Cyber Security News

Hackers Use Malicious Chrome Extension to Break Out of Browser Sandbox

A sophisticated malware campaign that combined a phishing lure, an obfuscated Windows JavaScript dropper, a malicious Google Chrome extension and a Native Messaging Host to effectively break the browser sandbox and execute arbitrary PowerShell commands on infected Windows machines.

The attack chain shows how legitimate platform features signed executables, enterprise extension deployment policies and Chrome’s Native Messaging API can be abused together to create a resilient, remote-command backdoor.

The campaign began with short Italian-language phishing emails claiming a ready invoice (subject line “Fattura #2818999851”).

Recipients saw what looked like a PDF but downloading it produced an obfuscated JavaScript file named with a deceptive extension, Fattura-2819889242.pfd.js, designed to appear like a .pdf at a glance.

When executed by Windows Script Host the JavaScript decoded and wrote two files into the user’s temporary directory: a legitimately signed Epic Games executable and a malicious d3d11.dll.

PowerShell prepared and installed a Chrome extension named Cloud vn105rkj64 (extension ID gghagmhimhgfeajfdmjkgmmehbokmglg) and modified Chrome enterprise policy keys ExtensionInstallAllowlist and ExtensionInstallSources so the extension appeared as an administrator-deployed component.

Extensions alone cannot spawn local processes, but Chrome’s Native Messaging feature provides a sanctioned bridge to native code.

D3 Lab said in a report shared with GBhackers, attacker used DLL side‑loading: the trusted signed executable launched normally but loaded the attacker-controlled DLL, which then launched a hidden PowerShell process

Chrome Extension to Break Out of Browser Sandbox

The attackers registered a Native Messaging Host (com.vn105rkj64.tr7qprrt7g) with a manifest pointing to a locally stored executable and permitted the malicious extension as an allowed origin.

The recipient that a requested invoice was ready for download and was signed by an accounting office. The visible document was presented as a 158 KB PDF.

The phishing message (Source : D3 Lab).

Chrome then connected the extension to the host via a stdio channel and passed length-prefixed JSON messages; the native host executed PowerShell and returned results over the same channel.

Operational telemetry captured the extension contacting a command-and-control server at ext2[.]info. Initial exfiltration included Google cookies, open tabs and URLs, browser user-agent, language settings and a stable victim identifier data sufficient for session hijacking and profiling.

Chrome policy settings (Source : D3 Lab).

Subsequent controller instructions caused directory listings of C:\ to be returned, demonstrating that the extension-native host pairing functioned as an interactive remote-command backdoor rather than a simple information-stealer.

This technique matters because each abused component is legitimate and widely used: signed binaries can be side‑loaded, enterprises can push extensions via policy, and Native Messaging exists for legitimate integrations such as password managers.

Combining them restores the extension’s ability to affect the host OS while preserving a plausible-looking browser-side presence and HTTPS-based C2 communication, complicating detection and response.

Defenders should inspect unexpected Chrome enterprise policy entries and ExtensionInstallSources values, and enumerate Native Messaging hosts under HKCU\Software\Google\Chrome\NativeMessagingHosts.

Manifests’ allowed_origins reveal which extension may launch a host; unknown extension IDs paired with executables in user-writable locations are high-risk.

Endpoint monitoring should correlate events such as temporary csc.exe compilation activity, hidden PowerShell processes, changes to Chrome policy keys, DLL side-loading indicators and outbound POSTs to unusual domains including the extension’s origin header.

Incident response should not stop at removing the extension. Teams must unregister and delete the Native Messaging host, remove malicious executables and DLLs, review PowerShell logs, invalidate exposed browser sessions and rotate credentials where cookies or sessions may have been hijacked.

Indicators of Compromise

TypeIndicatorContext
Email subjectFattura #2818999851Italian invoice lure
Displayed filenameFattura-26189991026.pdfDocument shown in the email
Payload filenameFattura-2819889242.pfd.jsObfuscated Windows JavaScript
Payload MD561f07213f2e54c54ec379714fd211c73Initial JavaScript
Payload SHA-1d7a2361877b9cd1f4b6ef56f59fb7adec72cc945Initial JavaScript
Payload SHA-256b11ef9f11c9bb6228582f38a61f4c04dc7160939d8c5b7ee4e467ffde6317f02Initial JavaScript
Dropped filenameclient_124578.exeSigned application used for DLL side-loading
Dropped EXE SHA-256e77747f06d1d3ee5b8466340a10416874439dd69a7e9cd8653647be7782899b6Side-loading launcher
Dropped filenamed3d11.dllMalicious side-loaded DLL
Dropped DLL SHA-25694f333cba95e76e6b8c0f8831bffc446b5f3c90db2c598c6079a98f1a0ef9701Malicious DLL
Extension nameCloud vn105rkj64Malicious Chrome extension
Extension version2.0Observed manifest version
Chrome extension IDgghagmhimhgfeajfdmjkgmmehbokmglgAllowed extension origin
Extension CRX SHA-256d05e03173d9c841a28af60f5dda8a7c7a39c0a0d7302ec412ac4638b8f9872a3Extension package
Native Messaging Hostcom.vn105rkj64.tr7qprrt7gBridge between Chrome and Windows
C2 domainext2[.]infoConfirmed command-and-control server
Observed IP2.27.5.53Resolution observed during analysis
C2 requestPOST https://ext2[.]info/time.php?q=ste_jstest2Exfiltration and command channel
Related domaincd-nwlins[.]siteContacted during execution; returned parked content
Chrome policy keyHKCU\Software\Policies\Google\Chrome\ExtensionInstallAllowlistExtension installation policy modified by malware
Chrome policy keyHKCU\Software\Policies\Google\Chrome\ExtensionInstallSourcesObserved value: http://localhost:8080/*
Native host keyHKCU\Software\Google\Chrome\NativeMessagingHosts\com.vn105rkj64.tr7qprrt7gExpected registration location for the host

Note: IP addresses and domains are intentionally defanged (e.g., [.]) to prevent accidental resolution or hyperlinking. Re-fang only within controlled threat intelligence platforms such as MISP, VirusTotal, or your SIEM.

Follow us on Google News, LinkedIn, and X to Get Instant Updates and Set GBH as a Preferred Source in Google.

Mayura Kathir

Mayura Kathir is a cybersecurity reporter at GBHackers News, covering daily incidents including data breaches, malware attacks, cybercrime, vulnerabilities, zero-day exploits, and more.

Recent Posts

AI-Assisted Hackers Compromise Government Linux Servers to Deploy Backdoors and Steal Credentials

Hackers used AI-assisted scripts to compromise government Linux servers, harvest administrator credentials, and establish persistent…

47 minutes ago

Researchers Demonstrate iOS 27.0.1 Untethered Jailbreak on iPhone 18 Pro

Security researchers from Calif (Calif.io) demonstrated a working untethered jailbreak on iOS 27.0.1 running on…

15 hours ago

Progress Warns of MOVEit Transfer Vulnerabilities Allowing Authentication Bypass and Privilege Escalation

Progress Software has released fixes for 12 vulnerabilities in MOVEit Transfer, its managed file transfer…

15 hours ago

MITRE ATT&CK v20 Adds New AI Attack Techniques for Prompt Injection and Data Theft

MITRE released ATT&CK v20 on October 13, 2026. It adds the first Enterprise techniques for…

17 hours ago

Google Chrome 155 Patches Critical Use-After-Free Flaw and Nine Additional Security Bugs

Google has released Chrome 155 to the Stable channel with 10 security fixes. One is…

17 hours ago

BAE Systems Cybersecurity Incident Disrupts Bankruptcy Court Operations Across the U.S.

A cybersecurity incident at defense contractor BAE Systems has taken offline the systems that send…

19 hours ago