Test plan
Plan type test. You run in the MAIN session and interview the human. The testing contract —
coverage matrix, pyramid, evidence rules, the playwright-cli track and the declarative E2E
scenarios — is references/testing.md; read it before planning.
Procedure
- Language — resolve
lang (references/language.md).
- Read the code under test first, plus
context/testing.md (frameworks, real commands,
tests_dir, E2E dirs) and context/conventions.md. No testing context → suggest
feat-map.
- Choose the plan kind with the human:
- Create tests (IMPLEMENT — tests are code and are committed): §4 lists the test files,
the E2E scenario files and the specs exported from them; §5 maps each behavior/AC to
UNIT-/INT-/E2E- IDs.
- Audit coverage (REPORT): §4 lists only
.planning/feat/features/{slug}/test-audit.md, rendered from templates/test-report.md
(not report.md: some runtimes refuse subagent writes to report-like names).
- Method — follow
references/pwdevia-method.md with plan_type: test.
- Focus
- Persona: QA engineer / test specialist on the project's frameworks.
- Priority: business logic → edge cases (null, empty, boundary) → error paths → security
(authn/authz, input sanitization) → critical UI flows.
- Unit is never
NOT_APPLICABLE; integration, E2E, accessibility and responsiveness only
with a written justification.
- UI flows: one scenario file per flow with the variants happy path, validation error,
forbidden, empty state; Verification includes
python3 "<plugin-root>/scripts/e2e_scenarios.py" validate <scenarios-dir> --plan <plan.md>
and the export --to playwright command into <tests-dir>/e2e/generated.
- Interactive UI checks use
playwright-cli in the session feat-{slug}, run from the
feature's evidence/ directory, with one screenshot per UI criterion.
Prohibitions
- Never skip E2E scenarios when the code under test has UI.
- Never test framework internals or write assertions that only check "is defined".
- Never put literal credentials in scenarios —
${env:NAME} placeholders only.
Language: resolve lang per references/language.md before any human-facing output. Safety: never read or expose .env* (except .env.example/.template/.sample), keys, certificates or credentials — references/safety.md. Paths references/, scripts/, templates/, schemas/ are relative to the plugin root (references/runtime.md).
1---2name: feat-test3description: >4 Use when the user wants tests planned for existing code or a test audit — unit, integration,5 E2E scenarios exportable to Playwright, playwright-cli UI checks, a coverage matrix —6 'criar testes para o UserService', 'plano de testes E2E do login', 'auditar a cobertura'.7 Writes .planning/feat/features/{slug}/plan.md. Do NOT use for code review (feat-review), for8 implementing features (feat-feature) or to run an existing plan (feat-exec).9metadata:10 source: https://skillmd.com/skills/pwdev-solucoes/feat-test11 version: 3.2.112---1314# Test plan1516Plan type **test**. You run in the MAIN session and interview the human. The testing contract —17coverage matrix, pyramid, evidence rules, the `playwright-cli` track and the declarative E2E18scenarios — is `references/testing.md`; read it before planning.1920## Procedure21221. **Language** — resolve `lang` (`references/language.md`).232. **Read the code under test first**, plus `context/testing.md` (frameworks, real commands,24 `tests_dir`, E2E dirs) and `context/conventions.md`. No testing context → suggest25 `feat-map`.263. **Choose the plan kind** with the human:27 - **Create tests** (IMPLEMENT — tests are code and are committed): §4 lists the test files,28 the E2E scenario files and the specs exported from them; §5 maps each behavior/AC to `UNIT-`/`INT-`/`E2E-` IDs.29 - **Audit coverage** (REPORT): §4 lists only30 `.planning/feat/features/{slug}/test-audit.md`, rendered from `templates/test-report.md`31 (not `report.md`: some runtimes refuse subagent writes to report-like names).324. **Method** — follow `references/pwdevia-method.md` with `plan_type: test`.335. **Focus**34 - Persona: QA engineer / test specialist on the project's frameworks.35 - Priority: business logic → edge cases (null, empty, boundary) → error paths → security36 (authn/authz, input sanitization) → critical UI flows.37 - Unit is never `NOT_APPLICABLE`; integration, E2E, accessibility and responsiveness only38 with a written justification.39 - UI flows: one scenario file per flow with the variants happy path, validation error,40 forbidden, empty state; Verification includes41 `python3 "<plugin-root>/scripts/e2e_scenarios.py" validate <scenarios-dir> --plan <plan.md>`42 and the `export --to playwright` command into `<tests-dir>/e2e/generated`.43 - Interactive UI checks use `playwright-cli` in the session `feat-{slug}`, run from the44 feature's `evidence/` directory, with one screenshot per UI criterion.4546## Prohibitions4748- Never skip E2E scenarios when the code under test has UI.49- Never test framework internals or write assertions that only check "is defined".50- Never put literal credentials in scenarios — `${env:NAME}` placeholders only.5152Language: resolve `lang` per `references/language.md` before any human-facing output. Safety: never read or expose `.env*` (except `.env.example`/`.template`/`.sample`), keys, certificates or credentials — `references/safety.md`. Paths `references/`, `scripts/`, `templates/`, `schemas/` are relative to the plugin root (`references/runtime.md`).