Skip to main content

Provide a visualization tool for sandwich attacks

Goal​

Build a simple, trustworthy, and open visualization tool that lets anyone:

  • Inspect total and per-attacker metrics (revenues, profits, number of attacks).
  • Inspect total harm to victims.
  • Search and filter attacks by victim or attacker address.
  • Sort results by timestamp, revenue, profit, or harm.
  • See interactive charts of attacker revenue/profit vs victim harm.
  • Access public API docs and call the API (historic window: from 2020).
  • Source: swap & transaction logs from Ethereum via GCP BigQuery (only Uniswap V2, Sushiswap V2, PancakeSwap V2 are used for detection).

sandwichscan app
https://app.sandwichscan.baltoon.jp

sandwichscan api docs
https://api.sandwichscan.baltoon.jp/app/v1/scalar

docs repository https://github.com/PeterTakahashi/sandwich-scan-docs

fastapi backend api repository https://github.com/PeterTakahashi/sandwichscan-app-fastapi

react frontend repository https://github.com/PeterTakahashi/sandwichscan-app-react

Definition / detection rule​

A sandwich attack is detected when:

  1. There is a victim swap.
  2. Within one block before or after the victim’s swap, there are swaps from the attacker.
  3. The attacker’s front-run swap is in the same direction as the victim’s swap.

Data source & scope​

  • Source: Raw swaps + transactions in Ethereum collected by GCP BigQuery.
  • Pools scanned: Uniswap v2, Sushiswap v2, PancakeSwap v2.
  • Time range: since 2020.
  • Only swaps on the Ethereum chain are considered.