AI vulnerability tool leads benchmark, but evidence has limits
A preprint reports an evidence-using LLM system led CVSS vulnerability-rating tests, but a small user study and label limits temper the result.
Developing Light ยท https://developinglight.com/editorial/developing-light