test-generation

name	Test Generation
description	Generate comprehensive tests with good coverage strategy. Use when writing unit tests, integration tests, or test plans for existing or new code.

Test Generation

You are writing tests. Focus on testing behavior and edge cases, not implementation details. If the task involves multiple test cases, fixtures, or validation steps, start with a todo plan before writing tests. Use it to track analysis, implementation, and verification rather than keeping the plan implicit.

Available Tools

• read — read the code under test and existing test files
• search — find related tests, test utilities, mock patterns in the codebase
• execute — run tests to verify they pass, check coverage
• edit — write new test files or add test cases
• todo — track test case creation progress

Process

1. Plan Test Coverage

Use read to analyze the code under test, then plan with todo:

todo({ todoList: [
  { id: 1, title: "Analyze code and map branches", status: "in-progress" },
  { id: 2, title: "Check existing test patterns", status: "not-started" },
  { id: 3, title: "Write happy path tests", status: "not-started" },
  { id: 4, title: "Write edge case tests", status: "not-started" },
  { id: 5, title: "Write error path tests", status: "not-started" },
  { id: 6, title: "Run and verify all tests pass", status: "not-started" }
]})

2. Analyze Code Under Test

• Use read to study the function/module thoroughly
• Map decision branches (if/else, switch, error throws, early returns)
• Identify external dependencies (DB, APIs, file system)
• Use search to find existing test utilities, mocks, and fixtures

3. Check Existing Patterns

• Use search to find how tests are written in this project
• Match the existing test framework, assertion style, and naming conventions
• Reuse existing test helpers and mock setups

4. Write Tests

Use edit to create test files. Cover:

Happy path: Normal usage with typical inputs Edge cases: Empty inputs, nulls, boundary values, single-element collections, unicode Error paths: Invalid inputs, missing fields, service failures, timeouts State transitions: Initial → after action → expected state (for stateful code)

5. Verify

• Use execute to run the new tests
• Confirm all pass
• Check that error path tests actually test the error (not just "no crash")

Rules

• Update todo as you complete each test group
• Treat todo as mandatory for multi-step test work; only skip it for truly tiny one-test changes
• Each test should test one behavior with a clear descriptive name
• Use realistic test data, not "test" or "foo"
• Mock at boundaries, not deep internals
• Don't test private implementation details that may change