Skip to navigation

Threat reportMalwareTL-2026-6461

Abyssos: New Modular C++ RAT with Hidden VNC, Browser Session Hijacking, Credential Theft and UAC Bypass (Zscaler ThreatLabz)

highACTIVE

Abyssos is a new modular Windows remote administration tool written in C++, first identified by Zscaler ThreatLabz in late June 2026. It provides hidden VNC, browser cookie hijacking via the Chrome debug port, keylogging, clipboard capture, UAC bypass and downloadable post-exploitation modules over a custom AES-GCM encrypted TCP protocol. First disclosed .

Key facts

IdentifiersTL-2026-6461
SeverityHIGH
First disclosed
Affected productsMicrosoft Windows (x64 endpoints)
MalwareAbyssos, Win64.PWS.Abyssos
ToolsPluto (LLVM obfuscator), RDP Wrapper, VNC
MITRE ATT&CK techniquesT1018 Remote System Discovery, T1021.005 Remote Services: VNC, T1027 Obfuscated Files or Information, T1027.007 Obfuscated Files or Information: Dynamic API Resolution, T1049 System Network Connections Discovery, T1055 Process Injection, T1056.001 Input Capture: Keylogging, T1057 Process Discovery, T1059.003 Command and Scripting Interpreter: Windows Command Shell, T1070.004 Indicator Removal: File Deletion, T1095 Non-Application Layer Protocol, T1134.001 Access Token Manipulation: Token Impersonation/Theft and 7 more in the attack flow
MotivationUNKNOWN
Detection rules9 (1 shown free)
Sources6 cited sources

Detection coverage

As of 2026-11-19, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-6461 across Splunk SPL, Microsoft KQL and Sigma, covering 13 indicator(s) of compromise. The whole corpus is readable without an account. Every threat comes with one complete detection in Splunk SPL, Microsoft KQL and Sigma, free for everyone; a free account adds daily TL Creds to unlock the rest. Paid plans add a larger daily allowance, correlation and the MCP server. Threadlinqs MCP server · View plans.

Free detection: Abyssos RAT — Chrome Debug-Port Cookie Hijack and Keylogger/Clone Artifacts (Exploitation/Execution)

Detects Abyssos (Win64.PWS.Abyssos) session hijacking: Chrome launched with --remote-debugging-port=9222 (HVNC_PROG chrome_cdp), the %TEMP%\fontconfigs\cookies.json cookie store created by HVNC_CLONE_START, and the windows_update_cache.json keylogger output in the temp folder. Cookie injection via Network.setCookie can bypass passwords and some MFA, so any hit warrants host isolation and session revocation.

Splunk SPL

index=sysmon sourcetype="XmlWinEventLog:Microsoft-Windows-Sysmon/Operational" ((EventCode=1 Image="*\\chrome.exe" CommandLine="*--remote-debugging-port=9222*") OR (EventCode=11 (TargetFilename="*\\fontconfigs\\cookies.json" OR TargetFilename="*\\Temp\\windows_update_cache.json")))
| eval signal=case(EventCode=1, "chrome_cdp_9222_launch", like(TargetFilename, "%fontconfigs%"), "hvnc_clone_cookie_file", like(TargetFilename, "%windows_update_cache.json"), "keylogger_output_file", true(), "other")
| stats count min(_time) as first_seen max(_time) as last_seen dc(signal) as signal_count values(signal) as signals values(ParentImage) as parent_images values(CommandLine) as command_lines values(TargetFilename) as files by Computer, User
| convert ctime(first_seen) ctime(last_seen)
| sort - signal_count

Microsoft KQL

DeviceProcessEvents
| where Timestamp > ago(24h)
| where FileName =~ "chrome.exe"
| where ProcessCommandLine contains "--remote-debugging-port=9222" or ProcessCommandLine contains "\\fontconfigs"
| where InitiatingProcessFileName !in~ ("code.exe", "node.exe", "chromedriver.exe", "msedgedriver.exe")
| project Timestamp, DeviceName, AccountName, FileName, ProcessCommandLine, InitiatingProcessFileName, InitiatingProcessFolderPath, InitiatingProcessCommandLine, InitiatingProcessSHA256
| order by Timestamp desc

Sigma

title: 'Abyssos RAT Chrome Remote Debugging Port 9222 Cookie Injection'
id: 6f1c2a5e-3b7d-5c1e-9a40-6461a0000001
status: experimental
description: 'Detects Chrome started with --remote-debugging-port=9222 outside developer tooling, the chrome_cdp option Abyssos RAT uses to inject stolen cookies with Network.setCookie.'
references:
    - https://www.zscaler.com/blogs/security-research/abyssos-technical-analysis-new-modular-rat
author: Threadlinqs Team
date: 2026-11-19
tags:
    - attack.credential-access
    - attack.t1539
    - attack.defense-evasion
    - attack.t1550.004
logsource:
    category: process_creation
    product: windows
detection:
    selection_img:
        Image|endswith: '\chrome.exe'
        CommandLine|contains: '--remote-debugging-port=9222'
    filter_dev_parent:
        ParentImage|endswith:
            - '\node.exe'
            - '\code.exe'
            - '\chromedriver.exe'
    condition: selection_img and not filter_dev_parent
falsepositives:
    - 'Browser automation and testing frameworks'
    - 'Developers debugging web applications'
level: high

Known false positives:

  • Developers or QA automation (Selenium, Puppeteer, Playwright) legitimately launching Chrome with --remote-debugging-port=9222
  • IDE debugging sessions that attach to Chrome over the DevTools protocol

Tuning notes: Exclude known automation parents (node.exe, code.exe, chromedriver.exe) and developer hosts via a lookup. Raise priority when Chrome is spawned by a binary under %TEMP% or %APPDATA%, or when fontconfigs/cookies.json appears in the same window. Ensure Sysmon EventCode 1 and 11 are collected into index=sysmon.

8 more detection rules (Splunk SPL, Microsoft KQL, Sigma) · unlock with TL Creds. Sign up free to unlock the detection rules or query them with the Threadlinqs MCP server (Purple)
  • Abyssos RAT — Chrome Launched With Remote Debugging Port 9222 (Cookie Injection via Network.setCookie) · unlock with TL Creds
  • Abyssos RAT — Post-Exploitation Module Artifacts and RDP Wrapper ServiceDll Swap · unlock with TL Creds
  • Abyssos RAT — C2 Network Connections to Known and Community-Related Infrastructure · unlock with TL Creds
  • Abyssos RAT — Cookie Clone, Keylogger and DataScan Module File Artifacts · unlock with TL Creds
  • Abyssos RAT — fodhelper ms-settings Registry Hijack UAC Bypass · unlock with TL Creds
  • Abyssos RAT — C2 Connection to Known and Community-Related Infrastructure · unlock with TL Creds
  • Abyssos RAT — UAC Bypass via fodhelper ms-settings Hijack and ICMLuaUtil COM (Post-Exploitation) · unlock with TL Creds
  • Abyssos RAT — Outbound Custom AES-GCM TCP C2 to Known and Community-Related Infrastructure (C2) · unlock with TL Creds

Indicators of compromise

As of 2026-11-19, Threadlinqs Intelligence tracks 13 indicators of compromise for TL-2026-6461.

Sample indicators (3 of 13)

  • IP address: 213.145.86[.]42
  • File hash: 52b400c5be1557a8df146f62fde76d906e7e0a92ed76788717ef61c758f315aa
  • IP address: 209.99.184[.]223

Defanged, TLP:CLEAR. Full set of 13 indicators: unlock with TL Creds.

13 indicators of compromise · unlock with TL Creds. Compare plans

Community OSINT corroboration

1 of this threat's indicators have also been reported by the open-source security community. Community sightings are unverified and are kept separate from Threadlinqs' curated indicators. Indicator values, reporters and campaign linkage are available to authenticated Red-tier users.

Technical analysis

Abyssos is a modular remote administration tool (RAT) written in C++ and targeting Windows. Zscaler ThreatLabz first identified the family in late June 2026 and published a technical analysis of version 2.4F on 2026-08-10. A second, older build (2.1F) is also referenced in the IOC table, which indicates the tool is under active development. Public reporting describes it as not widely used. The source reporting does not state the delivery chain, initial access vector, victim sectors, geographic targeting or attribution to a named actor, so none is asserted here.

On start, the sample resolves Windows APIs dynamically and checks exports by CRC32 hash. It creates a mutex named Global\[UUID4], with the suffix _Admin or _User depending on privileges. The suffix is _Admin when the -elevated argument is present. It connects to its C2 and registers with a HELLO|...|v2.4F|... message that carries host details such as architecture, hostname, username, integrity level, public IP and country. It then runs a network thread that sends PING every few seconds. Some builds perform anti-analysis: CPUID checks for VMware, KVM, Xen and VirtualBox, and exit if processes such as vmtoolsd.exe, VBoxService.exe, qemu-ga.exe or vdservice.exe are running. The newest version omits these checks. The code is obfuscated with control flow flattening, bogus control flow, constant encryption and stack-based strings. ThreatLabz assesses with medium-to-high confidence that an open-source LLVM obfuscator such as Pluto was used.

The C2 protocol is a custom TCP protocol. Most traffic is AES-GCM encrypted with a hardcoded 32-byte key. The packet layout is a 4-byte data size, a flag byte (default 1), a 12-byte IV, the data and a 16-byte tag. Commands are pipe-delimited, for example PM_KILL|1234. Downloaded modules use AES-CBC (16-byte key/IV) or XOR, and most use the key 1234567890abcdef.

Built-in capabilities include: hidden VNC (HVNC_START, HVNC_INPUT, stop and maximize; 1920x1080 default) that can launch browsers, email clients and shells inside the hidden desktop; browser session hijacking, where HVNC_CLONE_START copies browser data to %TEMP%\fontconfigs and HVNC_PROG chrome_cdp launches Chrome with remote debugging on port 9222 and injects cookies from fontconfigs\cookies.json using Network.setCookie; UAC bypass through fodhelper and the ICMLuaUtil COM interface; a clipboard capture thread (every second); a keylogger whose output is read from windows_update_cache.json in the Windows temp folder; a GRABBER file scanner (by extension and size); in-memory or on-disk ZIP archiving (FM_ARCHIVE); process listing (every 3 seconds) and TCP/UDP connection collection (every 2 seconds) with kill/suspend/resume; file list/delete/upload/download/execute; hosts-file edits, shutdown, reboot, sleep and self-deletion; payload execution (EXECURL deleted 5 seconds after running, hex-encoded executables, and AES-encrypted shellcode injection into a named process); and a remote cmd.exe shell (C2CMD).

Additional modules are downloaded from the C2 into the temp folder with randomized names and prefixes (klog, rcv, rvg, plg, dcf, vul, ds, gc, rdp; DATASCAN uses a datascan.png suffix). They are: KEYLOGGER, RECOVERY (Chrome credentials), RECOVERY_GECKO (Firefox), SENDTXT/SENDTXT2, HDRPFILE, DCFINDER (domain controller discovery), VULNSCAN (vulnerability scanner), ELEVATE_SYS_TOKEN (escalation to SYSTEM), DATASCAN, GRABCOOKIES and RDPWRAP (RDP Wrapper). Zscaler's detection name is Win64.PWS.Abyssos.

Defender impact: cookie injection through the Chrome debugging port can bypass password prompts and some MFA, and the DC finder, vulnerability scanner and RDP wrapper modules point to hands-on-keyboard follow-on activity. Severity is the analyst's judgment, not a vendor rating. Useful detection surfaces are Chrome launched with --remote-debugging-port=9222, files in %TEMP%\fontconfigs, windows_update_cache.json, Global\<UUID>_Admin/_User mutexes, the -elevated argument, fodhelper registry hijack activity, and outbound custom TCP to the listed C2 IPs.

MITRE ATT&CK techniques by tactic:

Discovery

T1018 Remote System Discovery; T1049 System Network Connections Discovery; T1057 Process Discovery

Lateral Movement

T1021.005 Remote Services: VNC

Defense Evasion

T1027 Obfuscated Files or Information; T1027.007 Obfuscated Files or Information: Dynamic API Resolution; T1055 Process Injection; T1070.004 Indicator Removal: File Deletion; T1497.001 Virtualization/Sandbox Evasion: System Checks

Credential Access

T1056.001 Input Capture: Keylogging; T1539 Steal Web Session Cookie; T1555.003 Credentials from Password Stores: Credentials from Web Browsers

Execution

T1059.003 Command and Scripting Interpreter: Windows Command Shell

Command and Control

T1095 Non-Application Layer Protocol; T1573 Encrypted Channel

Privilege Escalation

T1134.001 Access Token Manipulation: Token Impersonation/Theft; T1548.002 Abuse Elevation Control Mechanism: Bypass User Account Control

lateral-movement

T1550.004 Use Alternate Authentication Material: Web Session Cookie

Collection

T1560 Archive Collected Data

Remediation

Immediate actions

  • Block outbound connections to 213.145.86.42 and 209.99.184.223 at the perimeter and hunt historical firewall/proxy logs for them
  • Search endpoints for SHA-256 52b400c5be1557a8df146f62fde76d906e7e0a92ed76788717ef61c758f315aa and ca94d95413210a2a325155740eb8a5c58627ad5c4e704478621e7fc8165fe173
  • Hunt for %TEMP%\fontconfigs, cookies.json, windows_update_cache.json and Chrome started with --remote-debugging-port=9222
  • Isolate affected hosts, and reset passwords and revoke active browser sessions/tokens for users on those hosts

Workarounds

  • Set UAC to always notify and remove local administrator rights from standard users
  • Block execution from user temp directories where feasible (application control)

Longer-term hardening

  • Deploy EDR with behavioral detection for hidden desktop (HVNC) creation, UAC bypass and process injection
  • Alert on registry writes under HKCU\Software\Classes\ms-settings\shell\open\command (fodhelper bypass) and ICMLuaUtil elevation
  • Restrict or alert on browsers launched with remote debugging enabled
  • Use phishing-resistant MFA and short session lifetimes or token binding to limit stolen-cookie replay

Affected and fixed versions

  • Microsoft — Windows (x64 endpoints)
    Vulnerable versions: Windows systems that execute the Abyssos payload; no vulnerability or CVE is exploited per the source

Timeline

  • Zscaler ThreatLabz first identifies the Abyssos RAT family in late June 2026 (exact day not stated in sources; date is approximate)
  • Zscaler ThreatLabz publishes 'Abyssos: Technical Analysis of a New Modular RAT' covering version 2.4F and listing IOCs for builds 2.4F and 2.1F
  • GBHackers reports on the Zscaler research, highlighting HVNC and Chrome debug-port cookie injection
  • SOC Prime releases IOC and behavioral detections associated with the Abyssos report
  • SOC Prime publishes its Abyssos active-threat brief rating it high bias
  • IT-Daily publishes German-language coverage of the Zscaler findings

Sources cited

Drafted by the Threadlinqs research pipeline from 6 sources; published ; last updated —.

Threadlinqs Intelligence — Real-Time Threat Detection Platform

// threat_feed
$ sort --newest
Showing all threats

Live intelligence console

Threat level

Latest Threats