Przejdź do treści
Biblioteka skilli
Programowanie · Testowanie

dester-cases

/dester-cases

Pisze przypadki testowe dla systemu lub ekranu z opisu, w pięciu kategoriach, prostym językiem i z odniesieniem do wymagań.

Licencja
MIT
Ostatnia zmiana repozytorium
29.10.2026

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
3
Skrypty
0
Komentarze HTML
0
Adresy URL w SKILL.md
0
Przestarzałość

Brak sygnałów. Repozytorium zmienione 29.10.2026, bez odwołań do wycofanych modeli i API.

Licencja

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

Co robi i kiedy

Pisze przypadki testowe dla systemu lub ekranu z opisu, w pięciu kategoriach, prostym językiem i z odniesieniem do wymagań.

Wyzwalacze z opisu

write test casesmake a test plancases for this screen
Oryginalny opis (EN)

Write test cases for a system, feature or screen from a derived spec, an exploration map, a link you can open, or a repo. Newbie-friendly, plain language. Produces cases in five categories (happy path, positive, negative, edge, boundary) with traceability to a requirement or spec row, in Markdown and CSV ready to import into a test tool. Use when the user says /dester-cases, 'write test cases', 'make a test plan', 'cases for this screen', or after dester-explore.

Jak działa

  1. Where the expectation comes from (this decides everything)
  2. Inputs (ask, with defaults)
  3. Method
  4. Case format
  5. Files (in the project folder; create it if missing)
  6. Self-check before handing over
  7. Hand-off
Sekcje z nagłówków SKILL.md, w oryginale · 99 linii, 932 słów
Podgląd SKILL.md
---
name: dester-cases
description: "Write test cases for a system, feature or screen from a derived spec, an exploration map, a link you can open, or a repo. Newbie-friendly, plain language. Produces cases in five categories (happy path, positive, negative, edge, boundary) with traceability to a requirement or spec row, in Markdown and CSV ready to import into a test tool. Use when the user says /dester-cases, 'write test cases', 'make a test plan', 'cases for this screen', or after dester-explore."
metadata:
  author: Destine April
  version: 0.1.0
  license: MIT
  suite: dester
---

# dester-cases v0.1.0

Author: Destine April. Read `references/core-rules.md` first; those rules win.

**What it does:** turns what is known about a system into test cases a person (or a Playwright spec later) can run.
Each case says what to do, what should happen, and **where that expectation comes from**.

## Where the expectation comes from (this decides everything)

A test case is only as good as its expected result. Never take the expected result only from what the app did.

| Source | Label on the case |
|---|---|
| Owner or developer confirmed it, or a real document | `Confirmed` |
| A derived spec row with confidence Seen, confirmed by the owner | `Confirmed` |
| A derived spec row Seen but not confirmed | `Observed (needs owner confirmation)` |
| From code or guess | `Hypothesis` |
| Common sense or standard behavior (a password field hides input, required field cannot be empty) | `Standard` |

Cases labelled Observed or Hypothesis are listed in `questions.md` so the owner can confirm them.

## Inputs (ask, with defaults)

1. **Scope:** which module, screen, flow or role? *Default: the highest-risk charter from `03_CHARTERS.md`.*
2. **Source:** `02_DERIVED_SPEC.md` and `01_MAP.md` if they exist (best). Otherwise a link (open it, see Access mode in
   `dester-explore`), a repo (read only), screenshots, or the user's description. If nothing exists yet, offer to run
   `dester-explore` first, or to do a quick look at only this screen.
3. **Detail level:** short (title, steps, expected) or full (add preconditions, data, priority). *Default: full.*
4. **Output:** Markdown, CSV for import (Testiny, TestRail, Jira Xray, Excel), or both. *Default: both.*
5. **Cases are for:** manual run now, handing to a client, or automation later. This only changes wording.

## Method

1. List the **requirements or spec rows** in scope. One row of the spec becomes one or more cases. Show the list
   and let the user add or remove rows.
2. For each row write cases in five categories (skip a category only if it truly does not apply, and say so):
   - **Happy path:** everything valid, normal user.
   - **Positive:** valid but different (optional fields empty, other valid options, other role that is allowed).
   - **Negative:** invalid input, forbidden role, wrong order, missing data.
   - **Edge:** unusual but valid (long text, special characters, duplicates, empty states, many records).
   - **Boundary:** just below, at and just over each limit (0, 1, max, max+1); dates and amounts at cut-offs.
3. Add **role cases**: for each role and each key action, allowed or denied, and what the user sees (menu, button,
   direct URL). Add **data accuracy cases**: the calculation done by hand (show the working) versus what the
   system shows.
4. Prioritize with risk (`~/.claude/skills/dester-explore/references/charters.md`): P1 money, data loss, security, core flow;
   P2 common features; P3 cosmetic or rare.
5. Never invent limits, formulas or message text. If unknown: write the case with `Expected: TBD, ask owner` and add a
   question. A guessed expected result is a trap.

## Case format

```
TC-<MODULE>-<nnn> | <title: action and condition> | P1/P2/P3 | Category
Source: <spec row, requirement or "Standard"> | Basis: Confirmed / Observed / Hypothesis / Standard
Preconditions: <role, data, state>
Steps:
1. ...
Expected: <specific, checkable result>
Data: <fictional values>
Status: Not run | Pass | Bug | Blocked | Flaky   (Pass only with evidence, core rules 5 and 7)
Evidence: <path or "-">
```

Automation-ready cases (when the cases will become Playwright checks with `dester-check`):
- One action per step, from a known start (a page, or sign-in as a role). Use the verb the user performs and the exact
  on-screen label in quotes: `Click the "Save" button`, `Type in the "Name" field`.
- Add a locator hint only when it was actually seen in the page or code: `{testid: save}`, `{role: button, name: Save}`.
  Never invent one.
- Say what to wait for before the next step ("wait for the 'Saved' message").
- **Never depend on one account's data.** Do not name records that exist only in the account you explored (customers,
  invoices, ids, emails, amounts copied from the screen). Make the case create its own data with a unique `QA ` name,
  or state a placeholder precondition ("an active customer with an unpaid invoice exists, set up by the tester").
  System reference data (fixed statuses, roles, dropdown options) is fine.

Rules for good cases: one idea per case; steps a newbie can follow; expected is observable (a message, a number, a
visible state), not "works fine"; fictional data only (`QA Test Merchant 01`).

## Files (in the project folder; create it if missing)

```
05_TEST_CASES.md        grouped by module, then category, with a coverage table at the top
05_TEST_CASES.csv       one row per case: ID, Module, Title, Priority, Category, Source, Basis, Preconditions, Steps, Expected, Data, Status
05_COVERAGE.md          spec row or requirement -> case IDs; rows with no case; cases with no source
```

## Self-check before handing over

- Every case has a Source and a Basis. Every spec row in scope has at least one case, or is listed as uncovered.
- No expected result is `Observed` without being flagged for confirmation.
- Counts add up: cases per category and per priority match the lists. CSV row count equals the case count.
- Fictional data only. No secrets. No case claims Pass without evidence.

## Hand-off

Show: counts by priority and category, what is uncovered, what needs owner confirmation (first). Ask whether to run
some cases now (record results with evidence), export, or continue to `dester-docs`. Never push to a test tool or
repo without a yes for that exact action.
Pliki w paczce (3)
  • SKILL.md 6 KB
  • references/core-rules.md 3 KB
  • references/glossary.md 2 KB

Wersje

W bazie nie ma innych skilli o nazwie dester-cases.

Migawka bazy z 13.11.2026 · wersja paczki a61942531105 (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

  • 100dester-explore
    Bada system bez dokumentacji i zamienia go w mapę, wyprowadzoną specyfikację oraz sesje testowe oparte na misjach.
  • 100dester-report
    Zamienia wyniki QA w raporty błędów i końcowy raport z rekomendacją wdrożenia oraz sekcją tego, czego nie sprawdzono.
  • 100test-cases
    Generuje pełny zestaw przypadków testowych z wymagania, historyjki lub opisu funkcji, z przypadkami granicznymi i walidacją skryptem.
  • 90write-tests
    Pisze same przypadki testowe z wymagań lub zgłoszenia, bez ich uruchamiania, i tworzy pliki testów wraz z planem testów.
  • 100test-scenarios
    Wyprowadza scenariusze QA i przypadki testów akceptacyjnych ze specyfikacji lub historyjki użytkownika: ścieżki szczęśliwe, przypadki brzegowe, błędy, warianty danych i przypadki ewaluacji ML.
  • 100test-cases
    Generuje czytelne przypadki QA dla zadania, PR lub diffu: ścieżkę poprawną, błędną i brzegowe, zebrane w tabeli.

Uwagi i zgłoszenia

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