Test

Who Test is for

For developers who intend to write tests and keep not doing it, and specifically for legacy code with no coverage where the blocker is deciding what to test first rather than writing the assertions.

What Test does

Test completes the development trio — Code writes it, Debug fixes it, Test verifies it. Four modes cover different needs: unit tests from code, a test plan from a specification, a coverage matrix showing what is and is not exercised, and a regression suite around a bug you just fixed. Prompts are calibrated per framework rather than generic, because pytest and Jest disagree about almost everything. Generated suites can feed Autoquiver for browser execution.

• →Unit, plan, matrix and regression modes
• →Framework-aware output
• →Works from code or from a specification
• →Feeds Autoquiver for browser execution

How Test works

1. 01

   Choose the mode

   Unit tests, a test plan, a coverage matrix, or a regression suite. Different problems, different outputs.

2. 02

   Supply code or a spec

   Both work. A spec produces a plan; code produces tests you can run.

3. 03

   Set the framework

   Output is shaped to it, because pytest and Jest disagree about nearly everything that matters.

4. 04

   Review before trusting

   A generated test that asserts the wrong thing passes forever and hides the bug it was meant to catch.

What Test does not do

Test generates tests from what you show it and cannot know your real edge cases, business rules or integration boundaries. Generated tests need review — one that asserts incorrect behaviour will pass forever and actively conceals the bug, which is worse than having no test. It does not execute anything or measure real coverage, and the coverage matrix reflects the code you supplied rather than an instrumented run.

Common questions

Works well with