The OTX intelligence feed reveals three convergent threats targeting enterprise credential stores and identity infrastructure. First, the Psychedelic Stealer campaign utilizes sophisticated ClickFix social engineering—fake Cloudflare CAPTCHA pages hosted on compromised Ukrainian SMB websites—to trick users into manually executing msiexec commands. This attack chain achieved a 14% success rate (79 infections from 426 clicks), delivering an infostealer that establishes a native-messaging bridge via browser extension for persistent credential theft.
Second, CVE-2026-73570 represents a critical infrastructure vulnerability: unauthenticated OS command injection in Zimbra Collaboration Suite's SNMP notification path. Exploitation occurred in the wild between July 20, 2026 (patch availability) and August 13, 2026 (public disclosure), with attackers deploying JSP webshells including China Chopper and GodzillaWebShell to establish persistent access to mail servers.
Third, the MadApe phishing kit targets Microsoft 365 environments through Device Code phishing—a technique that bypasses traditional credential harvesting by abusing legitimate OAuth flows. Multi-stage campaigns combine business-themed PDF lures with fake document-sharing pages to capture session tokens without triggering password-based detection controls.
Additionally, unauthorized AI relay services are compromising developer environments by installing legitimate OpenAI Codex and Anthropic Claude Code CLI tools, then reconfiguring them to route traffic through adversary-controlled infrastructure. This affects semiconductor, manufacturing, and technology sectors across China, Taiwan, Malaysia, and South Korea.
Collectively, these campaigns target the credential lifecycle: initial theft (Psychedelic), lateral movement via compromised mail servers (CVE-2026-73570), cloud account takeover (MadApe), and intellectual property exfiltration via AI service abuse. The convergence suggests a mature criminal ecosystem monetizing enterprise access through multiple vectors.
Threat Actor / Malware Profile
Psychedelic Stealer
Distribution: ClickFix social engineering via compromised Ukrainian websites (uasputnik.com, fsputnik.com). Victims encounter fake Cloudflare verification pages instructing them to press Win+R and execute msiexec commands pointing to https://uasputnik.com/elita.msi.
Payload Behavior: The MSI package installs psychedeliclove.exe, which deploys a browser extension utilizing native-messaging bridge architecture. This enables the malware to bypass browser sandboxing and communicate directly with the native host process for credential exfiltration from browser storage, cryptocurrency wallets, and password managers.
C2 Communication: The malware communicates with IP-based infrastructure (107.175.82.242:9000) using HTTP/S. The /wilow/ path suggests directory-based payload staging.
Persistence: Browser extension installation ensures survival across browser updates and provides continuous access to session cookies and stored credentials.
MadApe Phishing Kit
Distribution: Multi-stage phishing campaigns utilizing business-themed PDF lures with configurable redirect chains through domains including sucdens.com, cepoffices.com, and zakosta.info.
Payload Behavior: Rather than harvesting credentials directly, MadApe abuses Microsoft's Device Code authentication flow (RFC 8628). Victims are directed to legitimate Microsoft login pages but enter device codes provided by the attacker, granting the adversary OAuth tokens without capturing passwords.
Anti-Analysis: The kit employs redirect chain obfuscation and dynamic landing pages that mimic legitimate document-sharing services.
Zimbra CVE-2026-73570 Exploitation
Attack Vector: Unauthenticated OS command injection via SNMP notification path in Zimbra Collaboration Suite.
Post-Exploitation: Deployment of JSP webshells (China Chopper, GodzillaWebShell) and reverse shells to establish persistent access. The psk1zim.abrdns.com hostname suggests DNS-based C2 or exfiltration channels.
Privilege Escalation: Attackers achieve SYSTEM/root-level privileges on mail servers, providing access to organizational communication archives and authentication infrastructure.
Unauthorized AI Relay Services
Mechanism: Installation of legitimate AI CLI tools (OpenAI Codex, Claude Code) followed by reconfiguration of API endpoints to route through adversary infrastructure (aiok.me, sorryios.io, nuoda.vip). This enables interception of proprietary code, API keys, and intellectual property while maintaining the appearance of legitimate tool usage.
IOC Analysis
Network Indicators:
- Domains: uasputnik.com, fsputnik.com (Psychedelic C2); sucdens.com, cepoffices.com, zakosta.info (MadApe redirect chains); aiok.me, sorryios.io, nuoda.vip (AI relay infrastructure)
- IPs: 107.175.82.242 (Psychedelic payload delivery)
- Hostnames: psk1zim.abrdns.com (Zimbra exploitation C2)
File Indicators:
- SHA256 hashes for psychedeliclove.exe variants and JSP webshells
- MD5 hashes for rapid signature-based blocking
URL Indicators:
- https://uasputnik.com/elita.msi (payload delivery)
- http://107.175.82.242:9000/wilow/psychedeliclove.exe (direct executable download)
Operationalization: SOC teams should implement these indicators in EDR network filters, proxy blocklists, and email security gateways. The ClickFix campaign domains should be added to DNS sinkholes. For MadApe, monitor Azure AD sign-in logs for device code authentication flows from suspicious IP ranges. CVE-2026-73570 indicators should trigger vulnerability scans for unpatched Zimbra instances.
Detection Engineering
---
title: Psychedelic Stealer ClickFix Execution
id: 8f3e2a1c-4b5d-4e6f-9a8b-7c6d5e4f3a2b
status: experimental
description: Detects ClickFix social engineering technique leading to msiexec execution of Psychedelic Stealer payloads
author: Security Arsenal
references:
- https://connect.securonix.com/threat-research-intelligence-62/the-psychedelic-stealer-when-the-captcha-is-the-installer-582
date: 2026/11/01
tags:
- attack.t1203
- attack.t1218
detection:
selection:
Image|endswith: '\msiexec.exe'
CommandLine|contains:
- 'uasputnik.com'
- 'fsputnik.com'
- '/i http'
condition: selection
falsepositives:
- Legitimate administrative software installation via MSI
level: high
---
title: Zimbra CVE-2026-73570 Exploitation Attempt
id: 9a4f3b2d-5c6e-4f7a-8b9c-0d1e2f3a4b5c
status: experimental
description: Detects exploitation attempts against Zimbra Collaboration Suite SNMP notification path
author: Security Arsenal
date: 2026/11/01
tags:
- attack.t1190
- attack.t1059
detection:
selection:
Image|endswith:
- '\java.exe'
- '\zmmailboxd'
CommandLine|contains:
- 'snmp'
- 'notification'
filter:
CommandLine|contains: 'zimbra'
condition: selection and not filter
falsepositives:
- Zimbra administrative scripts
level: critical
---
title: MadApe Device Code Phishing Indicators
id: 1a2b3c4d-5e6f-4a7b-8c9d-0e1f2a3b4c5d
status: experimental
description: Detects suspicious process execution associated with MadApe phishing kit infrastructure
author: Security Arsenal
date: 2026/11/01
tags:
- attack.t1566
- attack.t1550
detection:
selection:
Image|endswith:
- '\chrome.exe'
- '\msedge.exe'
- '\firefox.exe'
CommandLine|contains:
- 'sucdens.com'
- 'cepoffices.com'
- 'zakosta.info'
- 'nsboards.com'
condition: selection
falsepositives:
- Legitimate browsing to these domains (unlikely)
level: high
// Hunt for Psychedelic Stealer ClickFix execution and C2 communication
DeviceProcessEvents
| where TimeGenerated > ago(7d)
| where FileName =~ "msiexec.exe"
| where ProcessCommandLine has_any ("uasputnik.com", "fsputnik.com")
| project TimeGenerated, DeviceName, AccountName, ProcessCommandLine, SHA256
| join kind=leftouter (
DeviceNetworkEvents
| where TimeGenerated > ago(7d)
| where RemoteIP == "107.175.82.242" or RemoteUrl has_any ("uasputnik.com", "fsputnik.com")
) on DeviceName
| summarize by DeviceName, AccountName, ProcessCommandLine, RemoteIP, RemoteUrl
// Hunt for MadApe Device Code authentication anomalies
let MadApeDomains = dynamic(["sucdens.com", "cepoffices.com", "zakosta.info", "nsboards.com", "adastraerrans.com"]);
DeviceNetworkEvents
| where TimeGenerated > ago(7d)
| where RemoteUrl has_any (MadApeDomains)
| project TimeGenerated, DeviceName, InitiatingProcessFileName, RemoteUrl, RemoteIP
| join kind=inner (
DeviceProcessEvents
| where TimeGenerated > ago(7d)
| where FileName in~ ("chrome.exe", "msedge.exe", "firefox.exe")
) on DeviceName
| summarize by DeviceName, RemoteUrl, InitiatingProcessFileName
// Hunt for Zimbra exploitation and webshell deployment
DeviceProcessEvents
| where TimeGenerated > ago(7d)
| where FileName =~ "java.exe" or FolderPath has "zimbra"
| where ProcessCommandLine has_any ("snmp", "notification", "cmd", "bash", "wget", "curl")
| project TimeGenerated, DeviceName, ProcessCommandLine, FolderPath, SHA256
| join kind=leftouter (
DeviceFileEvents
| where TimeGenerated > ago(7d)
| where FolderPath has ".jsp"
| where SHA256 in ("dee5af1c0f76b45d28bafd6e60c07bb8e391d98addf81ef8f13d073acdb3c48a", "aea991f694911e321b0ab97534f2ad0291c392c0a43dabff664c563618bd036d")
) on DeviceName
# OTX Threat Intelligence IOC Hunt Script
# Checks for Psychedelic Stealer, MadApe, and Zimbra exploitation artifacts
$ErrorActionPreference = 'SilentlyContinue'
# Psychedelic Stealer Registry Persistence
Write-Host "[+] Checking for Psychedelic Stealer registry artifacts..." -ForegroundColor Cyan
$PsychedelicKeys = @(
"HKCU:\Software\Microsoft\Windows\CurrentVersion\Run\Psychedelic",
"HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\Run\Psychedelic",
"HKCU:\Software\Classes\psychedeliclove"
)
foreach ($key in $PsychedelicKeys) {
if (Test-Path $key) {
Write-Host "[!] SUSPICIOUS: Found registry key $key" -ForegroundColor Red
Get-ItemProperty $key | Select-Object *
}
}
# Check for psychedeliclove.exe in common locations
$MalwarePaths = @(
"$env:TEMP\psychedeliclove.exe",
"$env:APPDATA\psychedeliclove.exe",
"$env:LOCALAPPDATA\psychedeliclove.exe",
"$env:ProgramData\psychedeliclove.exe"
)
foreach ($path in $MalwarePaths) {
if (Test-Path $path) {
Write-Host "[!] SUSPICIOUS: Found file $path" -ForegroundColor Red
Get-FileHash $path -Algorithm SHA256
}
}
# Browser extension check for native messaging bridge
Write-Host "[+] Checking browser native messaging hosts..." -ForegroundColor Cyan
$NativeMessagingPaths = @(
"$env:LOCALAPPDATA\Google\Chrome\User Data\NativeMessagingHosts",
"$env:APPDATA\Mozilla\NativeMessagingHosts",
"HKLM:\SOFTWARE\Google\Chrome\NativeMessagingHosts",
"HKCU:\SOFTWARE\Google\Chrome\NativeMessagingHosts"
)
foreach ($path in $NativeMessagingPaths) {
if (Test-Path $path) {
Get-ChildItem $path -Recurse -Filter "*.json" | ForEach-Object {
$content = Get-Content $_.FullName -Raw | ConvertFrom-Json
if ($content.path -like "*psychedelic*" -or $content.name -like "*psychedelic*") {
Write-Host "[!] SUSPICIOUS: Found native messaging config $($_.FullName)" -ForegroundColor Red
$content
}
}
}
}
# Network connections to known C2
Write-Host "[+] Checking active network connections to C2 infrastructure..." -ForegroundColor Cyan
$C2IPs = @("107.175.82.242")
$Connections = Get-NetTCPConnection -State Established
foreach ($conn in $Connections) {
if ($C2IPs -contains $conn.RemoteAddress) {
Write-Host "[!] SUSPICIOUS: Active connection to C2 IP $($conn.RemoteAddress):$($conn.RemotePort)" -ForegroundColor Red
$conn | Select-Object LocalAddress, LocalPort, RemoteAddress, RemotePort, OwningProcess
Get-Process -Id $conn.OwningProcess | Select-Object ProcessName, Path
}
}
# Check DNS cache for malicious domains
Write-Host "[+] Checking DNS cache for malicious domains..." -ForegroundColor Cyan
$MaliciousDomains = @("uasputnik.com", "fsputnik.com", "sucdens.com", "cepoffices.com", "zakosta.info", "psk1zim.abrdns.com", "aiok.me", "sorryios.io", "nuoda.vip")
$DNSCache = Get-DnsClientCache
foreach ($entry in $DNSCache) {
foreach ($domain in $MaliciousDomains) {
if ($entry.Entry -like "*$domain*") {
Write-Host "[!] SUSPICIOUS: DNS cache entry for $($entry.Entry)" -ForegroundColor Red
}
}
}
# Check for Zimbra webshells
Write-Host "[+] Checking for suspicious JSP files (Zimbra webshells)..." -ForegroundColor Cyan
$WebDirs = @("C:\zimbra\jetty\webapps", "C:\opt\zimbra\jetty\webapps", "/opt/zimbra/jetty/webapps")
foreach ($dir in $WebDirs) {
if (Test-Path $dir) {
Get-ChildItem $dir -Recurse -Filter "*.jsp" | ForEach-Object {
$hash = Get-FileHash $_.FullName -Algorithm SHA256
$SuspiciousHashes = @("dee5af1c0f76b45d28bafd6e60c07bb8e391d98addf81ef8f13d073acdb3c48a", "aea991f694911e321b0ab97534f2ad0291c392c0a43dabff664c563618bd036d")
if ($SuspiciousHashes -contains $hash.Hash.ToLower()) {
Write-Host "[!] SUSPICIOUS: Found webshell $($_.FullName) with hash $($hash.Hash)" -ForegroundColor Red
}
}
}
}
# Scheduled tasks check
Write-Host "[+] Checking scheduled tasks for persistence..." -ForegroundColor Cyan
$Tasks = Get-ScheduledTask | Where-Object {$_.TaskName -like "*psychedelic*" -or $_.TaskName -like "*update*"}
foreach ($task in $Tasks) {
if ($task.Actions[0].Execute -like "*powershell*" -or $task.Actions[0].Execute -like "*cmd*") {
Write-Host "[!] SUSPICIOUS: Scheduled task $($task.TaskName) - $($task.Actions[0].Execute) $($task.Actions[0].Arguments)" -ForegroundColor Red
}
}
Write-Host "[+] IOC Hunt Complete" -ForegroundColor Green
Response Priorities
Immediate (0-24 hours):
- Block all IOCs at network perimeter: 107.175.82.242, uasputnik.com, fsputnik.com, sucdens.com, cepoffices.com, zakosta.info, aiok.me, sorryios.io, nuoda.vip, psk1zim.abrdns.com
- Hunt for msiexec execution with HTTP arguments in process logs
- Audit browser extensions across the enterprise for unauthorized native messaging hosts
- Verify Zimbra Collaboration Suite patch level; if unpatched, isolate immediately
- Check for JSP webshells in Zimbra web directories
24 Hours:
- Force password resets for users who may have interacted with ClickFix pages or MadApe phishing lures
- Review Azure AD sign-in logs for device code authentication anomalies
- Audit Microsoft 365 mailbox rules and OAuth app consents for unauthorized additions
- Check for unauthorized AI CLI installations (OpenAI Codex, Claude Code) and verify API endpoint configurations
- Review email gateway logs for PDF attachments from MadApe domains
1 Week:
- Implement application whitelisting to prevent msiexec execution from user-initiated contexts
- Deploy browser isolation for untrusted websites to prevent ClickFix-style social engineering
- Harden Zimbra instances by restricting SNMP access and implementing network segmentation
- Enable Microsoft 365 Device Code Flow restrictions via Conditional Access policies
- Implement DNS filtering to block newly registered domains (NRDs) commonly used in phishing campaigns
- Conduct security awareness training focusing on ClickFix tactics and device code phishing recognition
- Review and restrict OAuth app permissions to prevent token 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.