← Back to home

VeriBot

Using zkML to prove the absence of vulnerabilities in (closed-) source code.

Screenshots

VeriBot screenshot 1
VeriBot screenshot 2
VeriBot screenshot 3
VeriBot screenshot 4
VeriBot screenshot 5
VeriBot screenshot 6

Problem Statement

Imagine you've poured countless hours into crafting the perfect smart contract, only to face the daunting task of proving its safety without revealing your confidential source code.Imagine you've poured countless hours into crafting the perfect smart contract, only to face the daunting task of proving its safety without revealing your confidential source code.Enter VeriBot! Developers can now confidently showcase the integrity of their contracts without compromising their code's confidentiality. Through Zero-Knowledge Proofs and Machine Learning, VeriBot empowers you to demonstrate your contract's reliability while keeping your code private. Say goodbye to sleepless nights worrying about rug pulls or vulnerabilities – VeriBot has got your back, ensuring your creations are as trustworthy as they are confidential.

Solution

Circuit : Utilizes Zero-Knowledge Proofs in Noir Language to validate machine learning inferences on smart contract bytecode, ensuring privacy and integrity without revealing the bytecode.Oracle : A Rust-built intermediary that securely connects off-chain machine learning predictions with on-chain smart contract decisions, enhancing contracts without exposing underlying data or models.Machine Learning: Analyzes smart contract bytecode to infer properties or vulnerabilities, acting as a privacy-preserving tool that abstracts complex contract logic for secure validation.

Hackathon

Circuit Breaker

2024

Prizes

  • 🏆

    Best use of Sindri in a ML x Web3 context1st place

    Sindri

  • 🏆

    🏆 Circuit Breaker Finalist

    ETHGlobal

Contributors