@techreport{han-bmwg-agent-security-benchmark-00, number = {draft-han-bmwg-agent-security-benchmark-00}, type = {Internet-Draft}, institution = {Internet Engineering Task Force}, publisher = {Internet Engineering Task Force}, note = {Work in Progress}, url = {https://datatracker.ietf.org/doc/draft-han-bmwg-agent-security-benchmark/00/}, author = {Yu Han and Meiling Chen and Yue Yu and Jianyu Lin}, title = {{Security Evaluation Benchmark for AI Agents}}, pagetotal = 43, year = 2026, month = jul, day = 5, abstract = {This document defines a security evaluation benchmark framework for AI agents. It is divided into two parts: evaluation metrics and evaluation methodology, comprising four first-level evaluation dimensions and 55 second-level metrics, as well as a comprehensive evaluation methodology system covering all scenarios, including static evaluation, dynamic evaluation, attack-defense evaluation, compliance evaluation, and quantitative evaluation. It aims to assess the risk profile of AI agents throughout their entire lifecycle, including perception risks, memory risks, decision-making risks, and execution risks.}, }