RugCheck AI

RugCheck AI is an AI-powered on-chain security analyst built on OKX.AI that helps users evaluate the safety of crypto tokens before they buy. Users simply paste a token contract address, and RugCheck instantly analyzes trusted on-chain security data from OKX Onchain OS to produce a clear risk verdict, a 0–100 risk score, liquidity metrics, holder distribution, market information, and security signals such as honeypots, mint authority, freeze authority, and other potential red flags.

Unlike traditional token scanners that overwhelm users with technical jargon, RugCheck combines a deterministic analysis engine with grounded AI explanations. The security engine determines the verdict using verified on-chain data, while the AI explains every finding in plain English without inventing facts, changing the verdict, or providing financial advice. Users can explore the report from the perspective of a beginner, trader, or developer, making complex blockchain security accessible to everyone.

RugCheck AI also includes an interactive AI analyst that answers follow-up questions about a specific report, helping users understand why a token received its score, what each risk means, and which issues deserve the most attention. Every scan can be shared as a professionally designed result card or posted directly to X, making security insights easy to share with the broader crypto community.

Built as an OKX.AI Agent Service Provider (ASP) with MCP support, RugCheck AI is not just a web application. It is a reusable AI service that can be invoked by both users and other AI agents, enabling secure, explainable, and accessible token risk analysis across the OKX.AI ecosystem.

• Ideated and designed RugCheck AI, an AI-powered on-chain token safety analyst.

• Built a complete web application with a modern, responsive interface.

• Integrated OKX Onchain OS to retrieve trusted on-chain security data.

• Developed a deterministic risk engine that generates a risk score, verdict, and security findings.

• Added AI-powered explanations that translate technical security signals into plain English without hallucinating or changing the underlying verdict.

• Implemented multiple explanation modes tailored for beginners, traders, and developers.

• Added an interactive AI chat to answer follow-up questions about each security report.

• Implemented shareable result cards and one-click posting to X for community sharing.

• Built and deployed the project as an OKX.AI MCP-compatible ASP.

• Successfully registered the ASP on XLayer and submitted it for OKX.AI marketplace listing.

• Created a complete demo video and documentation for the final submission.

This highlights both the technical work and the fact that you successfully built and deployed an actual OKX.AI agent during the hackathon.

NA