Przejdź do treści
Biblioteka skilli
Programowanie · Przegląd i audyt

e2e-prune

/e2e-prune

Przegląda zestaw testów Playwright i proponuje, które testy usunąć i które kroki skrócić, z uzasadnieniem dla każdej zmiany; niczego nie zmienia.

Licencja
MIT
Ostatnia zmiana repozytorium
13.11.2026

Ocena i bezpieczeństwo

100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
1
Skrypty
0
Komentarze HTML
0
Adresy URL w SKILL.md
0
Przestarzałość

Brak sygnałów. Repozytorium zmienione 13.11.2026, bez odwołań do wycofanych modeli i API.

Licencja

MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.

Co robi i kiedy

Przegląda zestaw testów Playwright i proponuje, które testy usunąć i które kroki skrócić, z uzasadnieniem dla każdej zmiany; niczego nie zmienia.

Oryginalny opis (EN)

Review a Playwright suite and propose which tests to drop and which steps to trim, with one reason per finding. Use when a suite has grown faster than anyone can read it, when asked which tests are noise, duplicates or theater, or before a maintenance pass. Proposes only; changes no test and no app file.

Jak działa

  1. Classes
  2. 1. Run the suite once
  3. 2. Read every test
  4. 3. Report
Sekcje z nagłówków SKILL.md, w oryginale · 50 linii, 658 słów
Podgląd SKILL.md
---
name: e2e-prune
description: Review a Playwright suite and propose which tests to drop and which steps to trim, with one reason per finding. Use when a suite has grown faster than anyone can read it, when asked which tests are noise, duplicates or theater, or before a maintenance pass. Proposes only; changes no test and no app file.
license: MIT
---

# e2e-prune

Input: a suite directory (`tests/`, `e2e/`) and the environment file `.e2e/e2e-env.md`, called ENV below. No `.e2e/e2e-env.md` in the repo: run the `e2e-setup` skill first. Output: a report that names every test to drop, every step to trim, and the reason. Nothing is edited: deletion is the team's call, the skill supplies the evidence.

A test earns its place by failing when the user-visible behaviour it names breaks. Everything else is cost: minutes per run, a line in every failure report, a thing to update when the page changes.

## Classes

Each dropped test gets exactly one class, the strongest that applies:

- **no-assert**: no `expect` on anything a user sees. `goto` then `waitFor`, `toHaveURL` alone, a page `toBeVisible`. It is green against an empty page. Catches nothing.
- **duplicate**: asserts the same user-visible outcome as another test, through the same flow. Different title, different locators, same facts checked. One copy stays: the one that asserts more (a subset is the copy to drop); on equal assertions, the shorter one with the more durable locators. Overlap alone is not duplication: two tests that share setup steps but check different outcomes both stay, since a suite is run in parts.
- **implementation-coupled**: asserts how the page is built, not what it shows. CSS classes, element ids, DOM structure (`div > span`), API response shape, request payloads, data attributes other than the test ids ENV names as stable. Breaks on a refactor that changes nothing for the user. Where it also checks a user outcome, the test may be worth rewriting instead of dropping; say so.
- **wrong-scope**: not an end-to-end check. Arithmetic on constants, a unit test of a helper, a pixel or CSS-value check, a test with no `page` interaction. Right thing, wrong layer.

**trim** is for tests that stay: steps that prove nothing the test is about. Title checks, a `reload` with no state to re-read, a detour to a page the test never asserts on, the same assertion repeated after an action that could not change it, a reset (per ENV) when nothing was changed. Name the step, keep the test.

## 1. Run the suite once

If the Playwright config does not load `.env.e2e`, load it into the shell first (`set -a; . ./.env.e2e; set +a`).

```bash
npx playwright test --reporter=line
```

A failing test gets reported as failing, not classified: triage comes first (the `e2e-triage` skill). Note runtime per test; a drop saves that time on every run.

## 2. Read every test

For each test write down: title, pages visited, outcomes asserted (what the user would see), locator style. Two tests with the same outcomes list are duplicate candidates; a test with an empty outcomes list is no-assert or wrong-scope; a test whose assertions name markup is coupled. Steps that add nothing to the outcomes list are trim candidates.

Reading the test is enough for the verdict. Open the page with `playwright-cli` only when a locator's meaning is unclear (does `.badge` carry the status the user reads, or decoration?).

## 3. Report

Summary line, then one line per finding, nothing else per finding:

```
Suite: 10 tests, 4 files, 14.2s. Proposed: drop 7, trim 1, keep 2.

- drop checkout page opens: no-assert: only toHaveURL after goto, green on an empty page
- drop shipping fee drops to zero once the cart passes $100: duplicate: same amounts on /cart as "socks alone pay $5.99 shipping..." (shipping.spec.ts:3), keep that one
- drop product grid renders eight cards with price tags: implementation-coupled: asserts article.card.product and span.price classes
- drop shipping and tax math: wrong-scope: arithmetic on constants, never opens a page
- trim guest cart survives sign-in and is gone after sign-out: toHaveTitle twice, reload on /, empty-cart check before anything is added: none bear on the merge
```

Title is the exact test title. Class is one of the four. Reason names the evidence from the test body. A keep needs no line.

Wersje

W bazie nie ma innych skilli o nazwie e2e-prune.

Migawka bazy z 13.11.2026 · wersja paczki 4237e25d6989 (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 0 · ranking

Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.

Podobne skille

  • 81burn-tests
    Przycina zbiór testów, usuwając rozdęte i wolne testy, aż zostaną te, które naprawdę mają znaczenie.
  • 100review
    Przegląda testy Playwright pod kątem jakości i dobrych praktyk.
  • 100review
    Przegląda testy Playwright pod kątem jakości i dobrych praktyk.
  • 90write-playwright-test
    Pisze lub zmienia testy Playwright w repozytorium, używając obiektów stron i pokazując, że test wykrywa błąd.
  • 100generate
    Generuje testy Playwright dla komponentów, stron i funkcji, w tym testy e2e, na prośby o napisanie testów.
  • 100generate
    Generuje testy Playwright dla komponentów, stron i funkcji, w tym testy e2e, na prośby o napisanie testów.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.