Codified practices for AI agents — teach your agents the way.
See the Manifesto for why this exists.
| Concept | Kata term | Description |
|---|---|---|
| Skills | kata | Codified practices agents follow |
| Rules | dojo kun | Always-on constraints applied to every session |
| Roles | sensei profiles | Agent role definitions with scoped permissions |
| Prompts | kata forms | Standardized templates for commits, PRs, reviews |
| ADRs | dojo records | Architecture decisions preserved for the school |
| MCP server | the dojo | Where kata are served and practiced |
| Framework | the ryu (school) | The complete system of practices |
Kata in this repo are platform-agnostic. They follow the Agent Skills specification and work with any agent that supports it — Kiro, Claude Code, OpenCode, or others. Skills avoid coupling to specific tooling, MCP servers, or platform-specific features so they remain portable across environments.
The fastest way to use kata is via the MCP server — add
https://openkata.dev/mcp to your agent config. Zero install,
always up-to-date.
For offline or local-first workflows, download and extract
skill folders into your project's .agents/skills/<name>/
directory. Your agent discovers them automatically.
- Skills live in
skills/, rules inrules/, profiles inprofiles/ - See the full catalog for all available artifacts
- See Getting Started for client-specific setup (Kiro, Claude Code, OpenCode, and others)
See CONTRIBUTING.md for how to add skills and rules. See AGENTS.md for build commands, code style, and commit conventions. See RELEASING.md for how to publish skills and deploy.
MIT — see LICENSE.