ARTEX: Autonomous Penetration Testing: LLM Agents Break In and Map Assets

Automate the entire penetration testing process with an LLM-based multi-agent architecture. From planning and executing attacks to analyzing results, the workflow proceeds without human intervention.

1 / 8

Manage asset and exploration graphs simultaneously, with immediate reflection of existing asset data via ScopeSentry integration. All traffic is recorded through a MITM proxy, and a Human-in-the-loop structure requires human approval for dangerous commands.

The Go backend and Next.js frontend are integrated into a single binary for easy deployment. Built on PostgreSQL, it reduces repetitive security operations tasks with automated retesting and status tracking for discovered vulnerabilities.