# What To Test > Use when writing tests, improving coverage, or deciding what to test in this CLI - Skill: `tomevault-io/what-to-test` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/what-to-test` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/what-to-test/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/what-to-test --- # What to Test ## Philosophy Test user-facing behavior. If a user would notice it's broken, it needs a test. All CLI commands must be tested - including commands marked "internal". Internal commands are still user-facing (Matt uses them daily). ## What Makes a Good Test - Tests behavior users depend on - Validates real workflows, not implementation details - Catches regressions before users do Do NOT write tests just to increase coverage numbers. Use coverage as a guide to find untested user-facing behavior. ## What NOT to Test Use `/* v8 ignore start */` for: - Integration boundaries (actual git calls, shell execution, filesystem) - Service layers that get mocked in tests - Entry points and DI wiring - Boilerplate, unreachable error branches ## Testing Pattern 1. Export a `runX()` function from command files for testability 2. Mock external services (GitService, PromptService) 3. Test error paths users will actually hit 4. Use v8 ignore for CLI formatting/presentation code --- > Source: [mattpocock/ai-hero-cli](https://github.com/mattpocock/ai-hero-cli) — distributed by [TomeVault](https://tomevault.io).