name	test
description	Generate test plan, test cases, test data fixtures, and Playwright E2E tests mapped to acceptance criteria.
disable-model-invocation	true
argument-hint	[--plan-only \| --e2e-only]

/test — Test Planning + E2E

Usage

/test                         # full test phase
/test --plan-only             # just generate test plan + cases
/test --e2e-only              # just generate Playwright tests

Prerequisites

specs/stories/ must exist
Source code in backend/ and frontend/ must exist (run /implement first)
Docker Compose stack must be deployable (run /deploy first for E2E)

Steps

Read .claude/skills/test-patterns/SKILL.md for test patterns.
Read .claude/skills/test-patterns/references/playwright.md for Playwright patterns.
Spawn test-engineer agent.
Agent generates to specs/test_artefacts/:
- test-plan.md — strategy, scope, environments
- test-cases.md — all cases mapped to acceptance criteria
- test-data/ — realistic fixtures (JSON/PDF)
- e2e/flows/ — Playwright test files

Copy Playwright config template:

cp .claude/skills/test-patterns/templates/playwright.config.ts playwright.config.ts

Install Playwright browsers:

npx playwright install --with-deps chromium

Verify Docker stack is healthy before E2E:

docker compose up -d --build
curl --retry 5 --retry-delay 3 http://localhost:8000/health
curl --retry 5 --retry-delay 3 http://localhost:3000

Run Playwright: npx playwright test
Generate report: npx playwright show-report

Verification

uv run pytest -x -q           # unit tests still pass
npm test                       # vitest still passes
npx playwright test            # E2E tests pass
npx playwright show-report     # visual report

Gotchas

Test cases not mapped to acceptance criteria. Every test case must reference a story ID and criterion. Unmapped tests are either missing coverage or testing nothing useful.
Playwright selectors using CSS classes. Use getByRole(), getByLabel(), getByText() — never CSS selectors or XPath. Class names change; roles don't.
Flaky waits. Never page.waitForTimeout(5000). Use expect(locator).toBeVisible() or page.waitForResponse() for deterministic waits.
E2E tests without Docker. Playwright config must use webServer to start Docker Compose. Tests that assume services are already running will fail in CI.
Missing test data fixtures. Use realistic data (real names, valid amounts) — not "test", "foo", "123".

Más de este repositorio

mismo repositorio

cross-provider-handoff

rlpatrao/claude_harness_forge

Switch model mid-session preserving thinking blocks. Used by /model command and by BRD §6.2 failover when the primary provider rate-limits or fails.

2026-05-194

extended-react

rlpatrao/claude_harness_forge

Six-phase ReAct loop (pre-check, thinking, self-critique, action, tool, post). Activates per-workflow via the thinking_level knob in config/workflows.yaml. Replaces standard ReAct for workflows where independent verification of the reasoning improves outcomes.

2026-05-194

instinct-extraction

rlpatrao/claude_harness_forge

Mines completed sessions for repeating {tool sequence → outcome} tuples. Scores by frequency × success_rate × novelty. High-scoring tuples become "instincts" — candidate skill seeds — in instincts/pending/.

2026-05-194

spec-backprop

rlpatrao/claude_harness_forge

Walk-back algorithm for the Spec-Auditor subagent. Given a failure at phase N, identifies the earliest phase whose spec, if tightened, would have prevented the failure. Proposes a surgical amendment.

2026-05-194

tree-sessions

rlpatrao/claude_harness_forge

Sessions stored as trees (not lists). /fork creates a branch from any point; /tree navigates; /branch labels a path; /export produces HTML for review. All branches live in one session file under sessions/<project>/<session_id>.json.

2026-05-194

iterative-retrieval

rlpatrao/claude_harness_forge

Progressive context refinement for subagents. Don't dump everything; retrieve in passes. Start with a list of file paths and one-line abstracts; load full content only when the abstract proves insufficient.

2026-05-194

/test — Test Planning + E2E

Usage

/test # full test phase /test --plan-only # just generate test plan + cases /test --e2e-only # just generate Playwright tests

Prerequisites

specs/stories/ must exist

Source code in backend/ and frontend/ must exist (run /implement first)

Docker Compose stack must be deployable (run /deploy first for E2E)

Steps

Read .claude/skills/test-patterns/SKILL.md for test patterns.

Read .claude/skills/test-patterns/references/playwright.md for Playwright patterns.

Spawn test-engineer agent.

Agent generates to specs/test_artefacts/:

test-plan.md — strategy, scope, environments
test-cases.md — all cases mapped to acceptance criteria
test-data/ — realistic fixtures (JSON/PDF)
e2e/flows/ — Playwright test files

Copy Playwright config template:

cp .claude/skills/test-patterns/templates/playwright.config.ts playwright.config.ts

Install Playwright browsers:

npx playwright install --with-deps chromium

Verify Docker stack is healthy before E2E:

docker compose up -d --build
curl --retry 5 --retry-delay 3 http://localhost:8000/health
curl --retry 5 --retry-delay 3 http://localhost:3000

Run Playwright: npx playwright test

Generate report: npx playwright show-report

Verification

uv run pytest -x -q # unit tests still pass npm test # vitest still passes npx playwright test # E2E tests pass npx playwright show-report # visual report

Gotchas

Test cases not mapped to acceptance criteria. Every test case must reference a story ID and criterion. Unmapped tests are either missing coverage or testing nothing useful.

Playwright selectors using CSS classes. Use getByRole(), getByLabel(), getByText() — never CSS selectors or XPath. Class names change; roles don't.

Flaky waits. Never page.waitForTimeout(5000). Use expect(locator).toBeVisible() or page.waitForResponse() for deterministic waits.

E2E tests without Docker. Playwright config must use webServer to start Docker Compose. Tests that assume services are already running will fail in CI.

Missing test data fixtures. Use realistic data (real names, valid amounts) — not "test", "foo", "123".