agentleFS
Sign inStart free
cogine-aicogine-ai/cogine-dev-skillset/skills/qa-only/SKILL.md

qa-only

Skill4 starsChanged 4 months agoApache-2.0View on GitHub

Use when the user asks for a report-only browser QA pass on a web application, live site, preview URL, authenticated app flow, or release candidate without making code changes.

We split this skill into the rules it gives an agent

9 rules. 0 are common, 8 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 29 edited versions we read keep these.

  • 8

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill to test a web application like a user and produce an evidence-based QA report.

    line 8
  2. Distinctive:

    Do not use this for unit-test-only debugging, backend-only compiler failures, or design critique without functional browser testing.

    line 23
  3. Very specific:

    If the site needs auth, use get-cookies or an existing real browser session.

    line 34
  4. Distinctive:

    If no browser automation helper exists, use the available local browser tooling and state the limitation.

    line 35
  5. Distinctive:

    Use a 0-100 health score.

    line 48
  6. Distinctive:

    Do not report an issue without a reproduction path.

    line 102
  7. Distinctive:

    Do not hide limitations; list what was not tested.

    line 103
  8. Distinctive:

    Prefer concrete evidence over broad impressions.

    line 104
  9. Distinctive:

    Do not call a site ready if critical paths were not tested.

    line 105

“Common” is measured: kept by at least 6 in 10 of the 29 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

QA Only

Use this skill to test a web application like a user and produce an evidence-based QA report. Do not edit code, commit changes, or apply fixes.

Core Rule

Report only. If you find a bug, document the evidence, reproduction path, severity, and suggested fix. Do not fix it unless the user explicitly switches to an implementation task.

Scope

Use this for:

  • Web apps, dashboards, SaaS tools, AI application frontends, and preview deployments.
  • Authenticated flows when cookies or a real browser session are available.
  • Release-candidate checks before shipping.
  • "Find bugs", "test the site", "QA this", or "report-only QA" requests.

Do not use this for unit-test-only debugging, backend-only compiler failures, or design critique without functional browser testing.

Workflow

  1. Establish target and mode.
    • Target URL or local app URL.
    • Auth state: public, login required, cookie import needed, or real-browser session.
    • Mode: Quick, Standard, or Focused.
    • Scope exclusions named by the user.
  2. Prepare the browser path.
    • If the site needs auth, use get-cookies or an existing real browser session.
    • If no browser automation helper exists, use the available local browser tooling and state the limitation.
    • Capture screenshots for every reported issue when tooling allows it.
  3. Orient before testing.
    • Identify primary user journeys, navigation, forms, stateful actions, AI interactions, and payments or account settings if present.
    • For AI apps, include empty input, long input, model failure, slow response, retry, cancellation, and malformed output paths when relevant.
  4. Execute the QA pass.
    • Quick: homepage plus top 5 user paths; console/network errors; major broken states.
    • Standard: primary journeys, key forms, mobile/desktop layout, auth/session behavior, loading/error/empty states.
    • Focused: user-specified page, flow, or feature with deeper edge cases.
  5. Score and report.
    • Use a 0-100 health score.
    • Separate blockers from polish.
    • Include reproduction steps precise enough for another agent to verify.

Severity Guide

SeverityMeaning
BlockerCore flow cannot complete, data loss, auth break, payment break, or severe AI failure.
HighImportant user path broken or misleading.
MediumFunctional issue with workaround or limited reach.
LowMinor visual, copy, or interaction defect.

Report Contract

## QA Report

### Target
URL:
Mode:
Auth:
Date:

### Summary
Health score: __/100
Overall status: Ship / Ship with known issues / Do not ship

### Paths Tested
- <path>

### Findings

#### QA-001: <title>
Severity:
Path:
Evidence:
Reproduction steps:
Expected:
Actual:
Suggested fix:

### Screenshots / Artifacts
- <artifact>

### Not Tested
- <area>

### Recommended Next Step
- <next step>

Quality Bar

  • Do not report an issue without a reproduction path.
  • Do not hide limitations; list what was not tested.
  • Prefer concrete evidence over broad impressions.
  • Do not call a site ready if critical paths were not tested.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free