Zscalerのブログ

Zscalerの最新ブログ情報を受信

Security Research

Attackers Exploit Optimizely Flaw To Inject Malicious JavaScript

image
ATINDERPAL SINGH
November 18, 2026 - 11 読了時間
>Attackers Exploit Optimizely Flaw To Inject Malicious JavaScript

Introduction

On November 16, 2026, an attacker exploited a vulnerability in Optimizely's Web Experimentation account signup flow. Optimizely is a platform used to create, manage, test, and personalize digital content and websites. The flaw allowed the attacker to edit the Project JavaScript (projectJS) field. Customers use this field to add custom code that runs on pages that load the Optimizely snippet. By abusing this mechanism, the attacker injected malicious JavaScript into customer projects. The code was then delivered to visitors’ browsers through the Optimizely snippets, including snippets served from Optimizely’s content delivery network (CDN).

The attack highlights the continued risk posed by third-party script dependencies and software supply chain compromises, particularly when websites rely on trusted third-party platforms to deliver JavaScript that runs in visitors’ browsers.

Optimizely reported that it blocked the vulnerable path, deployed a fix, removed the injected code from affected projects, ended active sessions, revoked tokens, reset access for affected accounts, and notified impacted customers.

The malicious payload was designed to replace cryptocurrency wallet addresses, skim payment card details, steal cookies, capture page screenshots, and deliver iOS exploits to iPhones and iPads. Organizations using Optimizely, or similar third-party JavaScript platforms, should treat this incident as a reminder to review third-party script governance, monitoring, and browser-side security controls.

ThreatLabz Observations

ThreatLabz identified more than 450 unique domains and more than 3,000 URLs serving the malicious script. These findings suggest a broader impact across organizations and industry verticals than previously reported. The affected domains span industries including the following:

  • Airlines and Travel
  • Automotive
  • Banking and Finance
  • E-commerce
  • Insurance
  • News and Media
  • Property and Real Estate
  • Retail and Fashion
  • Smart Home
  • Sports
  • Technology and SaaS
  • Telecommunications

Recommendations

For end users who visited affected sites

  • Monitor bank and credit card statements for suspicious activity.
  • Review cryptocurrency transactions made on or around November 16, 2026.
  • Be cautious of unexpected account activity on sites visited during the incident.
  • Update iOS and iPadOS devices to the latest available security version.
  • Change passwords for accounts used on potentially affected websites, especially if the same credentials are reused elsewhere.

For organizations using Optimizely

Review Optimizely project configuration

  • Audit all Project JavaScript fields for unauthorized or unexpected code.
  • Review recent changes, user activity, account creation events, and administrative access.
  • Confirm whether any project snippets were modified during the exposure window.
  • Verify that production pages load only approved versions of Optimizely snippets.

Rotate and reset access

  • Reset credentials for affected Optimizely administrative users.
  • Rotate API tokens, session tokens, and integration keys associated with Optimizely.
  • Enforce multi-factor authentication (MFA) for all users with access to experimentation or web personalization platforms.

Deploy Content Security Policy (CSP) in enforcement mode where possible

  • Restrict script execution to approved domains.
  • Limit where pages can send requests and load images and frames using connect-src, img-src, frame-src, and related CSP directives.
  • Use CSP reporting to identify unexpected browser-side activity.

Reduce third-party JavaScript risk

  • Maintain an inventory of third-party scripts loaded across websites.
  • Review whether each third-party script is still required.
  • Apply least-privilege principles to third-party tags and tag management platforms.
  • Monitor third-party script changes and unexpected behavior.

How It Works

Optimizely Web Experimentation allows site owners to run A/B tests, personalization logic, and other front-end experiments by loading JavaScript on customer websites. The Project JavaScript field is designed to let customers add custom JavaScript that executes on any page that loads the Optimizely snippet.

Public reporting indicates that the attacker injected malicious JavaScript into the Project JavaScript field of affected customer projects. The injected code was then delivered through legitimate Optimizely snippets. As a result, when visitors loaded a page with a modified snippet, the injected JavaScript could run in their browsers.

Scope of script activity: The observed JavaScript activity is primarily limited to the browser context of the website where the scripts are loaded. The wallet replacement, clipboard manipulation, form monitoring, card-skimming, screen capture, and network-hooking behaviors operate inside the current page, its accessible DOM, same-origin frames, page-initiated requests, and JavaScript-accessible cookies or browser APIs. By design, this JavaScript cannot directly inspect other browser tabs, unrelated websites, other sites’ cookies, local files, or global browser activity. The main exception is the iOS/iPadOS exploit-delivery branch: if one of the delivered exploit chains succeeds, the resulting impact could potentially exceed normal web-page sandbox restrictions.

Attack chain

The figure below shows how malicious JavaScript reached website visitors through Optimizely snippets.

Attack chain showing how a threat actor injected malicious JavaScript into Optimizely customer projects.

Figure 1: Attack chain showing how a threat actor injected malicious JavaScript into Optimizely customer projects.

Initial access and injection

As stated above, the attacker exploited a vulnerability in Optimizely’s Web Experimentation account signup flow. This flaw allowed unauthorized write access to the Project JavaScript field for multiple customer projects. Modified snippets were served through existing delivery paths, including Optimizely’s CDN (including cdn.optimizely.com) and, in some cases, customer domains. This made the attack more effective because many affected websites already trusted and allowed Optimizely scripts.

Malicious JavaScript payload

In most affected snippets, the injected code acted as a loader that fetched a second-stage JavaScript payload named track.js from attacker-controlled infrastructure such as 1688360[.]cc. The JavaScript payload provided the main malicious functionality, and ran in the visitor’s browser within the context of the affected website.

The following flowchart shows what track.js does after it runs in a visitor’s browser. 

Clipboard hijacking, card skimming, cookie theft, screen capturing, and attempted exploit delivery on iOS and iPadOS.

Figure 2: Clipboard hijacking, card skimming, cookie theft, screen capturing, and attempted exploit delivery on iOS and iPadOS.

Cryptocurrency address replacement 

The script monitors page content, clipboard operations, and network-transmitted data for cryptocurrency wallet addresses. Targeted cryptocurrencies include Bitcoin, Ethereum, Tron, Solana, Litecoin, Toncoin, Bitcoin Cash, and Dogecoin. When the script detects a valid wallet address, the script attempts to replace it with an attacker-controlled wallet address. 

Cookie theft and screen capture 

Cookie theft and screen capture happens only when a valid crypto wallet address is replaced and the element is visible. In that case, the JavaScript payload script loaded the legitimate html2canvas.min.js to capture a screenshot of the page, along with the page URL, cookies, original wallet address, and other metadata. This data is sent to the attacker endpoint https://1688360[.]cc/hit.php, likely so the operator can validate successful replacement activity and capture the victim’s session context. The JavaScript payload could collect browser cookies from the affected website, with some exceptions such as if the HttpOnly flag was set. 

This could cause users to unknowingly send cryptocurrency to the attacker instead of the intended recipient.

Payment card skimming

Later versions of the payload added payment card skimming functionality. The script monitors input and textarea elements, keyboard events, form submissions, and web requests for payment card numbers, expiration dates, and CVV values on the compromised website. Captured card data was to an attacker-controlled endpoint (1688360[.]cc/card.php). Note the skimmer could not read payment details entered inside secure cross-origin payment iframes, such as hosted payment fields from providers like Stripe or Adyen.

Attempted iOS exploit delivery 

For visitors using iPhones or iPads, the malware reportedly attempted to load the Coruna and DarkSword exploit kits. The payload used device and browser checks to decide whether to attempt iOS-specific exploitation. 

When the payload identified a visitor’s device as an iPhone or iPad, it added two script elements to the page to load external JavaScript files. Indicators of these pages are the elements data-channel-code, data-chain-url, and data-cfasync attributes.

The script injected a hidden iframe in some cases (hxxps://eil786npto9u5bq[.]icu/JIUJIU888/weifile[.]html) on devices it identified as running iOS 18.6 or earlier. The iframe is positioned off-screen with zero width and height, suggesting it is not intended to be visible to the user. 

iOS exploit delivery components

  • track.js is the browser-side payload. It checks for iOS or iPadOS and loads additional components when it detects either operating system.
  • boot.js is the main iOS loader. It uses Safari and iOS version information to select the next delivery path.
  • weifile.html provides a separate delivery path for selected iOS and iPadOS versions. It loads index.js.
  • index.js initializes an obfuscated loader, decodes embedded modules, reports device and domain metadata, and may fetch additional modules (if required). 

The following table summarizes the two modules embedded in index.js.

Embedded Module Identifier

Description

57620206d62079baad0e57e6d9ec93120c0f5247

Utility/runtime helper module providing encoding, binary parsing, numeric conversion, URL normalization, and typed-array/DataView helpers.

14669ca3b1519ba2a8f40be287f646d4d7593eb0

Browser environment profiling and anti-analysis module that checks runtime capabilities, storage behavior, WebRTC, MathML rendering, WebAssembly behavior, and automation indicators such as navigator.webdriver.

Table 1: Functions of the two modules embedded in index.js.

Conclusion

The Optimizely incident highlights how attackers can compromise trusted third-party JavaScript supply chains to reach users across many websites at once. By abusing a legitimate experimentation feature, the attacker delivered injected JavaScript through snippets that affected websites already trusted. Note the malicious JavaScript code can run in visitors’ browsers without any obvious change to the page. The card-skimming and wallet-replacement behavior should be understood as page-context activity, not browser-wide monitoring. These scripts act within the compromised site where they are loaded. Broader impact would depend on successful exploitation by the iOS/WebKit exploit chain.

Zscaler Coverage

The Zscaler ThreatLabz team has deployed protections for the indicators of compromise associated with this campaign.

Zscaler Advanced Threat Protection

How Zscaler Can Help

The following table shows Zscaler’s recommended mitigations by attack stage.

Attack Stage

Recommended Mitigation

Prevent compromise

  • Enable SSL/TLS inspection for all traffic, including trusted sources.
  • Use Advanced Threat Protection to detect malicious scripts. 
  • Configure policies to block Newly Registered and Observed Domains.
  • Route server egress through Zscaler Internet Access (ZIA) to detect and block post-compromise activity.
  • Use Advanced Cloud Firewall to extend controls across all ports/protocols.
  • Use Zscaler Zero Trust Browser to isolate high-risk browsing sessions and reduce endpoint exposure to malicious web content.

Prevent lateral threat movement

  • Use Zscaler Private Access (ZPA) to enforce least-privilege user-to-application segmentation.
  • Use Zscaler Deception to detect attacker movement after initial compromise. 
  • Monitor for suspicious access to internal applications following web application crashes or abnormal behavior.

Detect post-exploitation activity

Prevent data loss

  • Inspect outbound traffic across channels with Zscaler Data Loss Prevention (DLP).
  • Monitor for unusual uploads, archive transfers, credential exposure, or data movement following suspected exploitation.


Indicators Of Compromise (IOCs)

Domains and URLs

IOC

Type

Details

1688360[.]cc

Domain

Main domain associated with the threat campaign. Also used for fingerprinting, reporting, and screenshot submission by track.js. Includes the following URLs:

  • 1688360[.]cc/hit.php
  • 1688360[.]cc/html2canvas.min.js
  • 1688360[.]cc/track.js
  • 1688360[.]cc/card.php

eil786npto9u5bq[.]icu

Domain

Hosts another iOS boot.js script and the weifile.html iframe path. Includes the following URLs:

  • eil786npto9u5bq[.]icu/boot.js
  • eil786npto9u5bq[.]icu/JIUJIU888/weifile.html

5vobauua3ojpklx[.]icu

Domain

Hosts one of the iOS boot.js scripts loaded by track.js. Includes the following URL:

  • 5vobauua3ojpklx[.]icu/boot.js

ios-hack[.]icu

Domain

Main loader/API domain used directly by boot.js.

q1ls1ho9x7wsxlu[.]icu

Domain

Secondary C2 base decoded by boot.js and passed into the hidden iframe as c2Base.


Script/file names

IOC

Description

track.js

Injected browser script that monitors crypto wallet activity and initiates the iOS/iPadOS path.

boot.js

iOS loader script that decodes configuration, fingerprints Safari/iOS, and selects the delivery path.

weifile.html

Hidden iframe stage loaded for selected iOS/iPadOS versions.

index.js

Loader script loaded by weifile.html; decodes modules, reports metadata, and may load follow-on modules.

html2canvas

Library loaded from 1688360[.]cc to capture screenshots when wallet replacement is visible.


iOS exploit-kit entry paths

IOC

Description

A8/a.html

Reported Coruna exploit-kit entry targeting iOS 13–17.

B9/b.html

Reported DarkSword exploit-kit entry targeting iOS 18.x.

D11/b.html

Reported DarkSword exploit-kit entry targeting selected iOS 15–17 / 17.2 ranges.

D12/b.html

Reported DarkSword exploit-kit entry targeting selected iOS 18.x ranges.

C10/c.html

Selected by boot.js for iOS 26.


Follow-on JavaScript module pattern

IOC

Description

/<40-hex>.js?=0

Pattern used by index.js to request follow-on JavaScript modules from the same directory as the loader.

/<40-hex>.js?=1

Pattern used by index.js to request follow-on JavaScript modules from the same directory as the loader.

sha256("cecd08aa6ff548c2" + module_id).substring(0, 40) + ".js"

Filename derivation logic used to generate the follow-on module path.


Referenced module IDs and download paths

Referenced Module ID

Download Path

477db22c8e27d5a7bd72ca8e4bc502bdca6d0aba

bef10a7c014b826e9dd645984e80baf313c1635f.js

29b874a9a6cc9fa9d487b31144e130827bf941bb

4a75f0551eba446b4fa35127024a84b71d9688d6.js

9db8a84aa7caa5665f522873f49293e8eebccd5c

3fd66b32c44150acff3dcb80f86c759574148ed5.js

171a7da1934de9e0efb9c1645f4575f88e482873

17480ecc0120292fb6b8b19f2fa134385dcfd0fd.js

91b278ddb2aec817b10c1535e0963da74f9b8eeb

f6377d5d458183d41c5fd99661c5a306b42c6255.js

b586c88246144bc7975ad4e27ec6d62716bf34ea

3bc0f6865c0476c0a98a76cb9924d6b3972df591.js

e3b6ba10484875fabaed84076774a54b87752b8a

6beef463953ff422511395b79735ec990bed65f4.js

57cb8c6431c5efe203f5bfa5a1a83f705cb350b8

8c4451cf1258f9a8d6a8af27864f111fd69a0e99.js

d11d34e4d96a4c0539e441d861c5783db8a1c6e9

ff4f3cb4711fb364b52de5ab04a8f83140466f89.js

ea3da0cfb0a5bdb8c440dd4a963f94cbd39d9e44

46ecd515ac9e99ef0603063db39303a0fd849632.js


Attacker wallets

Wallet Address

Cryptocurrency / Network

0xC4026b06D82b0F0B363eda0642D61CfbDc17884e

Ethereum

THvs6p56XkUiMkJeSc51cAP64u2RsTYqxa

TRON

19g9rZcPeiz8rUvyWpTRyXXhPnjsWrc3kW

Bitcoin

UQAgM1VqSZQiJLSG5w-ovurq5CZdxLEK3_7CMmwALTA5e09p

TON

FVtL8NMYS9PA1v3HE67XUGCtGA9JaZUvqav2fU2LrV5n

Solana

ltc1q3wp7g6ds0ggg6fl6kncpa4sepn0q3qfacgred5

Litecoin

qzkaaf8n6grmpr0d9lx8xutn3r9lk722f566hhpe58

Bitcoin Cash

DBEGrxRF7KMSuhTS4c1MFviStU6efdMnu5

Dogecoin

0:2033556a49942224b486e70fa8beeaeae4265dc4b10adffec2326c002d30397b

TON

form submtited
お読みいただきありがとうございました

このブログは役に立ちましたか?

免責事項:このブログは、Zscalerが情報提供のみを目的として作成したものであり、「現状のまま」提供されています。記載された内容の正確性、完全性、信頼性については一切保証されません。Zscalerは、ブログ内の情報の誤りや欠如、またはその情報に基づいて行われるいかなる行為に関して一切の責任を負いません。また、ブログ内でリンクされているサードパーティーのWebサイトおよびリソースは、利便性のみを目的として提供されており、その内容や運用についても一切の責任を負いません。すべての内容は予告なく変更される場合があります。このブログにアクセスすることで、これらの条件に同意し、情報の確認および使用は自己責任で行うことを理解したものとみなされます。

Zscalerの最新ブログ情報を受信

このフォームを送信することで、Zscalerのプライバシー ポリシーに同意したものとみなされます。