What is SLOBAC?¶
The Suite Life of Bobs and Code is a point of view about what tests should be, what test suites should be, and the named ways both go wrong.
SLOBAC ships several things:
- A Manifesto: a self-contained read. No software is required to use it. A reader who finishes these pages should be able to audit their own test suite by hand and classify real tests against the named failure modes.
- Principles — what a test should be, and the bounds a disciplined test-suite refactor must respect.
- Taxonomy — a catalog of ways tests and test suites can go wrong.
- An Agent Skill: Instructions for AI agents to audit a codebases's test suite against the manifesto.
What This Is Not¶
- Not a linter. Where existing ecosystem tools already lint syntactically, SLOBAC defers.
- Not a mutation engine, codemod runner, or test generator — those are mature; SLOBAC orchestrates them, never reimplements.
- Not a smell-count scoreboard. Optimizing for raw smell counts is an explicit anti-goal (Panichella et al., Test Smells 20 Years Later, EMSE 2023).
How To Start¶
- Read the manifesto:
- Principles — what a test should be, and the bounds a disciplined test-suite refactor must respect. This is the vocabulary every other page cites.
- Workflows — the RED-GREEN-MUTATE-KILL-REFACTOR cycle that the taxonomy's prescribed fixes assume. You are doing mutation testing, right? No? Ah, well, you can still use SLOBAC, no worries.
- Taxonomy — the catalog of named failure modes, one entry per smell. Start with
taxonomy/README.mdfor the curated reading order across the 15 entries; each entry is a standalone page you can jump to directly. - Glossary — definitions and citations for terms the other pages rely on (mutation testing, kill-sets, tautology theatre, canonical location, etc.).
- Install the audit skill:
- Running the SLOBAC audit — instructions for installing and invoking the audit skill.
- Have your AI agents audit your test suite:
/slobac-auditand it'll walk you through the process.