Qa

QA & User Testing Guide

Examine the codebase to determine current build state, then generate the appropriate test plan.

Session Structure

1. First Impressions (5 min, no instructions given)

Observe: Can the user accomplish the primary task without guidance? Where do they get confused? Do they recover or give up? What do they try first?

2. Core Flow (10 min, free exploration)

Track: Task completion rate, time to complete key flows, error encounters, workarounds attempted, points of friction.

3. Edge Cases (10 min, guided)

Test specific scenarios: empty states, error states, slow network, invalid input, concurrent actions, undo/redo, back navigation.

4. Cross-Environment (if applicable)

Different browsers, devices, screen sizes, OS versions, network conditions.

Observation Checklist

| Category | What to Watch | |----------|---------------| | Confusion | Pauses, re-reads, wrong clicks, "what does this do?" | | Frustration | Repeated attempts, sighing, giving up | | Delight | "Oh nice", speed increases, exploring voluntarily | | Errors | Error messages shown, unhandled states, data loss | | Performance | Visible lag, loading states, janky animations |

Red Flags (Stop and Fix)

• User can't complete the primary task without help
• Data loss or corruption under normal usage
• Error with no recovery path (stuck states)
• Performance so bad it affects usability
• Security issue exposed to the user (visible tokens, other users' data)

Success Metrics

Define before testing:

| Metric | Target | How to Measure | |--------|--------|----------------| | Task completion rate | > 90% | Users who complete primary task without help | | Time to complete | < Xs | Stopwatch from start to success | | Error rate | < 5% | Unhandled errors per session | | Satisfaction | Positive | Post-session feedback |

Post-Session Questions

1. "What was that experience like?"
2. "What was confusing or frustrating?"
3. "What did you expect to happen that didn't?"
4. "Would you use this? Why or why not?"
5. "What's the one thing you'd change?"

Bug Report Format

FOUND: [what happened]
EXPECTED: [what should have happened]
STEPS: [how to reproduce]
ENVIRONMENT: [browser, device, OS]
SEVERITY: [blocks usage / annoying / cosmetic]

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: pvnarp
• Source: pvnarp/agent-skills
• License: MIT

Install and usage instructions live in the source repository linked above.