PROVISO: Provenance-Informed Action Selection for Secure LLM Agents
arXiv:2610.27115v1 Announce Type: new Abstract: LLM agents act on the external world through tools and third-party skills, so a single executed action can cause irreversible harm. Existing defenses judge a proposed action without tracking...
Read more: https://arxiv.org/abs/2610.27115
Scan with your phone
Scan with your phone