test-standards
Testing best practices and anti-patterns. Use when writing tests, reviewing test quality, or understanding testing methodology for this project.
When & Why to Use This Skill
This Claude skill provides a comprehensive framework for software testing standards, focusing on best practices and the identification of common anti-patterns. It helps engineering teams improve code quality by enforcing principles such as test validity, comprehensive coverage, and maintainable methodology, ensuring that tests verify requirements rather than implementation details.
Use Cases
- Writing New Tests: Use as a reference to ensure new unit and integration tests follow proper isolation, mocking, and setup/teardown procedures.
- Code Reviews: Evaluate the quality of peer-submitted tests to identify and eliminate anti-patterns like logic duplication or testing private methods.
- Coverage Optimization: Identify and fill gaps in testing suites, including edge cases, error scenarios, and critical boundary points between client and server.
- Technical Onboarding: Provide new developers with a clear, standardized methodology for the project's testing expectations and maintainability goals.
| name | test-standards |
|---|---|
| description | Testing best practices and anti-patterns. Use when writing tests, reviewing test quality, or understanding testing methodology for this project. |
Test Standards
Refer to test-standards.md for detailed testing guidelines.
Key Principles
- Test Validity - Tests verify requirements, not implementation details
- Coverage - Happy path, edge cases, error scenarios, integration points
- Methodology - Appropriate mocking, test isolation, clean setup/teardown
- Maintainability - Readable tests, minimal duplication, clear test data
Anti-Patterns to Avoid
- Logic duplication (re-implementing production logic in tests)
- Module-level mocking (prefer fetch-level mocks)
- Private method testing (test through public interface)
- Boundary testing gaps (client-server, null vs undefined)
- Form testing gaps (schema-only tests are insufficient)