One install. A whole team.
Everything lands in your .claude/ folder and Claude Code picks it up automatically.
One commandNo copy-pasting, no global install
A team you call by name102 specialists across 6 kits
Updates that respect your workFingerprinted, never clobbered
Pick the teams you need.
Every kit is a set of Claude Code agents, skills, and slash commands. Start with one, or install the whole roster.
Build
Full-stack engineering: tech-lead, backend-architect, react-specialist, code-reviewer, security-auditor, plus DevOps, testing, and RAG/agent-design skills.
Grow
Content, SEO + AEO (answer-engine optimization), CRO, growth, sales intelligence, and a landing-page generator — same context as your code.
Product
PM toolkit with RICE prioritizer, code-to-prd, roadmap and discovery, UX research, plus scrum, Jira, and Confluence automation via the bundled Atlassian MCP.
Audit & Compliance
SOC 2, ISO 27001, GDPR, MDR/FDA, and CAPA specialists, plus the skill-security-auditor that gates any skill for malicious code before install, with evidence and audit-readiness workflows.
Operator
Founder-mode C-level advisors (CEO/CTO/CFO/CMO/CRO/COO…), pricing and deal-desk, DCF and SaaS-metrics finance, RevOps, vendor management, and procurement.
Research & Focus
Academic and enterprise research orchestrators (litreview, market/product research), plus daily-driver productivity: capture, inbox triage, weekly review, and deep-work planning.
Three steps to a full team.
Install
Run npx rosterkit-cli init --kit both. It fetches from your private repo using your own git credentials — nothing global, nothing to configure.
Teach it your project
Fill in the generated CLAUDE.md once — your stack, conventions, and brand voice. Every agent reads it before doing anything.
Ship
Open Claude Code and go: use tech-lead to build auth or /blog-post our launch. That's the whole workflow.
Free while we're in early access.
We're onboarding our first testers by hand. Request access to any kit and we'll send you a GitHub invite and the install command — free while we're in early access.
The Whole Roster
All six kits — 362 skills, 102 agents, 116 commands. Every team, every workflow, one install.
- Build · Grow · Product
- Audit & Compliance · Operator · Research & Focus
- Both CLI + all CLAUDE.md templates
- Private repo + lifetime updates
Free for early-access testers · prices shown are launch pricing once we open to everyone
Good to know.
Do I need Claude Code? Does it work with Cursor?
Roster is built specifically for Claude Code. The kits are agents, skills, and slash commands that live in your project's .claude/ folder. It doesn't run inside Cursor, Copilot, or the ChatGPT app. Any Claude Code plan works — Pro, Max, Team, or API.
What exactly do I get?
Access to a private repo with your kit(s): a .claude/ directory of agents, skills, and commands, CLAUDE.md templates, the roster installer CLI, and docs. Drop it into any project with npx rosterkit-cli init.
What happens to my files when I update?
Roster fingerprints every file it installs. On roster update, files you never touched refresh to the latest version; files you edited are detected and left exactly as they are. Your CLAUDE.md is never modified after it's first created. Pass --force only if you want kit content to win.
Does it lock me into a framework?
No. Agents adapt to your stack — Next.js, Django, Rails, Go, Rust, anything — by reading your CLAUDE.md. There's no forced architecture.
Where does the content come from?
Roster is a curated distribution. Components are sourced from open-source projects under MIT/Apache-2.0 licenses, with full attribution preserved in a CREDITS.md file in your repo. You're paying for the curation, the installer, and lifetime updates.
Is this affiliated with Anthropic?
No. Roster is an independent, unofficial product and is not affiliated with, endorsed by, or sponsored by Anthropic. "Claude" and "Claude Code" are trademarks of Anthropic.