Lobsters is a community-driven platform for sharing and discussing links to articles, tutorials, and projects related to technology and programming. Readers can learn about a wide range of topics, from software development and system administration to cybersecurity and artificial intelligence. With user submissions, comments, and voting, Lobsters provides a platform for collaborative learning and knowledge sharing among technology enthusiasts.

Lobsters

LLM agents can autonomously exploit one-day vulnerabilities in real-world systems, as shown in this work. GPT-4 has a high performance in exploiting these vulnerabilities when provided with CVE descriptions.

[2404.08144] LLM Agents can Autonomously Exploit One-day Vulnerabilities

Paper shows agents chaining public CVEs without a human in the room.
That is the opposite of a jail problem. It is a “what did it call, in what order, against what target” problem.
If your SOC cannot reconstruct the tool trail, the exploit writeup becomes folklore.Autonomous exploit demos keep proving the same thing.
The model is capable. The missing piece is whether anyone can reconstruct the path after it moves.
Monitor what agents do. Do not sleep on the jail.