General FAQ

What is a persona?

A RolePlugin persona is a Claude plugin built around a specific role, like a business analyst or a financial analyst. It packages that role's knowledge, workflow, and guardrails so Claude can do the work the way a practitioner would, without you re-explaining context every time. See Introduction.

Do I need a GitHub account?

Only for the marketplace route. If you would rather not connect GitHub, download the persona zip directly from your library and extract it into ~/.claude/skills/. Both routes install the same persona. See Installing personas.

Can I use more than one persona at a time?

Yes. Install as many as your subscription includes. Claude routes each request to the persona whose capabilities match it, see Auto-activation.

Is my data used to train anything, or shared with other customers?

Personas themselves are generic by design: they ship with no client names, no proprietary data, and no real personal information baked into their instructions, that is checked before every release (see Safety & guardrails). What you type into your own Claude Code sessions is your conversation with Claude, not something other RolePlugin customers can see. RolePlugin does not insert other customers' content into your sessions, or vice versa.

Does a persona ever take action on its own, like sending an email or posting a comment?

No. Anything outward-facing or hard to undo stops for your explicit confirmation first, personas draft, you decide. See The confirmation gate.

What happens if I ask for something a persona doesn't cover?

Claude either answers normally without the persona, or the persona's hard input gate asks clarifying questions if your request is close to something it covers but under-specified. It will not fabricate an answer in a role it was not built for.

How do I know a persona's exact capabilities?

Check its detail page in your library, or the relevant category page in the persona library in these docs. Every persona also has a built-in help guide, see Help guides.