Aegis-Net is a decentralized AI Smart Contract Auditor designed for the Bittensor ecosystem. It crowdsources intelligent threat detection models using the compute power of decentralized LLM Miners (su
🌟 Overview Aegis-Net is a decentralized security network built on the Bittensor ecosystem. It crowdsources intelligent threat detection models using the compute power of decentralized LLM miners to inspect, simulate, and expose vulnerabilities in deployed Smart Contracts.
It aims to replace the slow, expensive, and centralized auditing processes currently prevailing in the Web3 space by creating an incentivized network of AI models competing to discover vulnerabilities with near real-time efficiency.
🚨 The Problem Web3 lost over $3.8 Billion to smart contract exploits last year alone. Traditional centralized auditing is fundamentally broken: it's too slow (2-6 weeks turnaround), too expensive ($50k-$500k per audit), and acts merely as a static snapshot rather than continuous protection. Protocols are forced to choose between shipping fast (and vulnerable) or staying safe (and losing market share).
🛡️ The Aegis-Net Solution (How It Works)
AI Miner Swarm: Hundreds of competing AI miners (specialized LLMs like DeepSeek-Audit-7B and Llama-Solidity-13B) race to detect zero-day vulnerabilities in seconds.
Validator Consensus: Findings are strictly verified via automated exploit simulation in forked environments (Anvil/Hardhat), ensuring zero false positives.
$TAO Incentives: Miners and Validators are rewarded purely based on accuracy, speed, and consensus—creating a self-improving security ecosystem.
🚀 What We Built for This Hackathon (Live Prototype) We didn't just write a proposal; we built a high-fidelity working prototype:
Live AI Smart Contract Scanner: A production-ready DApp featuring a Dual-AI Engine (Groq LPU LLaMA-3 70B & DeepSeek V3).
Instant Auditing: Users can paste any Solidity code and receive a brutal, instant vulnerability report (catching Reentrancy, Flash Loans, Logic Flaws) in under 8 seconds.
Interactive Dashboards: Featuring 'The Intercept' (Client Portal for bounties) and 'The Hive' (Subnet Explorer for miner leaderboards).
🔗 Official Links
Live DApp Demo: https://aegis-net-pi.vercel.app/
GitHub (Full Technical Proposal): https://github.com/panzauto46-bot/AEGIS-NET
Designed the Subnet Architecture (Miner/Validator/PoA), built the UI/UX Dashboards, and successfully integrated a working Live AI Scanner using Groq LLaMA-3 and DeepSeek API to demonstrate Miner intelligence.
Bootstrapped (Self-funded). > Currently seeking seed funding or ecosystem grants to accelerate Phase 2 (Core Engine Development), scale our AI Miner operations, and deploy our Validator infrastructure on the Bittensor Testnet.