Documentation
NZT documentation: what it is, how the method works, how to use it with Claude Code and Codex CLI, how to install and uninstall it, and the CLI reference.
Everything you need to understand NZT, install it and work with it.
Introduction
What is NZT
NZT is a working method for coding agents. It gives the model shape through method, memory and brakes so it performs steadily, from idea to product.
Why NZT
A coding agent can get to the wrong place very fast. See what changes with NZT: analysis before code, reviewable units and honest reports of what happened.
Concepts
Architecture
Three layers, and only what is needed gets loaded. How NZT routes each request from the kernel to a phase router, and from there to the one leaf the task needs.
The cycle
Understand, propose, execute, verify and record. The NZT five-step cycle is always the same; what changes with the size of the request is how deep it goes.
Brakes and state
The brakes are a feature. Learn the three NZT STOPs and how Plan/state.json keeps the work so a new session can resume without rebuilding the chat history.
The seven phases
Discovery, Architecture, UX, Build, Verify, Ship and Learn. What each NZT phase covers, and why a phase comes in only if it changes what is going to be built.
Stack layer
The .NET 10, Blazor, EF Core and Minimal APIs skills are always installed, but NZT applies them only when the component's stack document has chosen them.
Artifacts and guardrails
Decisions and evidence outside the chat. The artifact each NZT phase leaves, where it lives in your repo and the rules that keep the agent from fooling itself.
What NZT does not do
NZT does not decide for you. What the method does not do: it does not limit the model, use subagents or MCP, guarantee perfect routing or replace your judgment.
Usage
Getting started
How to use NZT once it is installed. Open your agent, ask for what you want to build, approve the plan and move forward one unit at a time. No new commands.
Example requests
Real requests for each moment of a project with NZT: the kickoff, the answers to discovery, the plan approval, the verification and the learning mode.
Providers
NZT in Claude Code
How NZT works in Claude Code: where the instructions and the skills are installed, and how it loads on its own every time you open claude in any project.
NZT in Codex CLI
How NZT works in Codex CLI: where AGENTS.md and the skills go, what happens when you change CODEX_HOME and what to check first if Codex does not see the skills.
Installation
Install NZT in Claude Code
Guide to installing NZT in Claude Code on Windows with a single PowerShell command: requirements, what it installs and where, how it loads and how to update it.
Install NZT in Codex CLI
Guide to installing NZT in Codex CLI on Windows with a single PowerShell command: requirements, where AGENTS.md and the skills go, loading and updating.
Uninstall NZT
How to uninstall NZT from Claude Code and Codex CLI. It removes only what its manifest lists and keeps your own content in CLAUDE.md and AGENTS.md intact.
Reference
CLI reference
Reference for the NZT CLI commands: install, uninstall, status, verify and the interactive menu, with their --provider and --dry-run options explained.
FAQ
Common questions about NZT: whether it improves the code, applies to all your projects, works outside .NET, overwrites CLAUDE.md, and how to update it.