A plugin bundles skills, tools, and connectors into one installable package — and the official marketplace already has 100+. Here's how to browse, install, and vet them as a non-developer.
A single skill teaches Claude one way of working. A plugin goes bigger: it bundles skills, tools, and connectors together into one package built around a job or an app. Instead of adding capabilities one at a time, you install the whole bundle and get everything it includes.
Think toolbox, not tool. If a skill is a wrench, a plugin is the full kit for a particular kind of work.
Plugins live in marketplaces — curated catalogs you can search. The official one already has around 100 plugins: roughly a third built by Anthropic and the rest from trusted partners like GitHub, Figma, Linear, Supabase, and Stripe. You search by the job you're trying to do, not by reading code.
It's effectively an app store for your AI assistant: browse, read what each one does, and pick the ones that match your work.
Installing is the easy part. You pick a plugin, add it, and it's ready — no setup files to edit, no code to write. The capability simply shows up and Claude can use it. For a non-developer this is the whole appeal: power without configuration.
Add what a real task needs rather than collecting plugins for their own sake. A lean, intentional set beats a cluttered one.
Browse the marketplace -> pick a plugin -> install -> it's ready. No config files. No code.
A plugin can do more than a single skill, so it deserves the same care — a bit more, if anything. Prefer official plugins and known partners, check the permissions it asks for like you would any download, and start with one so you can see how it behaves before adding more.
Same five-check instinct as installing a skill: source, maintainer, permissions, network, and a careful first run.
A short, sane list of plugin categories to look at first — plus how to vet any plugin before you install it.