test-designlisted
Install: claude install-skill app-vitals/shipwright
# test-design skill
## Purpose
Design the **target** test system. Greenfield. The blueprint Phase 3 reconciles existing tests against.
## When invoked
By the `/test-design` command. Requires `docs/test-readiness/test-inventory.md` from Phase 1.
## Process
### Step 1 — load the inventory
Read `docs/test-readiness/test-inventory.md`. If missing, abort with a clear "run /test-inventory first" message. Extract:
- Stack profile
- Per-layer test counts (how many unit, integration, smoke, E2E)
- External dependencies named (DBs, queues, third-party APIs)
- Canary suite candidates (canary-eligible roster)
### Step 2 — recommend frameworks per layer
Based on the stack profile, select one framework per layer. Use the matrix below as the default; override only if the inventory reveals a specific reason (e.g., existing build infra makes one choice radically cheaper).
| Stack | Unit | Integration | Smoke | E2E |
|---|---|---|---|---|
| TypeScript (Bun) | `bun test` | `bun test` + testcontainers | supertest / Hono test client | Playwright |
| TypeScript (Node) | Vitest | Vitest + testcontainers | supertest | Playwright |
| Python | pytest | pytest + testcontainers / docker-compose | httpx / FastAPI TestClient | Playwright (Python) |
| Go | `go test` | `go test` + testcontainers-go | `httptest` | Playwright |
| Ruby | RSpec | RSpec + database_cleaner | Rack::Test | Capybara / Playwright |
### Step 3 — local execution architecture
For every external dependency named in the invent