Unit tests¶

R-UNIT-1 - Deterministic logic has meaningful assertions¶

Rule. Deterministic logic must have tests that assert outputs, state transitions, errors, or observable side effects for representative inputs.

Why. Logic with no assertions over behavior can break while the suite stays green.

vcqa. Flag modules with exported functions/classes and no nearby tests, and flag tests whose assertions only check existence for behavior-bearing code.

References.

• Vitest Guide: https://vitest.dev/guide/

R-UNIT-2 - Tautology and empty tests are failures¶

Rule. Tests must not be empty, tautological, or unrelated to the unit under test.

Why. expect(true).toBe(true) and empty test bodies increase confidence metrics while proving nothing. This is a common failure mode in AI-generated test suites.

vcqa. Flag empty it/test bodies, assertions over constants, tests with no expectation or awaited failure, and tests whose only assertions are tautologies.

References.

• Vitest Guide: https://vitest.dev/guide/

R-UNIT-3 - Error and edge paths are represented¶

Rule. Unit tests for parsing, validation, transformation, permission, and branching logic must include at least one error or edge case for each material branch.

Why. Happy-path-only tests miss the branches that most often fail under real input.

vcqa. Flag behavior-bearing test files where all inputs look like happy-path fixtures and no assertion covers thrown errors, invalid input, boundary values, or rejected promises.

References.

• Vitest Guide: https://vitest.dev/guide/