QA Skills
Fifty testing skills for Claude Code, Codex and Cursor: Playwright, strategy, API, visual and load checks.
- 266 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Skills write and run tests and project commands, and the runner executes checks in your code.
Install
- Works in
Manual install
npx skills add petrkindlmann/qa-skills playwright-automation test-strategySelective install from the README; a full clone with `git clone https://github.com/petrkindlmann/qa-skills.git .skills/qa-skills` also gives the local runner.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Skills write and run tests and project commands, and the runner executes checks in your code.
This is third-party code. Review the repository files before installing.
What it does
QA Skills bundles 50 skills for test automation and test organization. Among them are `playwright-automation`, `test-strategy`, `ai-test-generation`, `ai-bug-triage`, `ci-cd-integration`, `qa-metrics`, `release-readiness`, `visual-testing`, `performance-testing` and `accessibility-testing`. It covers Playwright and Cypress, API, unit and mobile tests, plus accessibility, email, payment and analytics checks. Chosen skills install selectively with `npx skills add petrkindlmann/qa-skills playwright-automation test-strategy`. A full clone of the repository also gives a local runner: `npm run qa -- run --project /path/to/app --goal "..."` finds the project test command, picks skills for the goal, runs bounded checks and writes `report.html` with evidence in JSON and logs.
Who it is for. QA engineers and developers who want an agent to write and triage tests from ready methods instead of generic prompts.
Good fit when
- Playwright tests are needed for a checkout flow.
- Tests must be generated from a PRD and a recurring bug classified and triaged.
- Before a release an answer is needed on what to check, plus a readiness report.
- Test reporting and quality metrics must be set up in GitHub Actions.
Not a fit when
- A single skill for one framework is needed: a specialized card is simpler.
- Installing individual skills must provide a ready CLI: it exists only with a full clone.
- Code must be checked without letting the agent change tests.
Example request
Write Playwright tests for the checkout flow with the playwright-automation skill and show which checks remain uncovered.Limitations
Skill choice depends on the agent: name a skill explicitly for ambiguous requests. Installing single skills does not install the CLI, and the `npm run qa` runner needs a full clone and `npm ci`. `init --wire-agent` adds an instruction to project files and takes effect only if the agent reads `AGENTS.md`. Some skills cover payment, email and load checks, which should run on test environments only. The author drew the patterns from six production sites and the latest commit is dated October 2026.
How to disable. Delete the installed skills from the agent directory or the `.skills/qa-skills` folder. Remove the instruction added by `qa init --wire-agent` from the project `AGENTS.md` by hand.
FAQ
How do I install single skills?
Run `npx skills add petrkindlmann/qa-skills playwright-automation test-strategy` with the names you need. Other ways: a clone in `.skills/qa-skills`, a git submodule or manual copies of folders from `skills/`.
What does the local QA runner do?
`npm run qa -- run --project /path/to/app --goal "Check the critical user flows"` finds the test command, picks skills, runs bounded checks and creates `report.html`, JSON and logs. Skipped checks stay visible.
Which agents are supported?
Claude Code, Codex, Cursor, Gemini CLI, VS Code and other agents that follow the Agent Skills standard.
How is a skill picked for a request?
The agent picks it from its description: "write Playwright tests for checkout" triggers `playwright-automation`, and "what should we test before the release" triggers `release-readiness`. Name the skill explicitly when in doubt.
Is there a way to test the tests themselves?
Yes, the lab plants 28 deliberate faults: 19 in a local shop and 9 in a pinned TodoMVC. Reports sit in the `docs/` folder.
Discussion
No questions or reviews yet.
Related
SkillsThirteen testing skills from requirements to report: strategy, cases, API checks, Playwright and flaky-test triage.
A library of 169 QA skills per language: test design, API, UI, regression, load, security and LLM testing.
Seven QA agents and ten skills for test automation: Playwright in TypeScript, Selenium in Java, API, accessibility and ISTQB.
Guides for agents writing and fixing Playwright tests, with advice on locators, waiting and diagnosing CI failures
The QASkills.sh directory MCP server: an agent searches, reads and installs testing skills without leaving the conversation.
57 reference files on Playwright tests in TypeScript: the agent pulls the one that fits the task, from locators to CI sharding
README in short
The document describes the set of 50 skills, four installation methods, a local runner and a table mapping phrases to skills. Separate reports analyze checks against planted faults. The skills follow patterns from the author real projects. MIT license.