用多智能体系统分阶段生成漏洞利用，打通漏洞挖掘、优先级评估与检测规则三者断层。
FORGE: Multi-Agent Graduated Exploitation and Detection Engineering
- 五类智能体协同作业，按四层深度逐步推进漏洞利用，全程可评估。
- 在603个CVE上实现67.8%的利用成功率，每条成本仅1.5美元，跨187种缺陷类型。
- 生成的检测规则精准度高，93.4%无误报，且深层利用产出的规则更可靠。
当前漏洞披露量远超组织评估能力，而概念验证生成、漏洞优先级排序和检测规则工程三个研究领域彼此孤立。现有自动化利用系统仅返回成功/失败二值结果，忽略中间进展，也无法为其他领域提供信号。本文提出FORGE，一个通过渐进式利用深度连接三者的多智能体系统。五个专用智能体（Intel、Generator、Planner、Exploit、Detector）按固定流程执行：(1) 从CVE元数据生成目标易受攻击应用；(2) 在LLM主导的四层评估体系下进行多轮引导式利用（L0：无证据至L3：完全控制）；(3) 基于OpenTelemetry追踪数据生成Sigma与Snort检测规则。渐进深度作为桥梁：更深的利用产生更丰富的行为痕迹用于检测开发，同时深度数据可用于优先级验证。层级知识架构积累评估经验，将构建与利用知识迁移至后续CVE。在603个来自CVE-GENIE的数据集上的评估显示，跨8种语言和187种CWE类型，端到端实现67.8%的L1+利用率，单个漏洞成本仅1.5美元。无论EPSS或CVSS等级如何，利用率均接近68%，表明模式可达性与基于元数据的优先级无关。源自L2+利用的检测规则在跨度归一化覆盖度上显著优于L1规则（p=0.035），且93.4%的生成Snort规则在合成良性数据集中无误报。
原文摘要 · Abstract (English)
Vulnerability disclosure volumes now far exceed organizational assessment capacity, yet three adjacent research communities (proof-of-concept generation, vulnerability prioritization, and detection rule engineering) operate largely in isolation. Existing automated exploit generation systems report binary pass/fail outcomes, discarding partial progress and producing no signal for the other two communities. This paper presents FORGE, a multi-agent system that bridges these three silos through graduated exploitation depth. Five specialized agents (Intel, Generator, Planner, Exploit, and Detector) execute in a fixed pipeline that (1) generates targeted vulnerable applications from CVE metadata, (2) conducts coached, multi-turn exploitation assessed by an LLM-primary oracle on a four-level taxonomy (L0: no evidence through L3: full compromise), and (3) produces Sigma and Snort detection rules grounded in OpenTelemetry exploitation traces. Graduated depth is the bridging mechanism: deeper exploitation yields richer behavioral traces for detection engineering, while depth data across scoring bands provides ground truth for prioritization validation. A tiered knowledge architecture accumulates intelligence across assessments, transferring build and exploitation experience to subsequent CVEs. Evaluation on 603 CVEs from the CVE-GENIE dataset achieves 67.8% end-to-end L1+ exploitation at USD 1.50 per CVE across eight languages and 187 CWE types. Exploitation rates remain near 68% regardless of EPSS or CVSS band, indicating that pattern-level reachability is orthogonal to metadata-based prioritization. Detection rules from L2+ exploitation achieve significantly higher span-normalized grounding than L1-derived rules (p=0.035), and 93.4% of generated Snort rules produce zero false positives against a synthetic benign corpus.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。