qa

Locate and reproduce app defects with screenshot evidence and health scoring.

Updated Apr 14, 2026
One-click install
npx skills add https://github.com/sr-ai-dev/sr-harness --skill qa-sr-ai-dev
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: qa
Source: https://github.com/sr-ai-dev/sr-harness/tree/main/skills/qa
Command: npx skills add https://github.com/sr-ai-dev/sr-harness --skill qa-sr-ai-dev

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you discover real user-facing issues (visual, functional, UX, and accessibility) and turns them into verified, evidence-backed fixes so an app is safer to release.

Core Features & Use Cases

  • Plan-first exploratory QA: builds a structured test plan, navigates systematically, and records screenshots as evidence.
  • Mode-aware testing: automatically selects browser (chromux/CDP), computer (MCP screenshot + pixel interactions), or CLI (tmux capture-pane) testing based on the target, with explicit overrides.
  • Health scoring and ship readiness: computes before/after health scores and produces a ship-readiness summary using a weighted rubric.
  • Fix + verify loop (optional): for report mode it only documents; otherwise it applies minimal atomic code commits per issue, then re-tests to verify (or reverts on regression).
  • Spec-drift friendly: designed to complement verification by catching hidden failures and inconsistencies through additional browser verification heuristics and drift checks.

Quick Start

Use the qa skill to test a website end-to-end and produce a report with evidence screenshots and a before/after ship readiness health score.

Frequently Asked Questions about qa

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

FAQPage Schema
How do I perform exploratory QA on a web app and capture screenshot evidence of bugs?

Exploratory QA on a web app is performed by building a structured test plan, navigating systematically via chromux/CDP, and recording screenshot evidence for every discovered defect. This process ensures visual and functional issues are documented with verified proof.

What is the best way to find and fix UI bugs before shipping a release?

Finding and fixing UI bugs before shipping is best handled by running systematic bug hunting, applying minimal atomic code commits per issue, and re-testing to verify the fix. It computes a before/after health score to produce a final ship-readiness summary.

Can I run automated browser testing on desktop and native applications?

Automated browser testing is not used for desktop and native applications. Instead, the system automatically selects computer-use mode, utilizing MCP screenshot and pixel interactions to test desktop interfaces effectively.

Does this QA tool work with CLI and REPL interfaces?

Yes, this QA tool works with CLI and REPL interfaces by using tmux capture-pane. It automatically selects this mode during Phase 0 to test command-line interactions and capture output for evidence.

How do I generate a health score and ship readiness summary for my application?

To generate a health score and ship readiness summary, you run a standard or exhaustive test tier. The system computes before and after health scores using a weighted rubric, producing a structured summary of application readiness.

What happens if a code fix causes a regression during bug hunting?

If a code fix causes a regression during bug hunting, the system automatically reverts the change. This ensures the application remains stable by only keeping verified, atomic commits that pass re-verification.