The DeFi Walkaway Test: Can Users Actually Leave?
2026-11-05 –, Lotus Stage 3

A protocol can have immutable contracts and still fail the walkaway test: remove the team, frontend, hosted indexer, privileged relayer, and default RPC. Can a fresh machine still discover your position and execute a complete exit using public artifacts alone? I run this experiment across 10+ major Ethereum DeFi protocols and present measured exit success, recovery time, hidden dependencies, and capital exposed.


"The contracts still exist" is a weak test. Users must find positions, construct calls, obtain data, and verify exits without the operator.

I'm building an exit harness for 10+ major DeFi protocols. Each test starts with a historical position and removes the frontend, company API, hosted indexer, relayer, and default RPC one dependency at a time. Public artifacts must find the position and return its assets.

This talk presents the measured results: exit success, recovery time, required infrastructure, and value exposed to each dependency, separating inconvenience from lost agency, and catching immutable contracts that rely on mutable services.

I show real failures and the small repairs that fix them: ABIs, deployment records, permissionless automation, local indexes, fallback interfaces.

Diego is a staff engineer at Zircuit. He uses Go to build secure and reliable distributed systems. His work covers Ethereum infrastructure, protocol security, rollups, and reproducible failure and recovery tests.