Guides red-green-refactor test-driven development in Stacks projects using bun test.
npx skills add https://github.com/stacksjs/stacks --skill stacks-tdd
Please help me install this Agent Skill. Skill: stacks-tdd Source: https://github.com/stacksjs/stacks/tree/main/storage/framework/defaults/ai/skills/stacks-tdd Command: npx skills add https://github.com/stacksjs/stacks --skill stacks-tdd
Writing tests that survive refactors and actually verify behavior is hard. This Skill enforces the discipline of test-driven development in a Stacks project: choosing the right seam to test at, avoiding anti-patterns like tautological assertions and implementation-coupled mocks, and running the red-green loop correctly with bun test and @stacksjs/testing.
bun test on the narrow path, implement the minimal code to go green, then hand off to code review.Ask the AI to build a new feature test-first in your Stacks project using the red-green-refactor loop with bun test.
High-intent search queries and answers about installing and using this skill.
Write a failing test at an agreed seam first, then implement only enough code to pass it. Run the narrow test file with bun test during the loop, typecheck with ./buddy typecheck, and work in vertical slices of one test and one implementation per cycle.
A seam is the public boundary where you observe behavior without reaching inside. Stacks defines four seams from highest to lowest: route, action, model, and function. Prefer the highest seam that can go red on the behavior you care about.
No. Use setupDatabase() to create the test SQLite database and refreshDatabase() to re-migrate it between tests. Mocking User.find() only tests the mock. Reserve stand-ins for third-party HTTP, time, randomness, and queue dispatch assertions.
Use fake() from @stacksjs/queue, perform the action, then call queue.assertDispatched('JobName'). Always call restore() in the same test or in afterEach, because fake() mutates global state and leaks into other tests otherwise.
A tautological test recomputes the expected value the same way the implementation does, so it passes by construction and can never catch a bug. Expected values must come from an independent source such as a known-good literal or a worked example.
Run the narrow file with bun test during the red-green loop for fast feedback. Run the full buddy test suite once at the end as a gate, and run /stacks-review after the slice is green.
involvexConfigure Bun's test runner for Jest-compatible tests in JavaScript and TypeScript projects.
afel6Automates pre-commit verification of staged Git code with security scans, tests, linting, and JSON reviewer verdicts.
ian-pascoeDesign deep modules and clean interfaces for codebases.
GitnappValidate uncertain technical ideas through focused throwaway experiments before production implementation.