PentestCode

PentestCode is an autonomous penetration-testing agent for the terminal. Given a target and a goal, it runs security tools, reads their output, updates a structured picture of the network, and decides what to do next, similar to a human operator. It is a hard fork of OpenCode, rebuilt for offensive security instead of code editing.

It uses a multi-agent architecture with persistent engagement state, so results such as open ports, credentials and vulnerabilities can be queried mid-run with commands such as /status, /vulns or /creds. It supports more than 20 LLM providers and records each step with its evidence chain.

It installs as a single self-contained binary for Linux and macOS, or can be built from source with Bun. The project is in beta.