We implement a service that tracks well-known Ethereum wallets, and sends real-time alerts of ETH and ERC20 token transfers via Telegram, using Python’s asyncio and websockets libraries. We subscribe to JSON-RPC events from a node over WebSocket connection, and parse transaction logs emitted by the events. We also give a concrete example of how a probabilistic data structure called bloom filter is used by the node to efficiently filter for relevant logs.

Since multiple state changes can be executed in one atomic transaction (where all or none of the changes are executed), one feature in blockchain contracts is the ability to conduct flash loans, where users can borrow unlimited amounts without collateral but must return funds within the same transaction. Predictably, flash loans have been applied towards arbitrage opportunities in automated market makers (AMMs, ie. protocols that facilitate trading through liquidity pools, instead of traditional order books). Below we demonstrate a minimal prototype of such technique.

My original post on setting up a local node using Geth no longer works after the Merge. Node operators now need to run both a consensus client in addition to the execution client (such as Geth). This post describes how I setup Geth and Prysm (one of the 5 consensus clients) using a Docker Compose script. After experimenting with installation via Ubuntu PPA, I found this method to be more elegant.

This is the second half of the Ethernaut series notes on how I solved each level using Foundry. For the first half, please see the previous post. The contracts deployed to solve the challenges, as well as test scripts, can be found on my Github.

Ethernaut is a Solidity game where each level needs to be “hacked” by finding some vulnerability in an Ethereum smart contract. I found this to be an engaging way to learn the basics behind Ethereum smart contracts. Currently, there are 27 levels of varying difficulty. Below are brief notes on how I completed each level. While there are already many solution write-ups out there, most use Hardhat while this writeup uses Foundry (see previous blog) to test and interact with the game.