SkillWorks
SkillWorks is a free index that automatically tests thousands of Claude Code skills, subagents, and plugins every night and scores whether they actually still work.
🔗 Visit SkillWorks
Description
Anyone who's browsed a big list of AI plugins or extensions knows the problem: half the entries are abandoned, broken, or just don't do what the description claims, and there's no easy way to tell which is which before you waste time installing one. SkillWorks solves that specifically for the Claude Code ecosystem — it doesn't just list skills, subagents, plugins, and marketplaces, it actually runs them every night to check whether they still work.
Technically, SkillWorks evaluates over 503,000 listings through automated nightly rebuilds, scoring each one on whether it Works (40% of the score), is Maintained (25%), is Adopted (20%), and is Documented (15%). It currently tracks more than 48,000 broken artifacts with specific, stated failure reasons rather than a vague "unmaintained" flag, integrates with the skills.sh registry, and monitors daily GitHub commit activity to catch when a previously-working skill breaks. It's built by Kynth Studios, which also runs Toolproof, BlockDex, and KitGrade, and it's free to use with optional sponsorship — the site explicitly states sponsorship has zero influence on rankings.
💬 Our review
The short version: if you use Claude Code skills, subagents, or plugins and are tired of installing something from a marketplace listing only to find it's broken or abandoned, SkillWorks' actually-runs-it verification is a real time-saver, and since it's free, there's no reason not to check a skill's score before installing it.
The honest caveat is that this is a very young, narrow tool serving one specific ecosystem (Claude Code), so its usefulness is entirely tied to how much of your workflow depends on that marketplace — it has no relevance outside it. There's also an inherent trust question with any self-reported scoring system: SkillWorks states sponsorship doesn't affect rankings, but as with any young index, it's worth spot-checking its verdicts against your own experience with a skill rather than treating the score as infallible, especially since the evaluation methodology (nightly automated runs) can't fully replace real-world usage testing.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Index gratuit, options de sponsoring optionnelles sans impact déclaré sur le classement. Pas d'abonnement.
Pros
Vérifie réellement le fonctionnement des skills (pas juste un annuaire statique)
Score composite : fonctionnement, maintenance, adoption, documentation
Suit plus de 48 000 artefacts cassés avec raison précise
Gratuit, sponsoring sans influence déclarée sur le classement
Cons
Utile uniquement pour l'écosystème Claude Code — zéro pertinence ailleurs
Outil jeune, méthodologie de scoring auto-déclarée à vérifier dans le temps
L'exécution automatisée ne remplace pas totalement un test d'usage réel