LLM Agents Can Autonomously Exploit One-Day Vulnerabilities

It does not develop the exploit from a description of the vulnerability. It extracts public exploit code from the web. https://struct.github.io/auto_agents_1_day.html