Skip to content

Phani3108/Cortex

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🧠 Cortex

One config. Every AI coding tool. Zero drift.

🌐 cortex1.vercel.app · 📦 GitHub

Cortex is a universal context engine that compiles a single .cortex/ source into native config files for 9 AI coding tools — so every tool gets the same rules, skills, and style, always in sync.


⚡ The Problem

You use Claude Code, Cursor, Copilot, Gemini, and others — but each has its own config format. You end up:

  • ✍️ Writing the same rules in 4 different files
  • 🔀 Configs drifting out of sync across tools
  • 🧑‍🤝‍🧑 Team members getting inconsistent AI behavior
  • 🕐 Spending 15+ min per tool on manual setup

Cortex fixes this in one command.


🚀 Features

🔧 Engineering

  • 🔁 One Source → 9 Outputs — Write rules once in .cortex/, compile to Claude, Cursor, Copilot, Windsurf, Gemini, Codex, Kiro, Antigravity, and OpenAI
  • 🧬 Model-Aware Formatting — Automatically adapts output format per model family (XML tags for Claude, numbered lists for Copilot, minimal scaffolding for reasoning models)
  • 📡 Signal Detection — Scans git diffs, linter output, and AI chat logs to detect patterns in your workflow
  • 🔄 Learning Loopcortex learn captures signals → evolves rules → recompile propagates to all tools
  • 🪝 Git Hooks — Auto-learn on commit, auto-recompile when .cortex/ changes
  • 📦 Import Existing Config — Already have a CLAUDE.md or .cursorrules? Import them into .cortex/ in one command
  • 👀 Watch Mode — Auto-recompile and auto-learn as you edit, in real time

📊 Product

  • 🤖 Guided Assistantcortex assist walks you through setup with questions, not docs
  • 📈 Impact Metrics — Tracks time saved, consistency score, and learning velocity across your project
  • 🧠 Session Memory — Remembers your goals, decisions, and progress between runs
  • 💰 Token Cost Analysis — See exactly how much context you're sending to each provider
  • 📤 Diff & Export — See what changed since last compile; export your entire context for sharing

🏢 GTM / Team Value

  • ⏱️ Setup in 60 secondscortex init + cortex compile and you're done
  • 👥 Team Consistency — Commit .cortex/ to git; every teammate runs cortex compile and gets identical AI behavior
  • 🔌 9 Providers, 1 Workflow — Switch tools freely without rewriting config
  • 📚 Shareable Skills — Package reusable expertise (TDD, security audits, debugging) as portable skill files
  • 🔄 Sync from Upstream — Pull community rules and skills from remote registries

🏛️ Stack Intelligence (Interactive — Live Site)

Three interactive modes to help you pick the right tech stack:

  • Pick My Stack — Browse 40+ technologies across Frontend, Backend, Database, Deployment, Tooling, and Architecture Patterns. Select what you use and see a scored build matrix of what you can build, with stack tips and a ready-to-use Cortex config.

  • Guide Me — A 4-step wizard that asks what you're building, what matters most, your team size, and your budget — then recommends a proven stack combination with real-world examples.

  • Architect My Project — A deep 9-step architecture advisor powered by real startup intelligence:

    • 9 adaptive questions: project type (with "similar to" real products), core user workflow, scale, team size & timeline, data requirements (10 options), authentication needs, infrastructure, team skills, and compliance
    • 33 live insight rules that fire in real-time as you answer — covering N+1 query traps, connection pooling pitfalls, multi-tenant data isolation, compliance constraints, and more
    • Output includes: Core Stack with per-tech reasoning (why each technology, what real companies use it), Architecture Patterns with implementation detail, Startup Traps (real post-mortem failure patterns specific to your choices), Compliance & Security warnings, Month-by-month Scaling Roadmap, and Alternative Approaches with "best when" guidance

🛠️ Installation

# Clone the repo
git clone https://github.com/Phani3108/Cortex.git
cd Cortex

# Install globally (makes `cortex` available everywhere)
npm install -g .

# Or run directly without installing
node bin/cortex.js

🏁 Quick Start

# 1. Initialize Cortex in your project
cortex init

# 2. Enable the tools you use (edit .cortex/config.yaml)
#    providers: claude, cursor, copilot, gemini, windsurf, codex, kiro, antigravity, openai

# 3. Compile — generates native config for every enabled tool
cortex compile

# 4. Done. Your AI tools now share the same intelligence.

📋 Commands

Command What it does
cortex init Initialize .cortex/ in your project
cortex compile Generate provider-specific config files
cortex learn Capture signals and evolve your rules
cortex watch Auto-recompile and auto-learn as you work
cortex assist Guided conversational setup
cortex import Import existing CLAUDE.md, .cursorrules, etc.
cortex diff See what changed since last compile
cortex cost Token cost analysis across providers
cortex hooks install Install git hooks for auto-learning
cortex add skill <name> Add a skill template (tdd, security-audit, debugging)
cortex sync Pull rules/skills from remote sources
cortex status Show current configuration
cortex export Export context for sharing or backup
cortex profile Manage your personal AI style globally

🎯 Supported Providers

Provider Output Files
🟠 Claude Code CLAUDE.md + .claude/commands/*.md
🟣 Cursor .cursor/rules/project.mdc
🔵 GitHub Copilot .github/copilot-instructions.md
🟢 Windsurf .windsurf/rules/project.md
🔴 Gemini CLI GEMINI.md + .gemini/style-guide.md
OpenAI Codex codex.md
🟡 Amazon Kiro .kiro/rules/*.md
🔵 Antigravity .agent/skills/*.md
OpenAI ChatGPT chatgpt-instructions.md

📁 Project Structure

.cortex/                  ← Your source of truth (commit this)
  config.yaml             ← Providers, language, preferences
  rules/                  ← Project rules (style, architecture, testing)
  skills/                 ← Reusable skills (TDD, security, debugging)

bin/cortex.js             ← CLI entry point
src/
  commands/               ← CLI command handlers
  core/                   ← Compiler, signals, metrics, assistant
  providers/              ← Provider-specific output generators
  utils/                  ← File system, logging, YAML helpers
templates/                ← Built-in rule & skill templates

🧩 How It Works

 .cortex/rules/    ─┐
 .cortex/skills/    ├──▶  cortex compile  ──▶  CLAUDE.md
 .cortex/config.yaml┘         │               .cursorrules
                               │               copilot-instructions.md
                               │               GEMINI.md
                               │               ... (9 formats)
                               ▼
                        Model-aware formatting
                        (XML, numbered lists,
                         minimal, etc.)

📄 License

MIT License — Copyright (c) 2026 Phani Marupaka


🌐 Live: cortex1.vercel.app · 📦 Source: github.com/Phani3108/Cortex

Created & Developed by Phani Marupaka. All rights reserved under applicable copyright law.

Any fork, derivative work, or redistribution must visibly credit the original author and include a link to linkedin.com/in/phani-marupaka.

See LICENSE for full terms.


Built for engineers who use more than one AI tool and want them all to be equally smart.

About

One config. Every AI coding tool. Zero drift. Cortex is a universal context engine that compiles a single .cortex/ source into native config files for 9 AI coding tools — so every tool gets the same rules, skills, and style, always in sync.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors