What problem does it solve? Teams often write tests by habit rather than by risk, ending up with brittle end-to-end suites, coverage metrics that hide real gaps, and flaky tests nobody owns. This Skill provides a language-agnostic framework for deciding what to test, at which level, and whether a test is worth writing at all. ## Core Features & Use Cases - Risk-Based Test Selection: A decision rubric maps risk types (pure logic, integration boundaries, critical user flows, bug fixes, refactors) to the cheapest reliable test level. - Flakiness and Suite Health Signals: Evidence signals for detecting unreliable suites, including pass/fail frequency, retry dependence, quarantines, and duration variance. - Review Checklist: Concrete questions for evaluating whether a test would catch a meaningful regression and survive harmless refactoring. - Use Case: When reviewing a pull request that adds a broad browser test for simple validation logic, use this Skill to justify moving the assertion to a unit test and reserving E2E coverage for the critical user workflow. ## Quick Start Ask the AI to evaluate the testing strategy for your current change and recommend which test levels to use based on risk.
