CK
GitHubnpmv0.13.0
Context Engineering + Agentic AI Pipelines

Ground your AI agents
in your conventions

Scaffold structured standards so every AI tool reads your project conventions automatically. Then run a grounded multi-role pipeline — PO to Architect to Dev to Tester to Reviewer — without agents hallucinating your patterns.

Get StartedRead DocsGitHub
$npm install -g @nolrm/contextkit

Key Features

Context-aware AI development that works seamlessly with your existing tools and engineering patterns.

Context Engineering
Provide structured markdown context to AI assistants, preventing hallucinations and ensuring code matches your exact patterns and architecture.
Project-Specific Standards
Auto-detect your tech stack and customize standards. Glossary, code style, testing patterns—all tailored specifically to your project's needs.
Multi-Platform Support
Works with Cursor (auto-loads), VS Code (Copilot), Aider, Continue.dev, UX Pilot CLI, Gemini, and more. One context source for all tools.
Squad Workflow
Multi-agent pipeline that simulates a full team: Product Owner → Architect → Dev → Test → Review. Each role reads and writes to a shared handoff file.
MD-First Development
Write a spec before writing code. Every component gets a colocated markdown file describing its purpose, props, and logic — giving AI full context before it touches a line.
Slash Commands
17 reusable AI workflows installed across all platforms: /analyze, /fix, /refactor, /spec, /squad, and more. One command set, every tool.

Structured Context Files

ContextKit generates a consistent folder structure for your project documentation. This allows LLMs to understand your project's architecture instantly.

  1. 1
    Initialize your projectRun npx contextkit init to create the structure.
  2. 2
    Define standardsEdit the generated markdown files to match your team's coding style.
  3. 3
    Generate better codeAI assistants read these files to produce code that actually compiles and fits in.
.context/architecture/tech-stack.md

# Tech Stack & Architecture

## Frontend Framework

- **Next.js 16** (App Router)

- **TypeScript** (Strict mode)

- **Tailwind CSS** for styling

## State Management

- **Zustand** for global client state

- **React Query** for server state

## Database

- **Supabase** (PostgreSQL)

- **Prisma** ORM

## AI Patterns

- Use Vercel AI SDK for streaming

- Prefer Server Actions over API routes

Trusted by developers building with

React
Vue
Node.js
Python
Rust