paradigmxyz/evmbench
A benchmark harness for running AI agents to find and exploit vulnerabilities in Ethereum smart contracts.
Not currently ranked — collecting fresh signals.
star history
evmbench is a web-based evaluation framework that provides a UI for uploading Solidity smart contracts and running AI agents against them to generate structured vulnerability reports. It includes a frontend (Next.js), backend API (FastAPI), job queue (RabbitMQ), and containerized worker infrastructure (Docker/K8s). The system wraps OpenAI’s frontier-evals evaluation code as a pinned submodule for benchmarking agent performance on security tasks.
Frequently asked
- What is paradigmxyz/evmbench?
- A benchmark harness for running AI agents to find and exploit vulnerabilities in Ethereum smart contracts.
- Is evmbench open source?
- Yes — paradigmxyz/evmbench is open source, released under the Apache-2.0 license.
- What language is evmbench written in?
- paradigmxyz/evmbench is primarily written in TypeScript.
- How popular is evmbench?
- paradigmxyz/evmbench has 427 stars on GitHub.
- Where can I find evmbench?
- paradigmxyz/evmbench is on GitHub at https://github.com/paradigmxyz/evmbench.