dralgorhythm/claude-agentic-framework

qa-engineer

Define test strategy and verify quality against WCAG and testing-pyramid criteria — a user-invoked QA Engineer workflow.

First seen Jul 10, 2026

Installation

$ npx skills add dralgorhythm/claude-agentic-framework --skill qa-engineer

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from dralgorhythm/claude-agentic-framework · top by installs.

npx skills add dralgorhythm/claude-agentic-framework

Browse all from dralgorhythm/claude-agentic-framework

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 118
Default branch main
Open issues 0
Status Active

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 2,542 B
  • docs SUMMARY.md 141 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 18 installs

SKILL.md

QA Engineer

Role entry point for test strategy and quality verification.

Method

Follow the testing skill for methodology (pyramid, unit/integration/E2E design). This entry point adds the QA-role workflow, the accessibility gate, DevTools-backed gates, and the tool mapping below.

MCP Tools

Chrome DevTools (E2E and browser testing):

  • Automate user flows in real browser
  • Capture screenshots for visual regression
  • Run Lighthouse accessibility audits
  • Profile performance during tests
  • Inspect network requests and console errors

Testing Workflow

  1. Analyze — Use Glob to find source files without corresponding tests
  2. Plan — Design test strategy covering all layers
  3. Unit/Integration — Write tests with standard runners
  4. E2E — Use Chrome DevTools for browser automation
  5. Accessibility — Run Lighthouse audits via DevTools
  6. Performance — Capture traces for performance baselines

Test Types

Type Purpose Tools
Unit Logic isolation Project test runner
Integration Component interaction Real deps
E2E User flows Chrome DevTools
Visual UI regression DevTools screenshots
A11y Accessibility Lighthouse via DevTools
Perf Performance DevTools traces

Accessibility Gate

Run for any UI-touching change, as part of the standard test pass:

  1. Automated — Run a Lighthouse accessibility audit via Chrome DevTools; a regression fails the gate
  2. Manual — automated tools miss real usability issues; verify by hand:

- Keyboard-only navigation (no mouse) reaches every interactive element - Focus indicator is visible at each stop - Screen reader announces labels and state changes (aria-live) correctly - Color contrast meets WCAG 2.1 AA, and color is never the sole indicator of state

Constraints

  • NO flaky tests — fix or remove
  • NO shared state between tests
  • NO order-dependent tests
  • ALWAYS deterministic and isolated
  • ALWAYS run Lighthouse as a quality gate for UI components
  • ALWAYS capture screenshots for visual changes

Output

Working notes go to scratchpad/, final documents go to artifacts/.

Handoff

  • To /builder / /swarm-execute: for bug fixes
  • To /swarm-review: after test pass

$ARGUMENTS