Wanted to share something I've been tinkering on for a bit.
The Problem:
Crypto is notorious for its hacks and uncanny ability to put user money at risk. There's a million companies focusing on securing the smart contract piece of the puzzle through audits (CodeArena, Sherlock, etc) and on-chain monitoring (Hypernative), but a glaring ignorance to frontend attacks.
This is a bit concerning given that we've seen millions of dollars go up in smoke through frontend attacks on BadgerDAO (Cloudflare worker), Curve (DNS), Sushi Miso (supply chain), and others.
The problem is historically there have been no good frameworks for E2E testing where those that did exist were either extremely convoluted to use, extremely flaky, or both. There have also been no systems for monitoring these kinds of E2E tests like there are in Checkly, LambdaTest, etc for traditional apps.
The Solution:
I built a E2E testing framework and a system for monitoring these E2E tests against production sites to alleviate this.
The framework - GuardianTest - is open source, has minimal dependencies, has no brittleness around wallet interactions, can perform actions against network forks (Ethereum, Arbitrum, Optimism, Polygon), can easily mock token balances, can validate the contract target of an app interaction, can make assertions about on-chain state, and has all the validation and action capabilities of Playwright
Link to the code: https://github.com/GuardianUI/GuardianTest/tree/v1.0.3
Link to NPM: https://www.npmjs.com/package/@guardianui/test?activeTab=rea...
The monitoring - Once these tests are written they can be submitted to be monitored as frequently as every 5 min, or as infrequently as once a day, will provide alerts are sent if something goes wrong, and alerts are categorized by their significance so there should be no false positives for critical issues
General docs: https://docs.guardianui.com/overview/readme