Test-Driven Development

Overview

Test-driven development is a way of writing code in very short loops: state one behaviour as an automated test, watch it fail, write just enough code to pass, then improve the design while the tests protect you. For an agent the discipline matters more than for a person, because the failing test is evidence that the test can fail at all and the passing run is evidence the change works; neither can be claimed without running something. This skill sets out the loop, what to show the user at each step, and how to apply it to new features, bug fixes and code that has no tests yet.

Instructions

1. Learn how tests run in this project

Before writing anything, find the runner, where tests live, how they are named, and the command for a single test. Read package.json, pyproject.toml, Makefile or the CI workflow; copy the project's own command instead of inventing one. Then run the whole suite once. If it is already failing, tell the user before going further: a red baseline hides whether your new test fails for the right reason.