vitest

I've been digging into how teams are handling the Vite testing ecosystem, and vitest keeps coming up as the pragmatic choice—especially given how it mirrors Jest's API but actually leverages your existing Vite config instead of duplicating it. Your skill scores a solid B (81/100), which honestly tracks with the execution, but I'm curious what drove some of your architectural choices here.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 81/100, solid B grade territory. This evaluation is based on Anthropic's best practices for agentic skills. Your strongest area is Writing Style (9/10) with clear, consistent technical voice—but Progressive Disclosure Architecture (20/30) is where you're leaving points on the table. The skill's foundation is good, but it needs better structure and error guidance to move into A territory.

What's Working Well

• Metadata triggers are solid - You've got specific, discoverable trigger phrases (writing/configuring Vitest tests, assertions, mocking, coverage). That's exactly what gets Claude to actually use this skill.
• Writing is genuinely clean - Imperative headers, no marketing fluff, just technical content. The assertions table is a nice touch for quick reference.
• Practical examples - Your Quick Start and config template show real, runnable code. The async testing and mocking patterns aren't hand-wavy.

The Big One: Missing Layered Architecture

Here's the blocker: everything's crammed into a single SKILL.md file with no external references. At 84 lines, you're still under the TOC threshold, but advanced topics (workspace setup, custom matchers, debugging) need their own homes.

Why it matters: Skills score points for reference depth (0/5 right now) and token efficiency. When Claude loads this skill, it loads *...