By hominhtuong
All-in-one QA toolkit for Claude Code; commands are namespaced /qa:<name>. Automation (Playwright Web + Appium iOS/Android), auto-routing by platform (web/android/ios) to load only the matching skill: exploratory/plan-tests/find-elements/cook/run/fix/analyze/count-cases/review/push-code/merge-request. Manual QA with distinct command names: analyze-spec/plan-gen-testcases/gen-testcases/count-testcases/log-bug/update-board.
Introduce & guide usage of the qa plugin — which commands, which skills, the automation vs manual workflows, and the platform router
Kill all running Appium servers using the kill-appium.sh script
Review the current git change (diff) against the design system & coding rules, output findings with code + how to fix
Review the ENTIRE codebase against the design system, coding rules and cross-file consistency
Analyze a requirement/spec doc from a QA perspective (testable items, ambiguity, risk, questions for the PO) into an md file used as input for /qa:plan-gen-testcases and /qa:gen-testcases (read-only, no edits)
Read Figma design screens and extract structured UI summary (layout, elements, states). Use when a Figma URL is provided and design context needs to be fetched. Returns structured markdown.
Read a single Lark wiki/docx document and extract structured content (text, tables, comments). Use when a Lark URL is provided and document content needs to be fetched. Returns structured markdown.
Discover elements from live app via Appium MCP, save as elements.json, then create/update Screen classes. Use when elements are unknown or need verification.
Reusable logic acting as a "green gate" — clean compile/build (auto-detects the project's build system) and (optionally) a quick run of the just-touched feature's tests. Use before finishing cook/fix/exploratory and before committing in push-code/merge-request. Do not raise timeouts / add workarounds just to "pass the test".
Reusable logic to commit & push safely — never push directly to main/master (auto-creates a feat/<slug> branch), commit per Conventional Commits with a Co-Authored-By trailer, push -u without force. Used by push-code and merge-request after the code is clean + build is green. Runs only when the user requests it.
Engine that writes mobile test code (Appium Java + TestNG + POM), used by the cook command when the platform is android/ios. Covers 3-layer element lookup, Screen declaration (via the declare-screen skill), writing atomic/flow Tests, TestNG XML (GoToHomeTest first), Utils.passCap/failCap with the [APP-BUG] triage label, build verification, sitemap updates, and recording elements missing an id. Keeps the cook command a thin router.
Reusable logic to write/maintain a Playwright Java Page Object (Screen) + <Feature>Tests class + smoke/regression runner following the 3-layer POM, with correct locator priority and isDisplayed(), and NO assert in the Screen. Used by the cook/plan-tests command (creating a new feature or extending one). Hooks the missing-ids skill for elements lacking data-testid; verifies via the build-verify skill.
Reusable logic to write/update a "Screen" Page Object following the project's POM standard (Appium Java, cross-platform Android/iOS/Flutter). Used when a command (cook, exploratory) needs to declare a screen as a Screen class — @MobileFindBy fields by locator priority, isDisplayed(), verb-style action methods, NO assert. Includes the strategy for choosing stable locators and hooks into the missing-ids skill for elements lacking an id.
Uses power tools
Uses Bash, Write, or Edit tools
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.
🌐 Tiếng Việt · English
Bộ plugin Claude Code mã nguồn mở cho mọi kỹ sư QA/QC dùng AI. Tự động hoá kiểm thử web & mobile, sinh test case từ tài liệu, exploratory testing, review code - biến Claude Code thành một "QA engineer" thực thụ ngay trong terminal/IDE của bạn.
Miễn phí & mở cho tất cả mọi người. Cài một lần, dùng cho mọi project có Claude. Mọi command gọi dạng /qa:<tên> (ví dụ /qa:exploratory).
/plugin marketplace add hominhtuong/qa-claude-plugins
/plugin install qa@qa-claude
Hoặc khai báo sẵn trong project (ai clone về cũng được hỏi cài) - .claude/settings.json:
{
"extraKnownMarketplaces": {
"qa-claude": { "source": { "source": "github", "repo": "hominhtuong/qa-claude-plugins" } }
},
"enabledPlugins": { "qa@qa-claude": true }
}
Bạn vẫn có thể thêm command riêng vào
.claude/commands/của project (gọi không prefix, vd/mycmd) - chạy song song, độc lập với command plugin (/qa:…).
Khi có bản mới:
/plugin marketplace update qa-claude
/reload-plugins
Chạy /qa:setup-plugin --update để làm mới các file .example/template trong .claude/qa-claude/ (file .env và log-bug.config.yml của bạn không bị ghi đè).
Cách gọi: tất cả command của plugin gọi dạng
/qa:<tên>(ví dụ/qa:exploratory). Đây là quy ước namespace bắt buộc của Claude Code cho plugin - không bỏ được. Những mục*-method/*-app/*-web… hiện trong menu là skill nội bộ, bạn không cần gọi trực tiếp - cứ gọi command.
| Command | Việc |
|---|---|
/qa:setup-plugin | Cài plugin vào project (1 lần): Claude tự chạy script tạo .claude/qa-claude/ + check toolchain. Không cần terminal. |
/qa:help [chủ đề] | Giới thiệu & hướng dẫn dùng plugin, liệt kê command/skill. |
/qa:status | Tóm tắt nhanh trạng thái project (git, device, Appium, coverage). |
/qa:ask <câu hỏi> | Hỏi đáp về codebase / kiến trúc / cấu hình / cách test (chỉ trả lời). |
/qa:missing-test-ids | Quản lý "nợ test-id" gửi dev (export / record / resolve). |
/qa:feedback <góp ý/lỗi> | Góp ý / báo lỗi plugin → mở sẵn 1 GitHub issue (kèm version/OS), bấm Submit là xong. |
Automation và Manual dùng tên riêng, không nhập nhằng:
cook(viết code) ↔gen-testcases(sinh test case) ·plan-tests↔plan-gen-testcases·analyze↔analyze-spec·count-cases↔count-testcases.
| Command | Việc |
|---|---|
/qa:exploratory <feature> [platform] | Khám phá màn hình như QA senior, săn bug, chụp bằng chứng, xuất bug report gửi dev. 🚦 Có [APP-BUG] → dừng, không viết test cho app sai. |
/qa:plan-tests <feature> | Thiết kế kế hoạch test automation (chỉ khi exploratory sạch). |
/qa:find-elements <màn> | Trích locator bền vững (tự rẽ web/android/ios). |
/qa:cook <plan|yêu cầu> | Viết Page Object + test code theo design pattern. |
/qa:run [platform] | Compile + chạy test + phân loại lỗi ([APP-BUG] vs [FRAMEWORK]/[ENV]/[DATA]). |
/qa:fix <bug> | Sửa lỗi (compile/test fail, flaky, vi phạm rule) đúng layer. |
/qa:analyze <code|kết quả> | Phân tích codebase / kết quả test / cấu trúc (chỉ đọc, không sửa). |
/qa:count-cases <feature> | Đếm testcase automation (@Test) hiện có + ước lượng tổng theo sub-feature. |
/qa:review-change | Soi git diff hiện tại theo design system & coding rules. |
/qa:review-codebase | Soi toàn bộ codebase theo rule + nhất quán cross-file. |
/qa:push-code | Review → tự fix Blocker → build xanh → commit & push branch. |
/qa:merge-request | Review → merge target an toàn → push → mở PR/MR (tự nhận GitHub gh / GitLab glab). |
/qa:kill-appium | Tắt mọi Appium server đang chạy. |
npx claudepluginhub hominhtuong/qa-claude-plugins --plugin qaTools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Comprehensive 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.