/honeypot-demo

An experiment to bait a generalized frontrunning MEV bot

Primary LanguageJupyter Notebook

Honeypot 🍯

A simple experiment that:

  • Creates a honeypot contract
  • Baits a generalized fronturnning MEV bot with a unique transaction
  • Analyze bot behaviour using a black box approach

Final project for ChainShort bootcamp Oct 2021 cohort.

Presentation Deck

The project presentation deck is in presentation directory. It gives an overview about the project.

Experiment addresses and txs

Honeypot contract address: 0x1e232d5871979eaa715de2c38381574a9c886bad

Bot contract: 0x31B7e144b2CF261A015004BEE9c84a98263E2F66

Bot operator: 0x0a04e8b4d2014cd2d07a9eaf946945bed1262a99

Failed tx 1 (block 13710082, index 22): 0xcc1172506d5b5fa09cbf66d2296deb24958181f186817eb29cbe8385fd55ed51

Frontrun tx 1 (block 13710082, index 0): 0x18ec2c2e5720c6d332a0f308f8803e834e06c78dcebdc255178891ead56c6d73

Failed tx 2 (block 13710542, index 80): 0xfce9b77a8c7b8544cb699ce646558dc506e030aaba1533c917d7841bcc3f206a

Frontrun tx 2 (block 13710542, index 0): 0x8cda6e76f9a19ce69967d9f74d52402afbafba6ca3469248fe5c9937ef065d47

Running contract tests

The contract tests are written in Solidity. To run them:

  1. Install dapptools on your machine
  2. Navigate to the project root directory in terminal, then dapp install ds-test
  3. Rename .dapprc.template to .dapprc and add your Ethereum RPC endpoint
  4. Use dapp test to run the tests.

PnL dataset

To create or update the PnL dataset:

  1. Make sure you have Python 3 and the relevant modules installed on your machine
  2. Rename config.template.py to config.py and add your Etherscan API key and Alchemy RPC endpoint
  3. Run python analysis/create_pnl_datasets.py in your terminal

Analysis

You can view the analysis files on GitHub. If you want to edit and run them, you need to run Jupyter Notebook server with Anaconda or something similar.

Known limitations

These limitaitons are known by the time of the final presentation:

  • Unoptimized performance and too many JSON-RPC calls in when fetching data
  • PnL computation is based on heuristic, not EVM state changes
  • Outlier detection is based on manual sample check
  • A few hardcoded simplifications like constant token prices
  • No test for pnl.py and calldata.py