Testing Best Practices

This is Yoni Goldberg's opinionated testing rulebook condensed into 40+ rules for AI assistants. It's built around six critical constraints, like keeping tests under 10 statements and never mocking internal application parts. The smoking gun principle appears repeatedly: any data you assert on must first appear in the arrange phase so readers can trace cause to effect. You also get specific guidance on data factories with faker, assertion minimalism, and the breadcrumb principle about keeping test dependencies visible. Use this when you want tests that read like simple stories rather than systems to maintain. It's prescriptive enough that you'll either love the constraints or find them too rigid for your context.