Back to Intelligence

Malicious Claude Code Extension in Cursor IDE + AD CS ESC1/ESC8 Exploitation: OTX Pulse Analysis — Enterprise Detection Pack

SA
Security Arsenal Team
November 2, 2026
5 min read

Two recent OTX pulses highlight emerging threats targeting enterprise environments:

  1. Fake Claude Code Extension in Cursor IDE: Between October 29 and November 2, 2026, malicious versions of the Claude Code extension (2.1.312 and 2.1.313) were distributed via Cursor's VS Code extension marketplace. These extensions installed SSH backdoors on macOS developer machines, establishing reverse SSH tunnels to 134.209.82[.]225 and persisting via LaunchAgents. The campaign was live for approximately 11 hours before removal.

  2. AD CS ESC1/ESC8 Exploitation via Teams Phishing: Human-operated intrusions leveraged Microsoft Teams messages and calls to impersonate IT personnel, socially engineering users into granting remote access through Quick Assist. Threat actors staged portable Python, Bun, or Java runtimes, performed network discovery, and assessed Active Directory Certificate Services for authentication paths. The incidents diverged at the certificate layer: one leveraged ESC1 certificate abuse, the other ESC8 NTLM relay.

Both campaigns demonstrate sophisticated attack chains targeting developers and enterprise identity infrastructure.

Threat Actor / Malware Profile

Fake Claude Code Extension (Cursor IDE)

  • Distribution Method: Malicious extension published on Cursor's VS Code marketplace.
  • Payload Behavior: Installs attacker SSH keys, establishes reverse SSH tunnels, creates persistence via LaunchAgents.
  • C2 Communication: Reverse SSH tunnel to 134.209.82[.]225.
  • Persistence Mechanism: LaunchAgents on macOS.
  • Anti-Analysis: Operated for only 11 hours, limiting detection window.

AD CS ESC1/ESC8 Exploitation

  • Distribution Method: Teams phishing, Quick Assist abuse.
  • Payload Behavior: Stages portable runtimes (Python, Bun, Java), performs network discovery, assesses AD CS for ESC1/ESC8 paths.
  • C2 Communication: Uses Rclone for exfiltration.
  • Persistence Mechanism: Certificate-based authentication via AD CS abuse.
  • Anti-Analysis: Uses legitimate tools (Quick Assist, Rclone) to blend in.

IOC Analysis

Indicator Types

  • File Hashes: SHA256, SHA1, MD5 hashes of malicious extensions and payloads.
  • IP Addresses: 134.209.82[.]225 (SSH C2).
  • Domains/URLs: Not present in this dataset.

Operationalization

  • Block Hashes: Add hashes to EDR/AV blocklists.
  • Network Controls: Block outbound SSH to 134.209.82[.]225.
  • Tooling: Use Sigma, KQL, and PowerShell for detection and hunting.

Detection Engineering

YAML
---
title: Malicious Claude Code Extension - SSH Backdoor Persistence
id: 1
status: experimental
description: Detects creation of LaunchAgents for SSH persistence
date: 2026/11/03
author: Security Arsenal
logsource:
  product: macos
  service: auditd
detection:
  selection:
    Image: '/bin/launchctl'
    CommandLine|contains: 'load'
    CommandLine|contains: 'com.apple.launchd'
    CommandLine|contains: 'ssh'
  condition: selection
fields:
  - Image
  - CommandLine
falsepositives:
  - Legitimate SSH agent setup
level: high
---
title: AD CS ESC1 Certificate Abuse
id: 2
status: experimental
description: Detects ESC1 certificate enrollment attempts
date: 2026/11/03
author: Security Arsenal
logsource:
  product: windows
  service: security
detection:
  selection:
    EventID: 4886
    CertificateTemplate: 'User'
    SubjectAlternativeName: '*'
  condition: selection
fields:
  - EventID
  - CertificateTemplate
  - SubjectAlternativeName
falsepositives:
  - Legitimate certificate enrollment
level: medium
---
title: Rclone Exfiltration
id: 3
status: experimental
description: Detects Rclone usage for data exfiltration
date: 2026/11/03
author: Security Arsenal
logsource:
  product: windows
  service: sysmon
detection:
  selection:
    Image: '*rclone.exe'
    CommandLine|contains: 'copy'
  condition: selection
fields:
  - Image
  - CommandLine
falsepositives:
  - Legitimate Rclone usage
level: medium
KQL — Microsoft Sentinel / Defender
// Hunt for malicious Claude Code extension SSH backdoor
DeviceNetworkEvents
| where RemoteIP == "134.209.82.225"
| where InitiatingProcessFileName =~ "ssh" or InitiatingProcessFileName =~ "launchd"
| project TimeGenerated, DeviceName, RemoteIP, InitiatingProcessFileName, InitiatingProcessCommandLine

// Hunt for AD CS ESC1/ESC8 exploitation
SecurityEvent
| where EventID in (4886, 4887)
| where CertificateTemplate == "User" or CertificateTemplate == "WebServer"
| where SubjectAlternativeName contains "*"
| project TimeGenerated, Computer, Account, CertificateTemplate, SubjectAlternativeName

// Hunt for Rclone exfiltration
DeviceProcessEvents
| where FileName =~ "rclone.exe"
| where ProcessCommandLine contains "copy"
| project TimeGenerated, DeviceName, FileName, ProcessCommandLine
PowerShell
# PowerShell IOC Hunt Script
# Check for malicious SSH keys
$sshKeyPath = "$env:USERPROFILE\.ssh\authorized_keys"
if (Test-Path $sshKeyPath) {
    $content = Get-Content $sshKeyPath
    if ($content -match "ssh-rsa.*134\.209\.82\.225") {
        Write-Output "Malicious SSH key found in $sshKeyPath"
    }
}

# Check for LaunchAgents persistence
$launchAgents = Get-ChildItem "~/Library/LaunchAgents" -ErrorAction SilentlyContinue
foreach ($agent in $launchAgents) {
    $plist = plutil -p $agent.FullName
    if ($plist -match "ssh" -and $plist -match "134\.209\.82\.225") {
        Write-Output "Malicious LaunchAgent found: $($agent.FullName)"
    }
}

# Check for Rclone usage
$rclone = Get-Process -Name "rclone" -ErrorAction SilentlyContinue
if ($rclone) {
    Write-Output "Rclone process detected: $($rclone.Path)"
}

# Check for AD CS ESC1/ESC8 abuse
$events = Get-WinEvent -FilterHashtable @{LogName='Security'; ID=4886,4887} -ErrorAction SilentlyContinue
foreach ($event in $events) {
    if ($event.Message -match "CertificateTemplate.*User" -or $event.Message -match "SubjectAlternativeName.*\*") {
        Write-Output "Potential AD CS abuse detected: $($event.Message)"
    }
}

Response Priorities

Immediate

  • Block hashes in EDR/AV.
  • Block outbound SSH to 134.209.82[.]225.
  • Hunt for SSH keys and LaunchAgents persistence.

24h

  • Verify identity of users who installed the malicious extension.
  • Review AD CS enrollment logs for ESC1/ESC8 abuse.

1 Week

  • Harden AD CS configuration (disable ESC1/ESC8 templates).
  • Implement application control for portable runtimes.
  • Educate users on Teams phishing and Quick Assist abuse.

Related Resources

Security Arsenal Incident Response Managed SOC & MDR Services AlertMonitor Threat Detection From The Dark Side Intel Hub

Is your security operations ready?

Get a free SOC assessment or see how AlertMonitor cuts through alert noise with automated triage.