MASTA:多智能体系统自动完成 Tamarin 安全协议形式化建模
MASTA: A Feedback-Scheduled Multi-Agent System for End-to-End Tamarin Modeling and Analysis
安全协议形式化验证以往要专家手写规则和引理,MASTA 用三个智能体分工加验证器反馈调度,让 LLM 自动产出 Tamarin 模型,用较弱的 Qwen 骨干还能跑赢 Codex 和 Claude Code,做安全分析的人可以试试。
论文提出 MASTA,一个面向 Tamarin 形式化验证的多智能体系统,用三个角色分工解决 LLM 单体建模的痛点:Architect 负责多重集重写规则建模,Auditor 负责时序逻辑性质建模,Arbiter 只做调度不参与写作。
Arbiter 实现了一个以 Tamarin 验证器输出为锚点的调度器,根据最新验证结论派发下一步任务,并通过带问题标签的修改列表路由修复,配合六类问题分类法和 LLM 去重检查避免死循环。团队还构建了包含 10 个协议、139 个变体的基准,配五项细粒度指标分别衡量正确性(M1–M3)和完整性(M4–M5)。
在 Qwen 3.5-plus 骨干上,MASTA 相比同骨干的 OpenCode 单项指标最高提升 +0.27,正确性与完整性两项平均分均排第一,超过了使用 GPT-5.5 的 Codex 和 Opus 4.7 的 Claude Code。消融实验显示,去掉 Arbiter 后性质建模与验证得分直接归零。
MASTA: A Feedback-Scheduled Multi-Agent System for End-to-End Tamarin Modeling and Analysis
Tamarin-based formal verification of security protocols requires expert-authored multiset rewriting rules and temporal-logic lemmas. Translating a natural-language protocol description into a complete formal model with an LLM is hard because syntactic, reachability, and property-level errors compound across the multi-step modeling process, and a single agent that both authors and self-evaluates cannot reliably localize them. We present MASTA, a multi-agent system that addresses these difficulties through three innovations. First, an operational--evaluative role decomposition assigns rule modeling to an Architect, property modeling to an Auditor, and pure dispatch to a dedicated Arbiter that holds no authoring duties. Second, the Arbiter implements a Tamarin-anchored scheduler that dispatches the next agent based on the latest verifier verdict and routes problem-tagged fixes through a modification list, with a six-class problem taxonomy and an LLM deduplication check preventing routing deadlock. Third, we construct a benchmark of 10 protocols paired with 139 mutation variants and five fine-grained metrics that jointly assess model correctness ($M_1$--$M_3$) and completeness ($M_4$--$M_5$). On this benchmark, MASTA on Qwen 3.5-plus improves over OpenCode on the same backbone by up to $+0.27$ on individual metrics and ranks first on both the correctness and completeness averages, surpassing Codex (GPT-5.5) and Claude Code (Opus 4.7) despite using a weaker LLM. Ablations on five protocols confirm that removing the Arbiter collapses property and verification scores to zero.
本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:Tamarin、MASTA、Qwen、多智能体,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明:本文原始线索采编自「arXiv cs.AI」(@Yongkang Xiao, Qiyi Deng, Jing Chen, Min Shi, Ju Ma, Ruiying Du),经 AITOP 多维评分管线结构化提炼。