explaingit

anhtester/claude-testing-kit

★ 16Audience · developerComplexity · 2/5ActiveSetup · easy

TLDR

Vietnamese-language Claude Code configuration bundle for QA work, with 16 slash commands, role skills, and rules for Playwright, Selenium, and Appium.

Mindmap

mindmap
  root((claude-testing-kit))
    Inputs
      Requirements docs
      Existing test cases
      Swagger specs
    Outputs
      Manual test cases
      Automation scripts
      Test plans
      Reports
    Use Cases
      Risk based test design
      Manual to automation
      Flaky test triage
    Tech Stack
      Claude Code
      Playwright
      Selenium
      Appium

Things people build with this

USE CASE 1

Add Vietnamese-language QA slash commands to a Claude Code project

USE CASE 2

Generate Page Object Model Playwright or Selenium scripts from manual test cases

USE CASE 3

Run risk-based manual test case design with a six-step recipe

USE CASE 4

Build cross-module pairwise or Cartesian test matrices for regression

Tech stack

Claude CodePlaywrightSeleniumAppiumTypeScriptJava

Getting it running

Difficulty · easy Time to first run · 30min

Prompts and rules are in Vietnamese, so non-Vietnamese teams will need to translate the templates before use.

In plain English

Claude Testing Kit is a Vietnamese language configuration bundle for Claude Code, the coding assistant from Anthropic. It is put together by a Vietnamese tester known as Anh Tester for the local QA community. The repo does not contain an application of its own. Instead it ships a set of files that Claude Code reads when you drop them into the root of a project, and those files steer the assistant toward software testing tasks rather than general coding help. The kit is aimed at both manual testers and automation engineers. The README says it covers the whole testing lifecycle, from analyzing requirements and writing test cases by hand to generating automation scripts, hooking up CI/CD, and producing reports. All of the prompts and rules are written in Vietnamese so that the assistant replies in Vietnamese too. The folder layout is simple. A .claude folder holds 16 slash commands (such as /generate_automation_from_testcases, /generate_manual_testcases_rbt, and /generate_cross_module_test_plan), a rules folder that forces the assistant to follow Page Object Model structure and safe locator strategies for Playwright, Selenium, and Appium, a skills folder with 10 specialised roles like automation engineer, manual tester, UI debugger, locator healer, and Jira integrator, and a settings.json that limits what Claude Code is allowed to do on disk. Longer jobs live under plans/, which holds six step recipes for manual test case design using risk based testing, for building automation scripts from context to review, and for cross module pairwise or Cartesian test matrices. Shorter one-off jobs live under prompt_templates/, with ready to paste prompts for requirements analysis, framework setup in Playwright TypeScript or Selenium Java, manual to automation conversion, flaky test analysis, and API tests from Swagger. Usage is straightforward: copy the .claude folder and the CLAUDE.md file into a project, run claude in the terminal or open the Claude Code panel in VS Code, and type / to see the testing commands.

Copy-paste prompts

Prompt 1
Copy the .claude folder and CLAUDE.md from anhtester/claude-testing-kit into my Playwright project and list the new slash commands.
Prompt 2
Use the /generate_automation_from_testcases command in claude-testing-kit to turn my Excel test cases into Playwright TypeScript Page Object scripts.
Prompt 3
Show me the rules in claude-testing-kit that enforce Page Object Model and safe locator strategies for Selenium.
Prompt 4
Run the /generate_cross_module_test_plan command from claude-testing-kit on a checkout flow that spans cart, payment, and order modules.
Prompt 5
Adapt the prompt_templates in claude-testing-kit so the assistant replies in English instead of Vietnamese for my team.
Open on GitHub → Explain another repo

Generated 2026-05-22 · Model: sonnet-4-6 · Verify against the repo before relying on details.