Generate Comprehensive Test Suites

The Test Generator is an autonomous agent (Claude sonnet, tools: Read, Glob, Grep, Write, Edit) that creates comprehensive, well-structured test suites. It first analyzes the code - the function or class interface, input parameters and return types, dependencies and side effects, edge cases and boundary conditions - then generates tests across four categories: Happy Path (normal, expected usage), Edge Cases (boundary values, empty inputs, max values), Error Cases (invalid inputs, exceptions, error handling), and Integration (interactions between components). Every test follows the Arrange-Act-Assert structure: Arrange sets up the test data, Act executes the code under test, and Assert verifies the result.

Use it when you have existing code that needs a test suite written for it and want coverage across happy-path, edge-case, error-case and integration scenarios rather than just the obvious cases. Its guidelines are explicit about testing behavior, not implementation, using descriptive test names, one assertion per test where possible, independent tests, fixtures/factories for test data, and mocked external dependencies - so it is aimed at conventional unit/integration test suites, not exploratory or manual QA scripts.

Input is the code to be tested, read via Read/Glob/Grep, with tests written back via Write/Edit. Output format depends on the stack. For Ruby/Rails:

require "test_helper"

class UserTest < ActiveSupport::TestCase
  setup do
    @user = users(:valid)
  end

  test "should be valid with all required attributes" do
    assert @user.valid?
  end

  test "should require email" do
    @user.email = nil
    assert_not @user.valid?
    assert_includes @user.errors[:email], "can't be blank"
  end

  test "should validate email format" do
    @user.email = "invalid"
    assert_not @user.valid?
  end
end

For JavaScript/TypeScript it produces an equivalent suite, nesting describe blocks by class and method (e.g. describe('UserService') > describe('createUser')) with it(...) cases such as "should create a user with valid data" asserting expect(user.name).toBe('John'), and "should throw error for invalid email" asserting await expect(UserService.createUser(userData)).rejects.toThrow() for the async, error-path case.

It does not run the tests it generates or wire them into CI itself - it only writes the test file in the target framework's own conventions, matching the syntax and structure of whichever stack the code under test uses. Running the suite or wiring the new file into a CI pipeline stays a separate step for the developer.

Developers who want a first-draft, ready-to-run test suite generated from existing code - covering happy path, edge cases, error handling and integration points in one pass - rather than writing every individual test case by hand.