nw-ab-critique-dimensi…
Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation
Test directory structure patterns by architecture style, language conventions, naming rules, and fixture placement. Decision tree for selecting test organization strategy.
$ npx -y skills add nWave-ai/nWave --skill nw-test-organization-conventions --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/nw-test-organization-conventionsContext preview
The summary Claude sees to decide when to auto-load this skill.
Test directory structure patterns by architecture style, language conventions, naming rules, and fixture placement. Decision tree for selecting test organization strategy.
name: nw-test-organization-conventions description: Test directory structure patterns by architecture style, language conventions, naming rules, and fixture placement. Decision tree for selecting test organization strategy. user-invocable: false disable-model-invocation: true
Test directory structure encodes architectural boundaries. If a developer cannot infer the architecture from the test tree alone, the organization is wrong.
What is the project's architectural style?
|
+-- Hexagonal / Clean Architecture
| --> Test-type-first: tests/{unit,integration,acceptance,e2e}/
| --> Domain concepts nested within each type
| --> Port contract tests in tests/integration/
|
+-- Vertical Slice
| --> Co-located: features/{slice}/tests/
| --> Cross-slice tests in top-level tests/cross_feature/
|
+-- Modular Monolith
| --> Module-first: tests/modules/{module}/{unit,integration}/
| --> Architecture tests per module (dependency rule enforcement)
| --> Inter-module tests in tests/inter_module/
|
+-- Microservices
| --> Per-service test tree: {service}/tests/{unit,integration,component,contract}/
| --> Contract tests: consumer writes in own repo, provider verifies in own repo
| --> Cross-service E2E in separate project: e2e-tests/
|
+-- Event-Driven
| --> Test-type-first with event-specific categories
| --> Unique: schema contract tests, idempotency tests, saga compensation tests
|
+-- CQRS
| --> Command/Query split within each test type
| --> Unique: projection tests (rebuild + idempotency)
|
+-- Layered (N-Tier)
| --> Mirror source: tests mirror src layer hierarchy
| --> Integration tests verify layer boundary contracts
|
+-- DDD (Tactical)
| --> Bounded-context-first: tests/{context}/domain/aggregates/
| --> Cross-context contract tests in tests/bounded_context_integration/| Architecture | Primary Axis | Secondary Axis | Unique Test Types | Co-located? | |-------------|-------------|----------------|-------------------|-------------| | Hexagonal | Test type | Domain concept | Port contract | No | | Clean | Test type | Architecture ring | Gateway | No | | Layered | Mirror source | Test type | Layer boundary | No (Java: yes) | | Vertical Slice | Feature | Test type | Cross-slice | Yes | | Modular Monolith | Module | Test type | Architecture, inter-module | No | | Microservices | Per-service | Test type | Contract (Pact) | No | | Event-Driven | Test type | Event flow role | Schema, idempotency, saga | No | | CQRS | Command/Query | Test type | Projection | No | | DDD | Bounded context | Building block | Aggregate, cross-context | No |
| Strategy | Best For | Strengths | Weaknesses | |----------|----------|-----------|------------| | Mirror source | Layered, Clean, Hexagonal | Predictable paths, IDE navigation, package-private access (Java) | Feature changes touch multiple dirs | | Feature co-located | Vertical Slice, React | All feature code in one place, team ownership | Cross-feature tests homeless, hard to run by tier | | Hybrid (recommended) | Most projects | Type-first for CI stages, feature-nested within | Slightly deeper nesting |
Hybrid pattern (type-first, then feature within):
tests/
unit/
features/
order/test_create_order.py
payment/test_process_payment.py
integration/
features/
order/test_order_repository.py
e2e/
test_checkout_flow.py| Language | File Convention | Discovery Rule | Co-located? | |----------|---------------|----------------|-------------| | Python (pytest) | `test_*.py` or `*_test.py` | Prefix/suffix match | Separate `tests/` (recommended) or inlined | | TypeScript/JS (Jest) | `*.test.ts`, `*.spec.ts`, `__tests__/*.ts` | testMatch pattern | Either | | Java (JUnit/Maven) | `*Test.java` in mirrored package | `src/test/java` mirrors `src/main/java` | Mirrored packages | | Go | `*_test.go` in same directory | Language-enforced co-location | Always co-located | | C# (xUnit/NUnit) | `*Tests.cs` in parallel project | Separate test project | Separate project | | Rust | `#[cfg(test)] mod tests` inline + `tests/` | Inline for unit, `tests/` for integration | Unit: inline, Integration: separate |
Place `conftest.py` at the lowest directory where fixtures apply. Pytest discovers from outermost to innermost, enabling hierarchical fixture scoping:
tests/
conftest.py # Session fixtures: DB engine, app instance
unit/
conftest.py # Unit-specific: auto-mock markers
integration/
conftest.py # Integration: real DB fixtures
acceptance/
conftest.py # Acceptance: feature file pathstests/
features/ # Gherkin .feature files by domain
order/
place_order.feature
payment/
process_payment.feature
step_defs/ # Step implementations by domain concept
conftest.py
order_steps.py
payment_steps.pyOrganize step definitions by domain concept, not by feature file. Shared steps across features prevent duplication.
| Tier | Location | Tests | Adapters | |------|----------|-------|----------| | Unit | tests/unit/ | Domain model, value objects, service layer | None (pure) or mock driven ports | | Integration | tests/integration/ | Individual adapters against real infra | Real infrastructure | | Acceptance | tests/acceptance/ | Use cases through driving ports | In-memory adapters | | E2E | tests/e2e/ | Full stack through HTTP/CLI | Real adapters |
Port contract tests: when multiple adapters implement one driven port, create shared contract test suite and run against each adapter.
tests/integration/ test_repository_contract.py # Abstract tests for RepositoryPort test_postg
AI agents that guide you from idea to working code, with human judgment at every gate. nWave runs inside Claude Code. It breaks feature delivery into seven waves (discover, diverge, discuss, design, devops, distill, deliver).
Repo: nWave-ai/nWave
Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation
Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation
Review dimensions for acceptance test quality - happy path bias, GWT compliance, business language purity, coverage completeness, walking skeleton…
Detailed 5-phase workflow for creating agents - from requirements analysis through validation and iterative refinement
5-layer testing approach for agent validation including adversarial testing, security validation, and prompt injection resistance
Architectural style selection decision matrices, trade-off analysis, structural enforcement rules, and combination patterns. Load when choosing or evaluating…