A Claude plugin is a packaged bundle of instructions and tools that teaches Claude a new, repeatable way of working, something you install once and then use from any relevant session after that.

What is a Claude plugin?

Claude Code, Anthropic's agentic coding tool, is available as a command line tool, a desktop app, and a web app. Claude Cowork is the desktop app built for non-developer work such as research, writing, and planning. Both support plugins. Normally, if you want Claude to follow a specific process, you explain it in the conversation, and that explanation disappears when the session ends. A plugin changes that. You install it once, and its instructions and tools stay available going forward, so you are not re-teaching Claude the same workflow every time you open a new chat.

What is actually inside a plugin?

A plugin can bundle several kinds of pieces together:

• Skills: written instructions that Claude loads automatically when a task matches what the skill covers, or that you can invoke directly as a slash command, such as /help-ux.

• Slash commands: a typed shortcut that triggers a specific skill or action right away, instead of waiting for Claude to recognize the task.

• Specialized agents: focused sub-processes that Claude can hand off part of a task to.

• Settings: configuration that ships with the plugin so it behaves correctly as soon as it is installed.

You do not need to configure any of this by hand. Installing the plugin wires all of it in at once, and Claude draws on it whenever it is relevant to what you are doing.

What is a plugin marketplace, and how does installing actually work?

Plugins are not installed one file at a time. They are distributed through plugin marketplaces. A marketplace is, at its core, a git repository, public or private, that lists the plugins available from it. You add a marketplace once. After that, you install individual plugins from it by name, and you can add more later without repeating that first step. In Claude Code, this is done with the /plugin command. In Cowork, there is a graphical install path that accomplishes the same thing for people who would rather not use a command line. Either way, once a plugin is installed, it becomes part of how Claude behaves in your sessions, not something you have to reference or re-explain each time.

What does a persona plugin add on top of a plain plugin?

A persona plugin is a plugin whose skills encode one professional role end to end, rather than a single narrow task. RolePlugin builds its plugins this way. Each persona bundles roughly three capability skills that together cover the real workflow of that role, plus a built-in help skill and a feedback skill, so the plugin can explain itself and take input from you directly. RolePlugin sells a library of 20 persona plugins for roles such as Business Analyst, Product Manager, Technical Writer, UX Researcher, Financial Analyst, Recruiter, and Customer Success Manager. Every persona goes through the same process before it ships: research with at least eight sources, an approved spec, two rounds of evaluation, adversarial review, and a human sign-off. You can browse the full list of personas to see what each one covers.

How do you try one?

A RolePlugin subscription includes a private plugin marketplace, hosted as a private GitHub repository, plus zip downloads through the members app at app.roleplugin.com. One subscription covers the whole library, not a single persona at a time. Pricing is $19 a month or $180 a year for an individual, and $24 per seat a month for a team, with a five seat minimum. Once installed, a persona plugin runs inside your own Claude subscription: RolePlugin does not proxy or store your conversations. You can compare options on the plans page, then add the marketplace and install a persona to see how it works in your own Claude Code or Cowork session.