qa-skills
Thirteen testing skills from requirements to report: strategy, cases, API checks, Playwright and flaky-test triage.
- 64 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The skills run scripts, browser tests and requests to the system under test.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/fishzjp-qa-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skills run scripts, browser tests and requests to the system under test.
This is third-party code. Review the repository files before installing.
What it does
qa-skills turns an agent into a tester that moves through stages: clarify requirements, assess risk, write executable cases, automate them and compile a report. The set has thirteen skills and a shared `core/` knowledge base. Run the whole chain through the `qa` skill or take a single stage. Individual skills write and review cases, build a strategy across ten test types, automate end-to-end checks with Playwright, test APIs from OpenAPI, run exploratory sessions, find the root cause of a confirmed bug and triage flaky tests. The repository has five with-and-without comparison examples.
Who it is for. Testers and developers without a dedicated QA who want a repeatable feature checking process instead of one-off agent prompts.
Good fit when
- A PRD arrives and cases must be written from it, with doubts raised as questions.
- Existing cases are reviewed for coverage and executability.
- A set of API checks is built from an OpenAPI description.
- The same test fails once every few runs and the cause is needed.
- A team decides what to test for a new feature and what to leave out.
Not a fit when
- A load test run is needed: the skills only prepare a handoff for k6.
- The project runs on native Windows without WSL.
- The team does not read Chinese: the instruction bodies are written in it.
Example request
Analyze this requirement, write test cases and propose what to automate with Playwright.Limitations
Instruction bodies are written in Chinese although skill descriptions are mostly bilingual. Native Windows was not tested and the author advises WSL. You need Bash 3.2 or newer and Python 3.9 or later, plus Playwright for end-to-end tests. Installation, correct triggering and output quality are verified to different degrees: for Codex only install and paths are checked and the effect is not yet evaluated systematically. A single skill without `core/` breaks relative links.
How to disable. Remove the installed skills with skills.sh or `./uninstall.sh --target <directory>` if you used the script. Generated test projects and reports stay in your repository.
FAQ
Can one skill be installed alone?
No. Without the `core/` directory relative paths break. If you installed partially, repeat the full install with `--skill '*'`.
Which skill to choose?
For a full cycle take `qa`, which orchestrates stages. For one step use `test-case-writing`, `api-testing`, `automated-e2e-testing`, `bug-analysis` or `test-reliability`.
Which agents are supported?
Claude Code is the primary target. Install records exist for Codex and DeepSeek Harness; other hosts fit the structure but need separate verification.
What is the ten-axis matrix?
The type matrix the `test-strategy` skill uses to decide which kinds of checks a feature needs before handing off to specialist skills.
How to upgrade?
Rerun the skills.sh install command. For a script install run `git pull` and repeat the same command.
Discussion
No questions or reviews yet.
Related
SkillsFifty testing skills for Claude Code, Codex and Cursor: Playwright, strategy, API, visual and load checks.
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.
A self-improving QA agent: natural-language tests for web and mobile that self-heal and remember runs
Guides for agents writing and fixing Playwright tests, with advice on locators, waiting and diagnosing CI failures
Eleven skills to verify AI-written code: change blast radius, defect scan, review report, test cases and browser flow replay
README in short
A set of testing skills with a shared knowledge base, an end-to-end pipeline and checkable quality gates, plus example results with and without the skills.