跳到主内容
@wquguru
精选88Cloudflare 博客(RSS)技巧与观点

Cloudflare 发布基于多 AI Agent 的安全运营框架详解

Building an evidence-grounded agentic security operations harness on Cloudflare

原文
发到 X
推荐理由

这是一篇极具价值的工程复盘,清晰拆解了多 Agent 协作处理安全警报的完整工作流与防幻觉机制,做安全或 Agent 开发的同学值得参考其架构取舍。

Security alerts rarely arrive one at a time. A single alert can cause a spike across the environment, requiring a human analyst to decide which alerts are related and what they mean. When multiple arrive at the same time, it can quickly overwhelm even a seasoned security analyst. Enter the alert paradox. Now, our built-in, multi-AI-agent security operations harness can handle more of this work at Cloudflare scale.

安全警报很少单独出现。单个警报可能导致整个环境中的警报激增,需要人工分析师判断哪些警报是相关的以及它们的含义。当多个警报同时到达时,即使是有经验的网络安全分析师也会很快被淹没。这就是所谓的“警报悖论”。现在,我们内置的多 AI 代理安全运营编排工具能够在 Cloudflare 规模下处理更多此类工作。

Our Cloudflare Managed Defense AI agent harness speeds up the process of gathering data, connecting and aggregating detections, and accounting for missing sources while new alerts continue to arrive. To further analyze context, we make use of the OpenAI Daybreak Defense Network and our partnership with Anthropic. Cloudflare uses approved OpenAI Daybreak and Anthropic models, including GPT-5.6 Cyber and Mythos, for deeper model-backed analysis. Initial analysis and scoring is done with Clef, Cloudflare’s open-source decision model.

我们的 Cloudflare Managed Defense AI 代理编排工具加速了数据收集、连接与聚合检测项的过程,并在新警报持续到达的同时弥补缺失的数据源。为了进一步分析上下文,我们利用了 OpenAI Daybreak Defense Network 以及与 Anthropic 的合作关系。Cloudflare 使用经过批准的 OpenAI Daybreak 和 Anthropic 模型,包括 GPT-5.6 Cyber 和 Mythos,以进行更深层次的模型支持分析。初步分析和评分由 Clef(Cloudflare 的开源决策模型)完成。

Collecting evidence to understand what each alert means requires a lot of time. Even in a highly sophisticated Security Information and Event Management (SIEM), too much is still left for a human to review. Human analysts must gather data, connect and aggregate detections, and account for missing sources while new alerts continue to arrive.

收集证据以理解每个警报的含义需要大量时间。即使在高度复杂的 Security Information and Event Management (SIEM) 中,仍有太多内容留给人类审查。人工分析师必须收集数据、连接并聚合检测项,并在新警报持续到达的同时考虑缺失的数据源。

Think about every time a human analyst reviews an alert: "Which should we silence? Which action should we take? Which alert should we ignore? Which should we resolve as false positives? Which should we resolve as true positives? Which should trigger our incident team?" We address this predicament with our AI agent strategy. Our approach reduces all of those questions and gives Managed Defense Analysts a quick and consolidated view, directly providing insight into related alerts, admitted evidence, visible gaps, and recommended next steps. The result: cutting back on the time needed to analyze, and creating a hyper focus on actually getting security alerts resolved and mitigations deployed.

回想一下每次人工分析师审查警报时的场景:“我们应该静默哪一个?我们应该采取什么行动?我们应该忽略哪个警报?我们应该将哪个标记为误报?应该将哪个确认为真实威胁?哪个应该触发我们的事件响应团队?”我们通过 AI 代理策略解决了这一困境。我们的方法简化了所有这些疑问,并为 Managed Defense Analysts 提供快速且统一的视图,直接提供对相关警报、已确认证据、可见缺口以及推荐下一步操作的洞察。结果:减少了分析所需的时间,并让人工分析师能够高度专注于实际解决安全警报和部署缓解措施。

Why a single agent fails

单一代理为何失败

Our first prototype showed the limits of one general-purpose agent. We provided the AI agent the whole investigation. It produced useful analysis, but it also hallucinated claims the evidence did not support. Telemetry, detector descriptions, policies, and threat intelligence were flattened into one prompt, which caused their distinct roles to merge together.

我们的第一个原型展示了通用单一代理的局限性。我们将整个调查过程提供给 AI 代理。它产生了有用的分析,但也出现了证据不支持的幻觉声明。遥测数据、检测器描述、策略和威胁情报被扁平化为一个提示词,导致它们各自独特的角色混合在一起。

We saw three recurring problems with our first single-shot AI agent harness:

我们在第一个单次执行的 AI 代理编排工具中发现了三个反复出现的问题:

  • Context became authority. A detection is a hypothesis, not proof that an exploit succeeded or an attack occurred. A broad AI agent can blur that distinction.
  • Scope drifted. An AI agent can query the wrong account, time range, or source. You can’t rely on a language model prompt to be a boundary.
  • Failure disappeared. If a lookup times out, the result may not distinguish "not checked" from "checked and not found."
  • 上下文变成了权威。检测只是一种假设,而非证明漏洞利用成功或攻击发生的证据。一个宽泛的 AI 代理可能会模糊这一区别。
  • 范围发生漂移。AI 代理可能会查询错误的账户、时间范围或数据源。你不能依赖语言模型提示词作为边界。
  • 失败变得隐形。如果查找超时,结果可能无法区分“未检查”和“已检查但未找到”。

To address these challenges we moved evidence collection and scope enforcement into application code, before model analysis begins.

为了解决这些挑战,我们在模型分析开始之前,将证据收集和范围执行移入了应用程序代码中。

Recon first, inference second

先侦察,后推理

It's tempting to put an AI agent at every step. The front half of our harness has none.

在每个步骤都放置 AI 代理很有诱惑力。但我们框架的前半部分没有任何 AI 代理。

Before we even call inference, deterministic code runs a fixed set of reconnaissance workflows with versioned API calls. It collects the customer's identity, detection history, traffic baseline, enforcement outcome, and network observations. Each piece of data is stored with its source, version, and timestamp.

甚至在调用推理之前,确定性代码就会运行一套固定的侦察工作流,使用版本化的 API 调用。它收集客户的身份、检测历史、流量基线、执行结果和网络观测数据。每条数据都与其来源、版本和时间戳一起存储。

Cloudflare sees both the request and the action applied to it. That lets the investigation connect the behavior that triggered an alert with both the control that fired and its outcome.

Cloudflare 同时看到请求及其上应用的操作。这使得调查能够将触发警报的行为与触发的控制措施及其结果联系起来。

The fixed recon snapshot also makes evaluation reproducible. If AI agents fetch their own data, two runs may disagree because their inputs changed. Here, the same snapshot can be replayed, so differences between specialist AI agents’ findings come from interpretation rather than retrieval.

固定的侦察快照也使评估可重现。如果 AI 代理自行获取数据,由于输入发生变化,两次运行可能会得出不同结论。在这里,相同的快照可以被重放,因此专业 AI 代理发现之间的差异源于解释而非检索。

Filter noise early

尽早过滤噪声

Most alerts are not incidents. The same rule often fires repeatedly on a known traffic pattern, and paging Managed Defense Analysts every time makes it easier to miss a real security incident.

大多数警报并非安全事件。同一规则通常会在已知流量模式上反复触发,每次都为托管防御分析师发出页面通知,这更容易导致遗漏真正的安全事件。

We needed a lightweight triage model to compare each alert with its reconnaissance data: Has this event been detected for the customer before? What did Managed Defense Analysts decide previously? Does the traffic look consistent with normal human behavior? Alerts scored with a high likelihood to be false positives skip analysis by the specialist AI agents.

我们需要一个轻量级的分诊模型,用于将每个警报与其侦察数据进行比较:该客户之前是否检测到过此事件?托管防御分析师此前做出了什么决定?流量是否与正常的人类行为一致?被判定为高概率误报的警报将跳过专业 AI 代理的分析。

Clef, running on Workers AI, was the perfect fit for this type of fast agentic reasoning.

Clef(在 Workers AI 上运行)非常适合这种快速的多智能体推理类型。

Known high-volume noise is deterministically classified as passive when it arrives. It remains available as context but does not enter the active queue.

已知的高容量噪声在到达时会被确定性地分类为被动状态。它仍可作为上下文存在,但不会进入活动队列。

Specialist AI agents handle the investigation

专业 AI 代理处理调查

For alerts that need deeper review, a coordinator AI agent runs four specialist AI agents in parallel:

对于需要更深入审查的警报,协调员 AI 代理会并行运行四个专业 AI 代理:

  • Traffic analysis reviews request behavior, historical changes, and enforcement.
  • Customer context reviews earlier alerts, dispositions, and Managed Defense Analysts’ decisions.
  • Global telemetry compares the activity with privacy-preserving Internet-wide signals.
  • Threat intelligence checks indicators already admitted to the alert or case.
  • 流量分析审查请求行为、历史变更和执行情况。
  • 客户上下文会审查之前的警报、处置结果以及托管防御分析师的决策。
  • 全局遥测将活动与保护隐私的互联网范围信号进行比较。
  • 威胁情报检查已纳入警报或案例中的指标。

A synthesis AI agent combines their typed findings into one advisory; it can’t fetch new evidence or choose a classification outside the approved vocabulary. Keeping each task narrow makes unsupported claims easier to catch, and recommendations easier to audit.

综合 AI 代理将它们的类型化发现整合为一份建议;它无法获取新证据,也不能在批准词汇表之外选择分类。保持每个任务的范围狭窄,有助于更容易发现缺乏支持的声明,并使推荐内容更易于审计。

Global context without customer data

不含客户数据的全局上下文

A security tool knows what happened inside the environment it is deployed in, but little about the world beyond it. Cloudflare compares an alert with patterns seen across its global network.

安全工具了解其部署环境内部发生的情况,但对环境之外的世界知之甚少。Cloudflare 将警报与其全球网络上看到的模式进行比较。

For example, an IP may be targeting one site, scanning thousands of sites, or appearing for the first time. Those patterns carry different weights. To preserve customer privacy, the global telemetry specialist works only with aggregates; it never receives another customer's individual records or identity.

例如,某个 IP 可能正在针对一个站点、扫描数千个站点,或是首次出现。这些模式具有不同的权重。为了保护客户隐私,全局遥测专家仅使用聚合数据工作;它永远不会接收其他客户的个人记录或身份信息。

This view combines features from Cloudflare's CDN, WAF, DDoS, Turnstile, Rate Limiting, and Cloudforce One threat intelligence. The synthesis AI agent weighs both global reputation and customer history. This ensures a globally common pattern can still be used on a per-customer basis, but does not automatically imply a widespread campaign for all customers.

此视图结合了 Cloudflare 的 CDN、WAF、DDoS、Turnstile、Rate Limiting 和 Cloudforce One 威胁情报的功能。综合 AI 代理权衡全局声誉和客户历史。这确保了全球常见的模式仍可按客户基础使用,但不会自动暗示所有客户都面临广泛的活动。

History is evidence

历史即证据

Every evaluation is conscious of what came before it: the alert, the pattern, and which customer. The recon dossier records the alert's own track record including how many times this service alert has fired, how many of those were dispositioned as false positives, and what the Managed Defense Analyst concluded. An attack pattern that has been benign every time your Managed Defense Analysts have seen it is a very different object from the first sighting of something new, and the specialist AI agents are told which one they are looking at. Approved background context is retrieved from previous alerts and cases, so yesterday's conclusions are carried into today's decision, instead of being rebuilt from scratch.

每次评估都意识到此前发生的情况:警报、模式以及涉及的客户。侦察档案记录了警报自身的历史记录,包括该服务警报触发的次数、其中有多少次被处置为误报,以及托管防御分析师得出的结论。每当托管防御分析师看到时都表现为良性的攻击模式,与首次发现新事物时的对象截然不同,而专门的 AI 代理会被告知它们正在查看的是哪一种情况。已批准的背景上下文从之前的警报和案例中检索,因此昨天的结论会带入今天的决策中,而不是从头重新构建。

Our system aggregates related alerts into a consolidated case. In each case, we store evidence, findings, and recommendations. Our system deterministically joins and correlates this data, but we leave it to a Managed Defense Analyst to confirm the actual scope. Over time, a case can connect network, application, and Zero Trust evidence together, while tracking the source of the evidence for additional context and reference purposes.

我们的系统将相关警报聚合为一个综合案例。在每个案例中,我们存储证据、调查结果和建议。我们的系统确定性地关联和关联这些数据,但将确认实际范围的工作留给托管防御分析师。随着时间的推移,一个案例可以将网络、应用程序和零信任证据连接在一起,同时跟踪证据的来源,以提供额外的上下文和参考目的。

From evidence to decision

从证据到决策

更进一步:量化金融体系

看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力

进入量化体系 →

相似阅读

关联信息,但可能不是同一事件