qa

Tests web applications systematically and produces structured QA reports with screenshots and health scores.

Updated Jan 13, 2026
One-click install
npx skills add https://github.com/psyche911/AntigravityTest --skill qa-psyche911
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: qa
Source: https://github.com/psyche911/AntigravityTest/tree/main/.claude/skills/gstack/qa
Command: npx skills add https://github.com/psyche911/AntigravityTest --skill qa-psyche911

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Manually testing a web application for bugs, broken links, console errors, and UX issues is slow and inconsistent. This Skill automates systematic QA testing of any web app, acting like a real user who clicks every element, fills every form, and documents every issue with screenshot evidence. ## Core Features & Use Cases - Three Testing Modes: Full mode for systematic exploration of every reachable page, quick mode for a 30-second smoke test, and regression mode to diff results against a saved baseline. - Evidence-Based Issue Reporting: Every bug is documented with annotated screenshots, reproduction steps, severity, and category, compiled into a structured Markdown report. - Health Score Rubric: Computes a weighted 0-100 health score across console errors, links, visual, functional, UX, performance, content, and accessibility categories. - Use Case: Before launching a new release of your Next.js app, run a full QA pass to catch hydration errors, broken forms, and mobile layout issues, then save the baseline to compare against the next release. ## Quick Start Ask the agent to QA test your site at https://myapp.com in full mode and produce a report with screenshots and a health score.

Frequently Asked Questions about qa

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I QA test a web application automatically?▼

Provide the target URL and the Skill explores every reachable page, clicks interactive elements, fills forms, and checks console errors. It documents 5-10 well-evidenced issues with screenshots and repro steps in a structured report.

What is the difference between full, quick, and regression QA modes?▼

Full mode systematically explores all pages over 5-15 minutes. Quick mode runs a 30-second smoke test of the homepage plus top 5 navigation targets. Regression mode runs a full pass and diffs issues and health score against a saved baseline.json.

Does this QA approach work with React, Vue, or other SPAs?▼

Yes. For single-page applications it uses interactive snapshots to find client-side navigation elements instead of relying on link extraction. It also checks stale state, browser history handling, and hydration errors for frameworks like Next.js.

How is the QA health score calculated?▼

Each category (console, links, visual, functional, UX, performance, content, accessibility) starts at 100 and loses points per issue by severity. The final score is a weighted average, with functional at 20% and console, UX, and accessibility at 15% each.

Can the QA skill test pages behind a login?▼

Yes. It supports credential-based login by filling the login form, cookie import from a JSON file, and can pause for 2FA codes or CAPTCHA completion. Passwords are always redacted as [REDACTED] in the report.

What are the limitations of automated browser QA testing?▼

It tests as an end user and never reads source code, so it cannot diagnose root causes. It also cannot bypass CAPTCHAs or 2FA without user help, and quick mode skips detailed per-page issue documentation.