最新资讯
stackoverflow.blog 欢迎来到芯片时代:半导体制造、AI 数据中心与芯片设计的未来 pub.towardsai.net Netflix 科学家建议重采样 LLM 评判:5 次仅将误差棒缩小 13% pub.towardsai.net Jev 作为 LLM 前的静默层:构建提示感知的模型路由器 reddit.com OpenAI 称 Astra 仅满足 10% 安全任务标准 dev.to FreeToken:在自有 GPU 上运行前沿 MoE 模型及其内存需求 dev.to 当 AI Agent 说“完成”时,究竟完成了什么? dev.to 部署 Underdog Bark 3.0 Pro:逐字转写、自定义后端与分块设置 dev.to 为艰难对话构建 AI:五个 Agent、一条安全正则与一位盲评监督者 techcrunch.com Palo Alto Networks 洽谈以 5 亿美元收购 AI Agent 安全初创公司 Clutch theinformation.com 黄仁勋被要求就涉华间谍指控向国会作简报 www.infoq.com Edera 推出 Object Capability Bounded Authority,试图将凭证与 AI Agent 工作负载隔离 wechat 这波杀疯了!旗舰级模型不限量、1个月视频额度免费领
多元动态
登录 注册
开源工具 dev.to · @kozhevniko

端口 102 上的 161,764 个资产:量化 AA26-231A 所描述的工控攻击面

原文 我的收藏

联合网络安全公告 AA26-231A 于 2026 年 8 月 19 日由 NSA、CISA、FBI、能源部和环保署发布,警告攻击者利用互联网扫描服务定位暴露的西门子 S7 可编程逻辑控制器,并使用基于开源 snap7.dll 和 python-snap7 的 AI 生成脚本,伪装成合法监控工具,通过 S7comm 协议(通常为 TCP 102 端口)读写 PLC 内存、配置和梯形逻辑。公告未给出资产数量,作者用 ZoomEye 补充了 161,764 这一计数,并强调查询范围需与数字本身一样谨慎说明。

AI推荐理由 文章结合 2026 年 8 月联合网络安全公告 AA26-231A,用 ZoomEye 数据量化暴露在 102 端口的西门子 S7 PLC 资产规模,并提及攻击者使用基于 snap7 的 AI 生成脚本,属于 AI 与工控安全交叉的具体案例;但数据来源为第三方扫描平台,口径需谨慎对待。

66

正文

161,764 Assets on Port 102: Sizing the Industrial Control Surface That AA26-231A Described Joint Cybersecurity Advisory AA26-231A warned in August 2026 that threat actors were using internet scanning services to locate exposed Siemens S7 programmable logic controllers. The advisory did not include a count. ZoomEye can supply one, provided the scope of each query is stated as carefully as the number itself. What the advisory established AA26-231A was issued on 2026-08-19 by the NSA, CISA, the FBI, the Department of Energy, and the Environmental Protection Agency. It describes actors using scanning services to find internet-exposed or poorly protected PLCs running outdated software, then employing AI-generated scripts built on the open-source snap7.dll and python-snap7 libraries and disguised as legitimate monitoring tools. Those scripts read and write PLC memory, configuration data, and ladder logic over the S7comm protocol, typically on TCP port 102. The advisory names the S7-200, S7-300, S7-400, S7-1200, and S7-1500 families, including F-series safety controllers. It lists no CVE identifiers and no indicators of compromise, and it refers generally to known vulnerabilities and misconfigurations rather than a single shared defect. Four measurements, four different meanings ZoomEye can be queried at several levels of specificity for this subject, and the resulting numbers describe different things. Presenting them without that distinction would be misleading. | Query | What it measures | Count | Collected | | --- | --- | ---: | --- | | app="Siemens S7" | Assets fingerprinted as Siemens S7 | 173 | 2026-09-16 | | app="Siemens SIMATIC" | Assets fingerprinted as Siemens SIMATIC | 10,160 | 2026-09-16 | | port="102" | Assets with port 102 reachable | 161,764 | 2026-09-16 | | device="plc" | Assets classified as PLC devices | 95,395 | 2026-09-16 | The narrowest query, app="Siemens S7", returns 173 assets. This is the most specific fingerprint and the smallest population. The broader app="Siemens SIMATIC" fingerprint, which covers a wider product family that includes S7 controllers among other devices, returns 10,160. The port-based query returns 161,764 assets with port 102 reachable. Port 102 is associated with S7comm, but reachability on that port does not establish that the responding service is a Siemens controller, and it certainly does not establish that the device is vulnerable. Other services and other vendors' equipment can occupy the same port. The device-class query returns 95,395 assets classified as PLCs. That classification spans all vendors, so it describes the general programmable-controller surface rather than Siemens specifically. Why the spread matters The gap between 173 and 161,764 is not a contradiction. It is the difference between a product fingerprint, which requires the scanner to identify specific product characteristics, and a port reachability check, which requires only that something answers on a port. For an organization assessing its own risk, the useful query depends on the question being asked. An asset owner asking "how many of my Siemens S7 controllers are reachable" needs the narrow fingerprint plus internal inventory data. A researcher asking "how large is the protocol surface associated with S7comm" needs the port query, with the caveat that the result is a protocol surface rather than a product population. The advisory's own framing supports the broader reading. It notes that the targeting activity is broader than Siemens devices and that all PLC owners and operators should apply relevant mitigations. The 95,395 device-class figure is consistent with that framing. What the advisory implies about exposure The advisory's central claim is that actors use internet scanning services to find exposed controllers. That claim makes external exposure the operative risk factor, and it makes exposure reduction the primary mitigation. A controller that is not reachable from the internet cannot be located by an internet scanning service. The advisory lists potential consequences of unauthorized access: disruption of industrial processes, safety incidents, equipment damage, data compromise, cascading effects, and compliance violations. It also notes that firmware updates addressing known vulnerabilities are available through Siemens ProductCERT, while Siemens has stated that the advisory does not describe a new vulnerability in the S7 series. Those two statements are compatible. The advisory addresses exploitation of known weaknesses and misconfigurations in exposed devices, not a single new defect. The response is therefore layered: reduce exposure, update firmware, strengthen authentication, and monitor for anomalous protocol activity. Practical measurement for asset owners External exposure data is most useful to an asset owner as a cross-check rather than as a primary source. An organization that believes its controllers are segmented can query its own address space and compare the result against its network documentation. A discrepancy between the two is actionable regardless of what the global figures show. For that purpose, the narrow product fingerprint is the appropriate query, because it targets the specific device family named in the advisory. The port query is appropriate for verifying that no unexpected service is answering on port 102 anywhere in an owned range. Both queries should be run against owned address space only. Scanning address space an organization does not control is not part of an exposure assessment. Limits The counts above describe what ZoomEye indexed at collection time. They are not counts of vulnerable devices, and they are not counts of compromised devices. Product fingerprints describe identified products. Port counts describe reachable services. Device-class counts describe a classification that spans vendors. The advisory itself supplies no CVE identifiers and no IOCs, so no detection rule in this article should be read as derived from a published indicator. The advisory's value is its description of an active reconnaissance campaign and its recommended mitigations. References NSA, CISA, FBI, DOE, EPA — Joint Cybersecurity Advisory AA26-231A, 2026-08-19 Siemens ProductCERT advisory SSB-104599 ZoomEye — app="Siemens S7" (173), app="Siemens SIMATIC" (10,160), port="102" (161,764), device="plc" (95,395); collected 2026-09-16

来源报道 · 1 篇 · 以下链接均前往原始媒体

← 返回精选动态 原文

更多内容