Break-Test Our ERC-4337 Platform
Budget: $750 – $1,500 USD
Our ERC-4337 gas-sponsorship backend is feature-complete at MVP stage; now I want to push it to the edge before we move to mainnet. Your first mission is pure stress: generate concurrent requests from many independent sources and see how the API stack behaves under sustained, spiky loads. Locust is the framework I’d like you to start with, but feel free to extend it or plug in complementary tooling if it uncovers deeper issues.
Here’s the landscape you’ll be attacking:
• Account-abstraction flow that routes across multiple paymaster providers with automatic fail-over.
• Budget enforcement logic that must never overshoot per-dApp or per-wallet caps.
• Settlement and reconciliation jobs running against Base, Arbitrum and Optimism testnets.
• Read-only widgets embedded in front-ends consuming the same endpoints.
What I need from you
1. A repeatable Locust (or compatible) test suite that spins up distributed workers and floods the API with realistic ERC-4337 calls.
2. Metrics and dashboards highlighting throughput limits, latency degradation, memory/CPU spikes and any non-linear failure behaviours.
3. A written report detailing every weakness you found, how you triggered it, logs or traces that prove it, and suggested fixes.
4. Optional but valued: scripts that intentionally misuse our TypeScript SDK to surface edge-case bugs.
Success criteria
• The suite must run against my staging environment and complete without manual babysitting.
• All issues must be reproducible via the steps you document.
• Recommendations should be actionable and prioritised by risk to a mainnet launch.
If you have a track record working with production ERC-4337 flows, understand L2 quirks, and enjoy trying to break things, let’s talk.
Here’s the landscape you’ll be attacking:
• Account-abstraction flow that routes across multiple paymaster providers with automatic fail-over.
• Budget enforcement logic that must never overshoot per-dApp or per-wallet caps.
• Settlement and reconciliation jobs running against Base, Arbitrum and Optimism testnets.
• Read-only widgets embedded in front-ends consuming the same endpoints.
What I need from you
1. A repeatable Locust (or compatible) test suite that spins up distributed workers and floods the API with realistic ERC-4337 calls.
2. Metrics and dashboards highlighting throughput limits, latency degradation, memory/CPU spikes and any non-linear failure behaviours.
3. A written report detailing every weakness you found, how you triggered it, logs or traces that prove it, and suggested fixes.
4. Optional but valued: scripts that intentionally misuse our TypeScript SDK to surface edge-case bugs.
Success criteria
• The suite must run against my staging environment and complete without manual babysitting.
• All issues must be reproducible via the steps you document.
• Recommendations should be actionable and prioritised by risk to a mainnet launch.
If you have a track record working with production ERC-4337 flows, understand L2 quirks, and enjoy trying to break things, let’s talk.