Cognitive Framework
Rug.ai has built its own proprietary Cognitive Framework with the aim of building a complete tool capable of detecting and quantifying token vulnerabilities. Our AI framework aims to resolve both deterministic (hard-rugs) & non-deterministic (soft-rugs) vulnerabilities. Our cognitive framework renders on our interface in two main ways: Firstly by encoding results in a score, and secondly by rendering the rug.ai AI Contract Report.
Our Scoring system, described in our scoring section, is calculated by our Cognitive Framework and takes into account scores across 4 different verticals: Supply, Transferability, Liquidity & Audit. The scoring system is designed to quantify the risk of engaging with specific token contracts, in an intuitive numerical system interpretable by anyone and everyone.
The second way in which our Cognitive Framework is used in our product is through the AI Contract Report. Mentioned earlier, the AI Contract Report makes up a part of the scoring system for tokens, but it's also a standalone product. Our AI Contract Report parses through token source code as soon as the source code is available on-chain, and utilising our advanced AI models outputs features of the token contract ordered by severity. It then prescribes a score to the token determining the safety of the token based on flagged elements:
The purpose of our AI Contract Report is to give an insight into the functionality of the token, what's possible and what's not. In many cases, some of these flagged functions may be harmless on the surface, but coupled with human context could give invaluable insight into the token mechanics. Outside of nuanced cases, the contract report can identify harmful functions, where they exist and explain in understandable language why this particular function is harmful. This can outright tell users whether tokens are a honeypot, or have dangerous qualities which make certain tokens worth avoiding. In the following pages, we explain how this Cognitive Framework works, what it's comprised of, and the advantages of the approach we've taken.
Last updated
