Package Exports

@cortex-local/cli
@cortex-local/cli/dist/cli.js

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (@cortex-local/cli) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

⚡ CORTEX

Sovereign Intelligence Infrastructure.
Private local AI for sensitive codebases.

🚀 Recent Updates (v4.0.1)

Workspace Cleanup: Removed legacy research documents, scratch files, and debug artifacts to minimize footprint.
Windows Compatibility: Fixed EPERM initialization bug in SharedContext.ts when running the CLI from a Windows drive root (C:\).

📖 What is CORTEX?

CORTEX is a local-first AI development environment that combines a multi-agent terminal interface with a neuro-symbolic cognition engine (BIGROCK ASI) in a single process.

It is designed for teams that need AI-assisted development without sending proprietary code to third-party cloud services — defense contractors, healthcare organizations, financial firms, and government agencies.

Core Capabilities

100% Local Execution: All cognition runs on your hardware. No HTTP calls to external AI services for STEM computation.
Deterministic STEM Proofs: Symbolic math, physics simulation, and formal logic — with full proof traces and 100% confidence.
Local LLM Absorption: Drop any GGUF model into the models/ directory. CORTEX probes it, profiles its domain strengths, and routes queries to the best model.
Multi-Agent Development: 7 specialized agents (Developer, Explorer, Planner, Quality, DevOps, Browser, Network) with automatic task routing.
Cloud Fallback: When local models aren't available, CORTEX routes to cloud LLMs (OpenRouter/OpenAI) with automatic retry and model cascade.

🧠 The Cognition Core

CORTEX's local intelligence is powered by BIGROCK ASI — a neuro-symbolic engine that runs in-process with zero network latency.

Engines

Engine	What It Does	Output
Symbolic Math Engine	Algebraic evaluation, differentiation, solving — with step-by-step proof traces	Deterministic
Physics Engine	First-principles simulation across 6 domains (Newtonian, Relativistic, Quantum, Thermodynamics, Electromagnetism, Orbital)	Deterministic
Formal Logic Prover	Exhaustive truth-table evaluation of propositional logic expressions	Deterministic
ODE Neural Network	Liquid Time-Constant network with RK4 integration and Hebbian learning — adapts during inference	Adaptive
LLM Devourer	Scans local GGUF models, probes domain performance, profiles and ranks them	Adaptive
Mythos Threat Engine	12-domain cybersecurity threat ontology — maps architecture descriptions to known TTPs	Heuristic
Episodic Memory	TF-IDF + dense vector semantic search over conversation and thought history	N/A
Self-Reflection Engine	7-check metacognitive auditor that adjusts confidence of outputs	N/A
Autonomous Reasoning Loop	Background cycles: system health patrol, self-directed curiosity queries, memory consolidation	N/A

Pipeline

Input → Security Scan → DAG Decomposition → Hypervisor Sandbox
  → [Symbolic | Physics | Logic | Neural | Kernel | LLM Devourer]
  → Self-Reflection → Episodic Memory → Output

🤖 The Agent Swarm

Agent	Responsibility
ExploreAgent	Filesystem reading and codebase analysis
PlanAgent	Task decomposition and planning
DeveloperAgent	Code authoring, editing, and refactoring
QualityAgent	Testing, linting, and verification
DevOpsAgent	Infrastructure and dependency management
BrowserAgent	Headless web navigation and data extraction
NetworkAgent	API testing and network diagnostics

Routing Logic

STEM queries (math, physics, logic) → Local CognitionCore (deterministic, zero latency)
System queries (CPU, memory, processes) → Kernel Interface (direct OS telemetry)
Code/language tasks → Cloud LLM with automatic 12-model fallback cascade
Security analysis → Mythos Engine + CyberSecurityKing (10-engine scan)

🏗️ Architecture

┌─── CORTEX v4.0 ─── Single Unified Process ────────────────────────────┐
│                                                                        │
│  ┌─ CortexKernel (Singleton) ───────────────────────────────────┐     │
│  │  CognitionCore │ ARL │ Mythos │ Absorber │ Hypervisor        │     │
│  │  BeastMode     │ EpisodicMemory │ SymbolicEngine │ Physics   │     │
│  └──────────────────────────────────────────────────────────────┘     │
│        ↕ direct function call (zero-latency)                          │
│  ┌─ AgentManager ──────────────────────────────────────────────┐     │
│  │  Smart Routing → Local CognitionCore OR Cloud LLM            │     │
│  │  7 Sub-agents │ Plugin System │ SharedContext                 │     │
│  └──────────────────────────────────────────────────────────────┘     │
│        ↕                                                               │
│  ┌─ React Ink TTY ─────────────────────────────────────────────┐     │
│  │  Chat UI │ /beast │ /dashboard Live Telemetry                │     │
│  └──────────────────────────────────────────────────────────────┘     │
│                                                                        │
└────────────────────────────────────────────────────────────────────────┘

🛠️ Installation & Setup

⚡ One-Line Install

Windows (PowerShell):

irm https://raw.githubusercontent.com/vimalspaceton618-afk/CORTEX/main/install.ps1 | iex

macOS / Linux:

curl -fsSL https://raw.githubusercontent.com/vimalspaceton618-afk/CORTEX/main/install.sh | bash

npm (if Node.js is already installed):

npm install -g cortex && cortex --setup

Manual Install

git clone https://github.com/vimalspaceton618-afk/CORTEX.git
cd CORTEX
npm install
cd BIGROCK_ASI && npm install && cd ..
npm run boot

5. Install Globally (Optional)

npm install -g .

💻 Usage

Terminal Commands

Command	Action
`/help`	Display all commands and agents
`/beast`	Toggle BEASTMODE (cranks all subsystem configs to max)
`/dashboard`	Toggle live real-time system monitoring
`/status`	Full system status report
`/think <query>`	Force route through local CognitionCore
`/arl`	Show Autonomous Reasoning Loop metrics
`/arl start`	Start ARL background cycles
`/arl stop`	Stop ARL background cycles
`/brain eat`	Scan and absorb local GGUF models
`/brain`	Show active brain info and model status
`/brain iq`	Show Collective Intelligence Quotient breakdown
`/brain wisdom`	Show domain coverage heat map
`/mythos <arch>`	Run Mythos threat analysis on an architecture description
`/cyberscan`	Full spectrum 10-engine security scan
`/cyberheal`	Generate autonomous remediation scripts
`/demo`	Run the automated showcase sequence
`/health`	Show runtime readiness checks
`/setup`	Brain configuration wizard
`/exit`	Quit

Examples

cortex> escape velocity from Earth
→ [Local CognitionCore] 11186.0 m/s (100% Deterministic Proof)

cortex> /think p AND q IMPLIES p
→ [Formal Logic Prover] TAUTOLOGY (always true) — 100% Confidence

cortex> /mythos A Node.js API connected to S3 and Stripe
→ ⚠️ 8 Myth Gaps detected (Cloud Security, Web & API, DevSecOps...)

cortex> Refactor the authentication module to use JWT
→ [AgentManager → DeveloperAgent → QualityAgent auto-verification]

🛡️ Security

Input Sanitization: SecurityProcessor scans all inputs before processing (injection, prompt attack, path traversal detection)
Integrity Chain: Every thought is cryptographically hashed (AES-256-GCM) into a sequential chain — tampering is detectable
Process Hypervisor: All engine executions are sandboxed with capability-based access tokens
Mythos Threat Ontology: 12-domain threat taxonomy mapping architecture descriptions to known TTPs
Air-Gap Compatible: The CognitionCore, physics engine, math engine, and logic prover require zero network access

📄 License

© 2026 SpaceTon.
This software is proprietary. Unauthorized copying, distribution, or modification is prohibited.