4 skills
accessibility-testing, security-testing, cross-browser-testing, database-testing
AI-Augmented QA
4 skills
ai-test-generation, ai-bug-triage, test-reliability, ai-qa-review
Infrastructure
5 skills
ci-cd-integration, test-environments, test-data-management, contract-testing, service-virtualization
Metrics
3 skills
qa-metrics, qa-dashboard, coverage-analysis
Process
5 skills
shift-left-testing, qa-project-bootstrap, release-readiness, quality-postmortem, compliance-testing
Production
3 skills
testing-in-production, synthetic-monitoring, observability-driven-testing
Knowledge
3 skills
ai-system-testing, chaos-engineering, test-migration

Built from real QA review work

Based on repeated Playwright and QA review problems: brittle selectors, boilerplate setup, weak assertions, and poor risk-based coverage.