Stop fighting your AI. Start shipping with it.
NEZAM is an open-source workspace kit that turns AI-assisted coding into a structured, repeatable system — from idea to production.
The Problem • Quick Start • Commands • How It Works • AI Sync • Design Hub
You use Cursor, Claude, or Gemini every day. But at some point you notice:
- The AI forgets context between sessions
- It generates code that breaks something else
- You have no record of why a decision was made
- Different team members get different results from the same prompt
- Design and code slowly drift apart
This is not an AI problem. It is a missing structure problem.
NEZAM gives you that structure.
NEZAM is a set of rules, commands, and agents that sit on top of your existing AI tools. It does not replace Cursor or Claude — it makes them work in a controlled, predictable way.
Here is what you get:
| Without NEZAM | With NEZAM | |
|---|---|---|
| Starting a project | Prompt → hope for the best | /START all → structured onboarding, PRD locked |
| Design decisions | Scattered in chat history | Locked in DESIGN.md, always visible |
| Development | AI writes whatever it wants | Gated by wireframes and specs — no drift |
| Multi-tool teams | Everyone has different settings | One source in .cursor/, synced to all tools |
| Releases | Manual changelog, easy to miss things | /DEPLOY generates everything automatically |
You need: Node.js ≥ 20, pnpm ≥ 9, Git
# 1. Clone
git clone https://github.com/iDorgham/Nezam.git
cd Nezam
# 2. Install
pnpm install
# 3. Start inside Cursor, Claude Code, or VS Code
# Type this in the AI chat:
/START allThat is it. NEZAM will walk you through the rest.
Already have a project? You can copy the
.cursor/,.nezam/, and.claude/folders into your existing repo and runpnpm install.
Type these commands directly in your AI chat (Cursor, Claude Code, Gemini CLI).
| Command | What It Does |
|---|---|
/START |
Set up the workspace. Lock your PRD and design profile. |
/PLAN |
Break your project into phases: research → design → build → ship. |
/WIREFRAME |
Set your page layouts. Locks the structure before any code is written. |
/DESIGN |
Apply a design profile. Generates your color, type, and spacing tokens. |
/DEVELOP |
Build. Only unlocks after wireframes and design are confirmed. |
/CHECK |
Validate that all required files and locks exist. |
/SCAN |
Check accessibility, performance, and styling issues. |
/FIX |
Fix lint, type errors, or styling drift automatically. |
/DEPLOY |
Run CI, generate changelog, create release tag. |
/GUIDE |
See where you are in the pipeline and what to do next. |
Not sure which command to use? Run
/GUIDE— it always tells you the next step.
NEZAM enforces a strict order. You cannot run /DEVELOP until the design is locked. You cannot lock the design until the wireframes are done. This prevents the most common cause of AI-generated mess: building without a plan.
flowchart LR
A["🏁 /START\nOnboarding"] -->|PRD locked| B["📐 /PLAN\nPhases & Research"]
B -->|Specs ready| C["🎨 /DESIGN\nTokens & Profile"]
C -->|DESIGN.md locked| D["🔒 /WIREFRAME\nLayout Contract"]
D -->|wireframes_locked.json| E["⚙️ /DEVELOP\nGated Build"]
E -->|All gates pass| F["🚀 /DEPLOY\nRelease"]
style A fill:#E0F7FA,stroke:#0097A7,stroke-width:2px
style B fill:#F3E5F5,stroke:#7B1FA2,stroke-width:2px
style C fill:#FFF3E0,stroke:#F57F17,stroke-width:2px
style D fill:#E8F5E9,stroke:#388E3C,stroke-width:2px
style E fill:#FCE4EC,stroke:#C2185B,stroke-width:2px
style F fill:#F1F8E9,stroke:#689F38,stroke-width:2px
Each step creates a file. That file is the proof that the step is done. The next step checks for that file before it starts. No file → no access.
| Gate | File It Creates | What Gets Unlocked |
|---|---|---|
/START |
onboarding.yaml — prd_locked: true |
Planning |
/PLAN |
Phase specs in .nezam/core/plans/ |
Design |
/DESIGN |
DESIGN.md with tokens |
Wireframes |
/WIREFRAME |
wireframes_locked.json |
Development |
/CHECK |
Gate validation report | Deploy |
/DEPLOY |
Changelog + release tag | Production |
NEZAM keeps all your AI tools in sync automatically. You edit one folder — .cursor/ — and NEZAM mirrors it to Claude, Gemini, Windsurf, and others.
flowchart LR
SRC[".cursor/\nCanonical Source\n\nEdit here only"]
SRC -->|pnpm ai:sync| CL[".claude/\nClaude Code"]
SRC -->|auto-mirror| GEM[".gemini/\nGemini CLI"]
SRC -->|auto-mirror| WS[".windsurf/\nWindsurf"]
SRC -->|auto-mirror| CX["AGENTS.md\nCodex / OpenCode"]
SRC -->|auto-mirror| KIRO[".kiro/\nKiro"]
style SRC fill:#6366F1,color:#fff,stroke:#4F46E5,stroke-width:3px
style CL fill:#0EA5E9,color:#fff,stroke:#0284C7
style GEM fill:#8B5CF6,color:#fff,stroke:#7C3AED
style WS fill:#EC4899,color:#fff,stroke:#DB2777
style CX fill:#F59E0B,color:#fff,stroke:#D97706
style KIRO fill:#10B981,color:#fff,stroke:#059669
# Sync all mirrors from .cursor/
pnpm ai:sync
# Check for drift
pnpm ai:check
# Check sync status
pnpm ai:statusA pre-commit hook runs ai:sync automatically on every commit, so your tools never go out of sync.
The Design Hub is a local visual tool that runs in your browser. Use it to map your pages, pick colors, and set your layout — before writing any code.
pnpm design-hub
→ Open http://localhost:4000
Inside you will find:
- Sitemap Canvas — drag and drop your pages and routes
- Wireframe Builder — design your layouts block by block
- Theme Studio — pick colors, type scale, and spacing
- Export — click Export to lock everything into
wireframes_locked.json
Once you export, your wireframes are locked. The AI will follow your layout — not invent a new one.
To apply a ready-made design profile:
pnpm run design:apply -- nezam-v3Profiles are in .nezam/design-hub/design/<brand>/design.md.
NEZAM includes 150+ specialized agents. Each agent has one job. They hand off to each other through a controlled bus — no agent goes outside its scope.
A few examples:
| Agent | Job |
|---|---|
swarm-leader |
Orchestrates all other agents |
frontend-lead |
Owns all UI decisions |
design-lead |
Enforces the design contract |
qa-test-lead |
Writes and validates test coverage |
seo-specialist |
Handles SEO and AEO optimization |
security-auditor |
Runs security checks and CVE reports |
All agents are in .cursor/agents/. You can add, remove, or customize them.
NEZAM/
├── .cursor/ # ← Edit here. Source of truth.
│ ├── commands/ # Slash command definitions
│ ├── agents/ # 150+ specialized agents
│ ├── skills/ # Reusable skill modules
│ └── rules/ # Enforcement rules
│
├── .nezam/
│ ├── core/prd/ # Product requirements
│ ├── core/plans/ # Phase progress files
│ ├── core/gates/ # Hardlock configs & CI matrices
│ └── design-hub/ # Visual Design Hub (Next.js app)
│
├── docs/reports/ # Audit reports, security, performance
├── DESIGN.md # Active design contract (auto-generated)
├── wireframes_locked.json # Layout lock (auto-generated)
└── CLAUDE.md / AGENTS.md # AI mirrors (auto-generated — do not edit)
pnpm install fails with dependency errors
Force the public npm registry and try again:
pnpm config set registry https://registry.npmjs.org/
pnpm install/DEVELOP is blocked
This is expected. /DEVELOP only unlocks after you complete the wireframes. Run pnpm design-hub, build your layout, and click Export. This creates wireframes_locked.json and unlocks development.
Git commit is rejected by pre-commit hook
The Husky hook found sync drift. Fix it in one command:
pnpm ai:sync && git add -A && git commit -m "your message"Not sure what to do next
Type /GUIDE in your AI chat. It reads your project state and tells you exactly what step to take next.
| Document | What Is It |
|---|---|
| PRD Specification | Full product requirements |
| ONBOARDING.md | Step-by-step setup guide |
| DESIGN_TO_CODE.md | How design tokens become code |
| RUNBOOKS.md | Operations and incident runbooks |
| TROUBLESHOOTING.md | FAQ and common fixes |
| Commands Reference | All slash command definitions |
| Agents Library | All 150+ agent definitions |
git clone https://github.com/iDorgham/Nezam.git && cd Nezam && pnpm installThen open the folder in Cursor or Claude Code and type /START all.