Agent-security research intensity (share of cs.CR)

Attack surface · source: arXiv cs.CR submissions

What this measures

The share of new arXiv cryptography-and-security (cs.CR) papers that engage with AI agents, LLMs, prompt injection and tool use. A rising share signals the security research field is being pulled toward AI-agent threats, which is evidence the threat surface is intensifying.

Recent values

Part of the Data Pressure Index. History before the first live day is reconstructed from current-method inputs.