Researcher builds multi-agent LLM system to automate security vulnerability hunting

A new approach uses multiple AI agents working together to discover and reproduce software vulnerabilities without human intervention. For indie developers and small security teams, automated vulnerability detection could shift the economics of security testing—though it's still early-stage research rather than a ready-to-use tool.