Set up
Skills are installed by placing their folder where your agent loads skills.
The download comes from github.com/iDorgham/Nezam itself, not from our server, so what you get is exactly what its author published.
Install instructions come from the README
README
synced Oct 27, 2026NEZAM
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
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.
✅ What NEZAM Does
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 |
🚀 Quick Start
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 all
That 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.
💬 Slash Commands
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.
🔄 How It Works
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 Reference
| 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 |
🔄 AI Sync
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:status
A pre-commit hook runs ai:sync automatically on every commit, so your tools never go out of sync.
🎨 Design Hub
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-v3
Profiles are in .nezam/design-hub/design/<brand>/design.md.
🤖 AI Agents
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.
📁 Repository Structure
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)
🛠️ Common Issues
Force the public npm registry and try again:
pnpm config set registry https://registry.npmjs.org/
pnpm install
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.
The Husky hook found sync drift. Fix it in one command:
pnpm ai:sync && git add -A && git commit -m "your message"
Type /GUIDE in your AI chat. It reads your project state and tells you exactly what step to take next.
📚 Documentation
| 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 |
Ready to start?
git clone https://github.com/iDorgham/Nezam.git && cd Nezam && pnpm install
Then open the folder in Cursor or Claude Code and type /START all.