Generates a comprehensive QA report with health scores and reproduction steps without modifying code.

Install

mkdir -p .claude/skills/qa-only && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/11598" && unzip -o skill.zip -d .claude/skills/qa-only && rm skill.zip

Installs to .claude/skills/qa-only

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Report-only QA testing — produces structured report with health score, screenshots, and repro steps. No fixes applied.
118 charsno explicit “when” trigger
Beginner

Key capabilities

  • Perform read-only QA testing
  • Generate structured QA reports
  • Capture screenshots of pages
  • List bugs with severity and repro steps
  • Report accessibility findings
  • Provide performance metrics

How it works

The skill executes QA steps (discover, smoke, deep QA, health score) in a strictly read-only manner. It then compiles findings into a detailed report without making any modifications.

Inputs & outputs

You give it
URL to test, mode (full or smoke)
You get back
QA-ONLY-{timestamp}.md report with health score, screenshots, bug list

When to use qa-only

  • Perform read-only QA audit
  • Generate bug report with repro steps
  • Check application health score

About this skill

qa-only

Report-only QA testing — produces structured report with health score, screenshots, and repro steps. No fixes applied.

Inputs

ParameterValuesDefaultDescription
urlstring--URL to test
modestringfullMode: full

/qa-only — Report-Only QA Testing

Same methodology as /qa but strictly read-only. No fixes. No code changes. Report only.

Step 0: Preamble

Run templates/PREAMBLE.md.

Execution

Follow the same Steps 1-4 as /qa (discover, smoke, deep QA, health score).

HARD-GATE: Do NOT dispatch any agent with Write or Edit tools. This is a READ-ONLY audit.

Output

Write report to .productionos/QA-ONLY-{timestamp}.md with:

  • Health score (0-100)
  • Screenshots of every page
  • Bug list with severity, repro steps, and file:line references
  • Accessibility findings
  • Performance metrics

Self-Eval

Run templates/SELF-EVAL-PROTOCOL.md on report completeness and evidence quality.

Error Handling

ScenarioAction
No target providedAsk for clarification with examples
Target not foundSearch for alternatives, suggest closest match
Missing dependenciesReport what is needed and how to install
Permission deniedCheck file permissions, suggest fix
State file corruptedReset to defaults, report what was lost

Guardrails

  1. Do not silently change scope or expand beyond the user request.
  2. Prefer concrete outputs and verification over abstract descriptions.
  3. Keep scope faithful to the user intent.
  4. Preserve existing workflow guardrails and stop conditions.
  5. Verify results before concluding.

When not to use it

  • When any fixes or code changes are required.
  • When the target is not found or missing dependencies.

Limitations

  • Do NOT dispatch any agent with Write or Edit tools.
  • Do not silently change scope or expand beyond the user request.
  • Keep scope faithful to the user intent.

How it compares

This skill provides a complete QA report with health scores, screenshots, and repro steps without altering any code or system state, unlike a full QA process that includes fixes.

Compared to similar skills

qa-only side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
qa-only (this skill)04moNo flagsBeginner
web-quality-audit76moReviewBeginner
roier-seo46moReviewIntermediate
web-testing16moReviewIntermediate

Try saying

Example prompts that trigger this skill in your AI assistant.

Search skills

Search the agent skills registry