For Claude Code · Any stack · License-audited

An engineering team
for Claude Code.

The open-source Claude Code ecosystem is huge, and mostly free. Kaairos is the curated, license-audited slice worth keeping — five kits covering engineering, security, configuration, DevOps and AI engineering, each installed with one command and never overwriting a file you own.

5Kits
321Components
1Install command
0Files overwritten
What you're buying

It started as curation.
It became a rewrite.

Kaairos began as an audit of the open-source Claude Code ecosystem. Reading it closely made the problem obvious — most components were a handful of lines any capable model already follows. So they were rewritten. Of 321 components, 292 are original: written to the depth a senior practitioner would actually give you, naming the commands, thresholds, and failure modes they would say out loud.

Some keep a name or a scope that came from an open-source project, and 29 components are open-source work shipped intact. Every one is credited by source, commit, and license in a CREDITS file you can read before you trust it. Permissive licenses only.

No secret techniques. Depth, licensing hygiene, and maintenance.

What's inside

One kit.
Five departments.

A full team configuration, not a pile of prompts. Each department is a set of agents, skills, and commands that already know how to hand work to each other.

01

Planning

A tech-lead that scopes and delegates, plus orchestrators that chain the rest — sequential, parallel, or reviewed.

02

Build

Backend and frontend architects with language specialists: TypeScript, Python, Go, Rust, SQL.

03

Data & infra

Postgres, Docker, Kubernetes, Terraform, CI — the unglamorous half of shipping.

04

Quality

Code review, test automation, debugging, security audit, performance — the gate before anything merges.

05

Commands & skills

Slash commands for scaffolds, tests, deploys, and docs; skills for the frameworks you actually use.

Three building blocks

Agents, skills,
and commands.

Written and audited down to the set that earns its place — every component follows one schema and passes the license gate. Agents you direct by name, skills Claude loads on its own, commands you type.

AGENTS
50

Focused specialists you direct by name — planning, backend, frontend, data, and quality roles.

SKILLS
37

Framework and workflow know-how Claude reaches for automatically, no prompting required.

COMMANDS
65

Scaffolds, tests, reviews, deploys, and docs — typed straight into the prompt.

How it works

Three steps. The third
one is the product.

01

Install

Run one npx command from your private repo's README. It copies the kit into .claude/ and never overwrites a file you already have.

02

Teach it your project

Fill in one CLAUDE.md — stack, conventions, don'ts. Every agent reads it first, so output matches your codebase, not a template.

03

Direct it

“Use tech-lead to plan auth.” /test-coverage. “Have security-auditor review the diff.” That's the whole interface.

Cherry-pick instead: kaairos add agent debugger. Update any time: kaairos update. Your files are never touched.

New · Security Kit · $49

Ship AI code without shipping
AI vulnerabilities.

Roughly 45% of AI-generated code fails security tests, and AI-written code now factors into about one in five enterprise breaches. Claude Code ships a free security-review command — a real start. This is the layered guardrail system around it, license-audited to the same standard as the kit.

12 agents · 12 skills · 12 commands · 5 guardrail hooks · hardened settings

  • GUARDRAIL HOOKSBlock secret leaks, dangerous commands, and prompt injection before they run.
  • REVIEW & THREAT MODELAgents read your diff for injection, authz, crypto, and supply-chain risk.
  • SCAN WORKFLOWSOrchestrates your own scanners — Semgrep, gitleaks, npm audit, Trivy. Nothing bundled.
  • HARDENED SETTINGSSafe-by-default permissions you opt into. Your files are never overwritten.
Provenance

Every source,
named and licensed.

These projects seeded Kaairos. Most of what they contributed has since been rewritten from scratch, but their licenses are preserved either way, and the exact source and commit for every component stays logged in the CREDITS file inside the kit.

obra/superpowersanthropics/skillsanthropics/defending-code-reference-harnessdwarvesf/claude-guardrailswshobson/agentswshobson/commandsbrennercruvinel/CCPluginscentminmod/my-claude-code-setup

Try before you decide — a free sampler of 8 agents, 5 skills, and 10 commands is public on GitHub . Same formats, same hygiene.

Pricing

Five kits.
One bundle.

One-time, no subscription, no seats. Buy once, install everywhere, updates included. Each kit is delivered as a private GitHub repo the moment you pay.

Engineer Kit$39The full team.
  • 50 agents, 37 skills, 65 commands
  • Planning, build, data & infra, quality
  • Installer CLI with cherry-pick + update
  • CLAUDE.md template
  • Full CREDITS + license files
Get Engineer Kit
Security Kit$49Ship it safely.
  • 12 agents, 12 skills, 12 commands
  • 5 guardrail hooks + hardened settings
  • Secret / injection / dangerous-command blocking
  • Secure-code review + threat modelling
  • Orchestrates Semgrep, gitleaks, Trivy
Get Security Kit
Power-ConfigNEW$29Tune the tool itself.
  • 6 productivity hooks (format, lint, test, notify)
  • Two settings profiles: flow and review
  • Statusline + 3 output styles
  • Vetted MCP set + plugin packaging
  • 4 skills, 5 commands
Get Power-Config
DevOps KitNEW$39Run it in production.
  • 10 agents, 9 skills, 17 commands
  • Incident command, postmortems, runbooks
  • SLOs and error-budget policy
  • GitOps, canary, rollback safety
  • Kubernetes triage, Terraform state, FinOps
Get DevOps Kit
AI EngineeringNEW$39Build LLM products.
  • 13 agents, 24 skills, 18 commands
  • RAG, chunking, retrieval evaluation
  • Eval harnesses + LLM-as-judge
  • Structured output, tool calling, agents
  • Tracing, guardrails, cost control
Get AI Engineering

Secure checkout via Polar · digital product, delivered instantly and in full — all sales final. The Complete bundle grants access to both private repos.

Questions

Before you buy.

Yes. Kaairos is built specifically for Claude Code — the kit lives in your project's .claude/ folder, which is a Claude Code format. Any plan works: Pro, Max, Team, or API billing. It does not run inside Cursor, Copilot, or the ChatGPT app.

Access to a private GitHub repo containing the kit, the installer CLI, the CLAUDE.md template, and docs. The exact one-line install command is in that repo's README. Node 18+ is the only requirement.

No. Agents adapt to your stack — Next.js, Django, Rails, Go, Rust, anything — by reading your CLAUDE.md. No forced architecture, no template code.

Run kaairos update (or git pull) whenever you like. Updates are included with your purchase. Your own files and your CLAUDE.md are never modified.

This is a digital product delivered instantly and in full at checkout, so all sales are final. Unsure whether it fits? Email hello@kaairos.com before buying — replies within a week.

No. Kaairos is an independent, unofficial product, not affiliated with, endorsed by, or sponsored by Anthropic. “Claude” and “Claude Code” are trademarks of Anthropic.