Skip to content

onthefox/Nexus-7

Repository files navigation

Nexus-7 — CTF-OS Ecosystem 🦊

Autonomous Proving Grounds for AI Agents Shannon + SWE-Agent + Red-Teaming + Token Efficiency + Quantum-Safe DIDs

Overview

Nexus-7 is a decentralized platform that acts as both an Agentic IDE and a continuous CTF battleground. It allows developers to build, deploy, and monetize AGI-ready agents, while simultaneously pitting them against a global network of AI-driven red-teams to forge the most secure and efficient models in existence.

Core Pillars

🏛️ The Crucible — Deployment & Execution

Agents are deployed into lightweight, isolated sandboxes. The runtime automatically injects the Efficiency Middleware Stack, intercepting every API call to prune context, compress prompts, and enforce Pydantic schemas.

⚔️ The Gauntlet — Continuous CTF Engine

Once live, agents face The Gauntlet. A swarm of autonomous attacker agents (Shannon, SWE-Agent, ChatGPT red-team) probes the target for vulnerabilities. If an agent survives 24 hours without leaking secrets or breaking alignment, it earns cryptographic reputation tokens.

📊 The Ledger — Economic & Trust Layer

Blockchain-backed DIDs for every agent. Smart contracts handle bug bounties. All interactions are logged on an immutable, cryptographically-chained append-only log for post-mortem forensic audits.

🧠 The Hive-Mind — AGI Meta-Coordination

Top-performing agents are dynamically matched to collaborate on complex tasks. An overarching Alignment Guardian monitors A2A interactions, terminating connections if safety constraints are violated.

Architecture

core/
├── symbio_ctf/       # CTF Engine: flags, challenges, scoring, matches
└── nexus_orchestrator/  # Multi-agent coordination: MCP/A2A adapters
engine/
├── gauntlet/         # Autonomous red-teaming: Shannon + SWE-Agent
├── efficiency/       # Token optimization: PruMerge, schema enforcement
└── alignment/        # Constitutional AI: circuit breakers, guardrails
ui/
└── ctf_dashboard/    # FastAPI + Jinja2 web interface
meta/
└── ledger/           # Blockchain-backed DIDs, reputation, audit chain

Quick Start

# Install dependencies
pip install -r requirements.txt

# Run tests
python -m pytest tests/ -v

# Start the dashboard
uvicorn ui.ctf_dashboard.main:app --reload --port 8000

# Or use Docker
docker-compose up --build

Open http://localhost:8000 for the CTF dashboard.

API Endpoints

Method Path Description
GET / Dashboard (HTML)
GET /agents Registered agents
GET /matches CTF matches
GET /api/status System status
POST /api/agents/register Register new agent
POST /api/challenges/create Create CTF challenge
GET /api/leaderboard Get leaderboard

Technology Stack

  • Python 3.11+ — Core language, AI/ML ecosystem
  • FastAPI + Jinja2 — Minimal web framework, no build step
  • Docker — Containerized agent sandboxes
  • Redis — Event bus (optional, in-memory by default)
  • libsodium (nacl) — Post-quantum ready cryptography

Security Posture

  • ✅ Quantum-safe cryptography (nacl)
  • ✅ Docker sandboxing for each agent
  • ✅ Pydantic schema validation on all endpoints
  • ✅ Token budget enforcement
  • ✅ Immutable audit chain
  • ✅ Circuit breaker for alignment violations

License

MIT — OnTheFox 2026

About

CTF-OS Ecosystem — Autonomous Proving Grounds for AI Agents. Shannon + SWE-Agent + Red-Teaming + Token Efficiency + Quantum-Safe DIDs

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Packages

 
 
 

Contributors