Przejdź do treści
Biblioteka skilli
Programowanie · Testowanie

stack-testing

/stack-testing

Podpowiada dyscyplinę testów na prawdziwym systemie dla pytest, Go, C++, Django, Spring Boot i FastAPI, także przy e2e w przeglądarce.

Repozytorium
proofpunk ↗
Licencja
MIT
Ostatnia zmiana repozytorium
14.09.2026

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
19
Skrypty
7
Komentarze HTML
0
Adresy URL w SKILL.md
0
Przestarzałość

Brak sygnałów. Repozytorium zmienione 14.09.2026, bez odwołań do wycofanych modeli i API.

Licencja

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

Co sprawdzić przed instalacją
  • 7 skryptów w paczce. Przejrzyj je przed instalacją, bo Claude może je uruchomić.

Skrypty: examples/condition_based_waiting.ts, examples/console_logging.py, examples/element_discovery.py, examples/static_html_automation.py, scripts/playwright/lib/helpers.js, scripts/playwright/run.js, scripts/with_server.py

Co robi i kiedy

Podpowiada dyscyplinę testów na prawdziwym systemie dla pytest, Go, C++, Django, Spring Boot i FastAPI, także przy e2e w przeglądarce.

Oryginalny opis (EN)

Real-system test discipline per stack — pytest/Go/C++/Django/Spring Boot gotchas that cause flaky CI, FastAPI HTTP/SSE testing with curl Playwright browser automation with server lifecycle management, and condition-based waiting to kill timing flakes. Use when writing debugging, or deflaking test suites in Python, Go, C++, Django, Spring Boot, or FastAPI projects; when tests pass locally but fail in CI; when browser e2e needs a dev server managed; or when any test uses sleep()/arbitrary timeouts. Iron Rule: mocks chapters in references are for understanding existing suites only; new tests run

Jak działa

  1. Run checklist
  2. Operating Rules
  3. Reference Routing
  4. Browser E2E Quickstart
  5. Anti-Patterns
  6. Bundled resources
  7. Example
  8. Skill calls
Sekcje z nagłówków SKILL.md, w oryginale · 99 linii, 677 słów
Podgląd SKILL.md
---
name: stack-testing
description: >
  Real-system test discipline per stack — pytest/Go/C++/Django/Spring Boot
  gotchas that cause flaky CI, FastAPI HTTP/SSE testing with curl
  Playwright browser automation with server lifecycle management, and
  condition-based waiting to kill timing flakes. Use when writing
  debugging, or deflaking test suites in Python, Go, C++, Django, Spring
  Boot, or FastAPI projects; when tests pass locally but fail in CI; when
  browser e2e needs a dev server managed; or when any test uses
  sleep()/arbitrary timeouts. Iron Rule: mocks chapters in references are
  for understanding existing suites only; new tests run against the real
  system. Not for end-user proof of a finished feature (use
  end-user-testing) or single-bug root causes (use root-cause-debugging).
---

# Stack Testing

## Run checklist

Copy this checklist and track your progress:

- [ ] Identify the stack and read its chapter in references
- [ ] Write the test against the REAL system (no new mocks)
- [ ] Start required services with lifecycle management
- [ ] Replace sleeps with condition-based waiting
- [ ] Run repeatedly to confirm no flakes (CI parity)

Test against the REAL system, per stack. This skill bundles the field-tested
"what goes wrong" knowledge for each major test stack and the browser-e2e
helpers, under one discipline: no mocks for new tests, no arbitrary sleeps
evidence for every green claim.

## Operating Rules

1. **Iron Rule applies to tests too** — per `../../references/end-user-actor.md`
   and the evidence contract at `../../references/evidence-contract.md`: a green
   test run you did not personally execute is UNVERIFIED. Never claim "tests
   pass" from reading code.
2. **No new mocks/stubs/test doubles.** Several references below document mock
   frameworks (GoogleMock, MockMvc, factory patterns) — those chapters exist to
   help you READ and FIX existing suites. New tests you write hit the real
   database, real server, real browser. Testcontainers-style real services in
   Docker are the gold pattern (see `references/springboot-testing-gotchas.md`).
3. **No `sleep()` in tests.** Replace every arbitrary timeout with condition
   polling — `references/condition-based-waiting.md` is the canonical pattern
   with `examples/condition_based_waiting.ts`.
4. **Flaky = bug.** A test that "sometimes passes" is a defect in the test or
   the system, never a reason to re-run CI until green. Diagnose per stack
   below; escalate to the `root-cause-debugging` skill when the cause is unclear.
5. **Test output is REGRESSION evidence, never VALIDATION evidence.** Pipe the
   final run to a file and seal it with the `end-user-testing` skill's
   `fresh_evidence.py` — labeled REGRESSION ("the suite is green"). The
   feature-level verdict comes only from driving the live system as the end
   user (the shared runbooks (`../../references/*-validation.md`): `curl` for JSON backends, browser for UI
   simulator for mobile). A green suite cited as proof a feature works is a
   doctrine violation — see `../../references/end-user-actor.md`.

## Reference Routing

| Situation | Read |
|---|---|
| pytest fixture/parametrize/async/coverage weirdness | `references/pytest-gotchas.md` |
| Go `t.Parallel()` races, TestMain, benchmarks, test cache | `references/go-testing-gotchas.md` |
| C++ GoogleTest/CTest, sanitizers, coverage, fuzzing | `references/cpp-testing-gtest.md` |
| Django TestCase vs TransactionTestCase, factory_boy, DRF | `references/django-testing-gotchas.md` |
| Spring Boot Testcontainers, context cache, @Transactional | `references/springboot-testing-gotchas.md` |
| FastAPI/OpenAI-compatible HTTP + SSE endpoints via curl | `references/fastapi-backend-testing.md` |
| Local web app e2e with dev-server lifecycle (Python Playwright) | `references/webapp-testing.md` + run `scripts/with_server.py` |
| Full browser automation (multi-viewport, forms, link checks) | `references/playwright-browser-automation.md` + import/read `scripts/playwright/` helpers |
| Any timing flake, race, "works on my machine" | `references/condition-based-waiting.md` |

## Browser E2E Quickstart

```bash
# single dev server + Playwright script, lifecycle managed
python3 scripts/with_server.py --server "npm run dev" --port 5173 \
  -- python3 your_check.py
# richer automation: universal Playwright runner (requires npm i playwright)
cd scripts/playwright && npm install && node run.js --help
```

`examples/` holds reconnaissance-then-action, console-logging, and static-HTML
patterns from the source skills.

## Anti-Patterns

- Writing a mock to "unblock" a failing real dependency → fix the real system.
- Re-running CI to make a flake disappear → diagnose with the stack reference.
- Asserting "tests pass" from a stale terminal scrollback → fresh evidence run.
- Copying fixture patterns between stacks → scoping rules differ per framework;
  read the matching gotchas file FIRST.


## Bundled resources

- `scripts/playwright/lib/helpers.js` — server lifecycle + condition-based waiting helpers; read as reference or import from Playwright specs.
- `scripts/playwright/package.json` — dependency manifest for the helpers.
- `references/playwright-api-reference.md` — full Playwright API reference; grep it when a specific API is needed, do not read end-to-end.

## Example

**Input:** User: 'This FastAPI SSE test flakes in CI.'

**Output:** The mock-based test is rewritten to run against the real server with curl, sleep(2) replaced by a condition-based wait on the stream, 20 consecutive CI-parity runs pass.

## Skill calls

| Calls | When | What it hands over |
|-------|------|--------------------|
| `root-cause-debugging` | flakes that survive condition-based waiting | root cause, never a retry |

Called by: `production-readiness`, `proofpunk`.
Pliki w paczce (19)
  • SKILL.md 6 KB
  • examples/condition_based_waiting.ts 5 KB
  • examples/console_logging.py 1 KB
  • examples/element_discovery.py 1 KB
  • examples/static_html_automation.py 1 KB
  • references/condition-based-waiting.md 5 KB
  • references/cpp-testing-gtest.md 11 KB
  • references/django-testing-gotchas.md 7 KB
  • references/fastapi-backend-testing.md 4 KB
  • references/go-testing-gotchas.md 6 KB
  • references/playwright-api-reference.md 16 KB
  • references/playwright-browser-automation.md 15 KB
  • references/pytest-gotchas.md 5 KB
  • references/springboot-testing-gotchas.md 7 KB
  • references/webapp-testing.md 5 KB
  • scripts/with_server.py 4 KB
  • scripts/playwright/package.json 1 KB
  • scripts/playwright/run.js 6 KB
  • scripts/playwright/lib/helpers.js 12 KB

Wersje

W bazie nie ma innych skilli o nazwie stack-testing.

Migawka bazy z 13.11.2026 · wersja paczki 7749ecdabc1b (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

  • 100generating-tests-from-code
    Analizuje istniejący kod API (Express, NestJS, Spring Boot, FastAPI, Django) i na podstawie tras, kontrolerów i middleware generuje skrypty testów obciążeniowych k6.
  • 100django-dev-test
    Pomaga pisać testy aplikacji Django w pytest-django z factory_boy do generowania danych testowych, w tym fixtures i pokrycie testami.
  • 100spring-boot-testing
    Pisze testy jednostkowe, integracyjne i slice dla aplikacji Spring Boot 3.
  • 100spring-testing-browser
    Testuje interfejsy webowe Spring Boot w przeglądarce: HtmlUnit, Selenium, formularze, kliknięcia i logowanie.
  • 95django-tdd
    Prowadzi testy Django przez pytest-django, TDD, factory_boy, mocki i pokrycie, także dla API w DRF.
  • 95pytest-django
    Obsługuje testy integracyjne Django w pytest-django: fixtures db, client i settings, DRF APIClient, factory_boy, widoki async, sygnały, komendy zarządzania i testy wielu baz.

Uwagi i zgłoszenia

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