Problem

The increasing cost of audits and the rise in the sophistication of hacks and exploits present a growing challenge in the blockchain space. Current testing solutions, such as fuzz testing, are helpful but inadequate to address the problem comprehensively. There is a market gap between affordable peer reviews and expensive professional audits, leading to smart contracts being deployed without sufficient testing, leaving them vulnerable to attack.

Solution

We propose an innovative platform leveraging recent advancements in LLM agents to automatically analyze smart contracts, identify security weaknesses, and suggest code improvements. By aggregating a comprehensive vulnerability knowledge base and equipping AI agents with the necessary tools, we can offer accessible and effective security services at as low a cost as possible.

Our platform, powered by Agent Smth's LLM agent infrastructure, aims to provide rigorous and secure smart contract development solutions, making web3 safer for all users.

Project Proposal

Our project plan focuses on delivering a user-friendly platform that streamlines smart contract security reviews and gas optimizations, ultimately promoting enhanced security and token engineering within the cryptocurrency ecosystem. The project comprises three main phases:

Phase 1: Building the security smart contract agents

We will develop AI agents to identify security vulnerabilities and suggest code amendments, utilizing Smth AI's LLM agent infrastructure. The agents will be accessible through a dedicated user interface on the antifragile.build website, where users can upload their smart contracts for analysis.

Deliverables

• Hosted security vulnerability agents on Smth AI platform
• User interface for smart contract analysis on antifragile.build

Phase 2: Add gas optimization agents

We will enhance the platform's capabilities by incorporating AI agents that identify gas optimization opportunities and suggest relevant code modifications, along with any required tools.

Deliverables

• Gas optimization agents and tools added to the platform

Phase 3: VitalikGPT

To further support our users, we will create educational materials and integrate chatbot functionality to help community members better comprehend smart contracts and the impact of token design choices on projects.

Deliverables

• Educational content covering smart contract security, gas optimization, and token engineering best practices
• Chatbot integration for real-time assistance and walk-throughs of smart contracts for investors and developers

Conclusion

AntiFragile.Build provides an accessible, cost-effective, and effective solution to address the growing demand for smart contract security and gas optimization in the web3 ecosystem. By harnessing the power of LLM agents, our platform aims to mitigate security risks and elevate the overall quality of smart contract development.

How would we spend the funds?

Based on the proposed project plan, the grant funds will be spent on the following activities:

1. Development of security vulnerability and gas optimization agents
2. Design and implementation of an intuitive user interface
3. LLM and any external API costs during development as well as an amount set aside to cover usage costs by public users
4. Initial marketing and community outreach initiatives

Team

Dec Kickham - A senior strategic designer from BCG Digital Ventures, Dec has helped commercialise AI in industry from computer vision startups through to DeepMind. Dec has an expertise in translating customer needs into crisp UX and has a Masters in Service Design from Royal College of Art.

Simon Dunkelman - Simon is a full-stack builder in web3 who’s worked on several projects to improve the safety and awareness of crypto investors. At his previous startup Health Potion he built a decentralised insurance product that protected investors against phishing scams and successfully insured $43K of assets for the MVP. He has a Masters in Risk Management & Financial Engineering from Imperial College London.