test-reportlisted
Install: claude install-skill YevheniiaQA/qa-agent-toolkit
# Test Report
A Playwright HTML report answers "which tests failed". It does not answer the
questions a team actually asks: what is covered, which failures are deliberate,
what can a manual tester execute, and is the suite trustworthy.
This skill produces one self-contained `REPORT.html` — no server, no build step,
no network. Double-click and it opens.
## Inputs
| Source | Provides |
|---|---|
| `results/run-*.json` (Playwright JSON reporter) | Per-test outcomes of each run: steps, errors, attachments |
| `05-metrics/events.jsonl` via `qa-metrics --json` | Stability across runs (`automation.suites.<suite>`: runs, clean runs, inconsistent tests, runtime), false failure rate, mutation score, defect counts |
| `02-analysis/<TICKET>/scenarios/api.json`, `ui.json` | Scenario text, data, expectations, defect refs |
| `02-analysis/<TICKET>/scenarios/RTM.md` or the suite's waiver block | Requirement coverage |
| `03-tests/*/TEST_CATALOGUE.md` | Human-readable case text |
| `RUN_REPORT.md` | Environment, contract version, dataset size |
Missing inputs degrade the report rather than blocking it — render the sections
you have and mark the others as unavailable with the reason. A report that
refuses to build because one file is absent is a report nobody gets.
## Building it
Use `report-template.html` in this skill folder. Do not write HTML from scratch:
a consistent report is one people learn to read quickly, and a regenerated layout
each time defeats that.
The template contains