Skip to navigation

Threat reportMalwareTL-2026-6604

Bazuka Android Malware Framework Uses Fake Screen Overlays for Credential Theft and Device Control

highACTIVE

ThreatFabric reports Bazuka, a modular Android fraud framework (dropper/downloader, core and 23 DEX modules) found in late October 2026 while investigating an NFC-relay banking trojan. It is delivered via droppers posing as a 'premium' TikTok app (com.tiktok.premium) and provides overlays, keylogging, SMS/2FA interception, crypto clipboard hijacking, remote control, ADB/Wireless Debugging abuse and NFC card interaction, with an LLM-assisted operator panel.

Key facts

IdentifiersTL-2026-6604
SeverityHIGH
First disclosed
Affected productsGoogle Android (devices permitting sideloaded apps and Accessibility/Wireless Debugging grants)
MalwareBazuka, RemoteControl
MITRE ATT&CK techniquesT1407 Download New Code at Runtime, T1417.001 Input Capture: Keylogging, T1417.002 Input Capture: GUI Input Capture, T1429 Audio Capture, T1437.001 Application Layer Protocol: Web Protocols, T1453 Abuse Accessibility Features, T1513 Screen Capture, T1517 Access Notifications, T1521.001 Encrypted Channel: Symmetric Cryptography, T1582 SMS Control, T1604 Proxy Through Victim, T1623.001 Command and Scripting Interpreter: Unix Shell and 8 more in the attack flow
MotivationFINANCIAL
Target sectorsfinance, cryptocurrency, consumer
Detection rules9 (1 shown free)
Sources7 cited sources

Detection coverage

As of 2026-11-20, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-6604 across Splunk SPL, Microsoft KQL and Sigma, covering 26 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: Bazuka C2 - brokerok.org, gargona.cash and Known Infrastructure IPs in DNS, Proxy and Firewall Logs

Detects egress to Bazuka C2: brokerok.org subdomains (c2ws, c2data, c2proxy, c2api) fronted by Cloudflare, gargona.cash, vendor-reported IPs 95.181.213.21 and 170.168.61.117 (ports 8082/8443/8444/8445), and community-observed sibling IPs tagged as lower-tier. Covers the Control and Data WebSocket channels (ChaCha20-Poly1305) and the reverse-proxy module.

  • Languages: Splunk SPL, Microsoft KQL, Sigma
  • ATT&CK: T1437.001, T1521.001, T1604, T1646
  • Data sources: index=dns (DNS query logs), index=proxy (web proxy logs), index=firewall (firewall traffic logs)
  • Severity: high
  • Confidence: high

Splunk SPL

(index=dns OR index=proxy OR index=firewall) earliest=-24h
``` Bazuka C2: brokerok.org subdomains (c2ws/c2data/c2proxy/c2api), gargona.cash, vendor IPs, plus community-observed sibling IPs ```
| eval host_ioc=rtrim(lower(coalesce(query, url_domain, dest_host)),".")
| eval ip_ioc=coalesce(dest_ip, dest)
| eval c2_domain=if(match(host_ioc,"(^|\.)(brokerok\.org|gargona\.cash)$"),1,0)
| eval vendor_ip=if(in(ip_ioc,"95.181.213.21","170.168.61.117"),1,0)
| eval community_ip=if(in(ip_ioc,"176.125.243.101","91.92.43.93","87.199.197.84","193.27.90.102","103.77.240.232"),1,0)
| where c2_domain=1 OR vendor_ip=1 OR community_ip=1
| eval ioc_tier=if(c2_domain=1 OR vendor_ip=1,"vendor_confirmed","community_observed")
| stats count as events dc(host_ioc) as distinct_hosts values(host_ioc) as c2_hosts values(ip_ioc) as c2_ips values(dest_port) as ports min(_time) as first_seen max(_time) as last_seen by src ioc_tier
| convert ctime(first_seen) ctime(last_seen)
| sort - events

Microsoft KQL

let c2_domains = dynamic(["brokerok.org", "gargona.cash"]);
let vendor_ips = dynamic(["95.181.213.21", "170.168.61.117"]);
let community_ips = dynamic(["176.125.243.101", "91.92.43.93", "87.199.197.84", "193.27.90.102", "103.77.240.232"]);
union isfuzzy=true
(DeviceNetworkEvents
 | where Timestamp > ago(24h)
 | where RemoteIP in (vendor_ips) or RemoteIP in (community_ips) or RemoteUrl has_any (c2_domains)
 | project Timestamp, DeviceName, RemoteIP, RemotePort = toint(RemotePort), RemoteUrl, Source = "DeviceNetworkEvents"),
(DnsEvents
 | where TimeGenerated > ago(24h)
 | where Name has_any (c2_domains)
 | project Timestamp = TimeGenerated, DeviceName = Computer, RemoteIP = tostring(IPAddresses), RemotePort = toint(int(null)), RemoteUrl = Name, Source = "DnsEvents")
| extend Tier = iff(RemoteIP in (vendor_ips) or RemoteUrl has_any (c2_domains), "vendor_confirmed", "community_observed")
| summarize Hits = count(), FirstSeen = min(Timestamp), LastSeen = max(Timestamp), Ports = make_set(RemotePort, 10), Hosts = make_set(RemoteUrl, 20), IPs = make_set(RemoteIP, 20) by DeviceName, Tier
| order by Hits desc

Sigma

title: Bazuka C2 - DNS Resolution of brokerok.org and gargona.cash Infrastructure
id: 0d93b6f8-2c47-4e1a-8a5d-9b4e7f31c6a2
status: experimental
description: Detects DNS queries for Bazuka C2 infrastructure - Cloudflare-fronted brokerok.org subdomains (c2ws, c2data, c2proxy, c2api) used for the Control and Data WebSocket channels and the reverse proxy module, and gargona.cash.
references:
    - https://www.threatfabric.com/blogs/bazuka-the-fraud-framework-that-can-do-it-all
    - https://thehackernews.com/2026/11/bazuka-android-malware-uses-fake-screen.html
author: Threadlinqs Team
date: 2026-11-20
tags:
    - attack.command_and_control
    - attack.t1437.001
    - attack.t1521.001
    - attack.t1604
    - attack.exfiltration
    - attack.t1646
logsource:
    category: dns
detection:
    selection_exact:
        query:
            - 'brokerok.org'
            - 'gargona.cash'
    selection_sub:
        query|endswith:
            - '.brokerok.org'
            - '.gargona.cash'
    condition: 1 of selection_*
falsepositives:
    - Security researchers or threat-intel sandboxes resolving the published IOCs
    - Threat-intel feed ingestion that logs lookups of blocklisted domains
level: critical

Known false positives:

  • Threat-intel sandboxes or researchers resolving published IOCs
  • Community sibling IPs may be reused or reassigned shared hosting and can be benign

Tuning notes: Field names (query, url_domain, dest_host, dest_ip, dest_port, src) follow CIM-style normalisation; align them to your sources. Treat ioc_tier=community_observed as hunt-grade only (community corroboration, two sightings on the vendor IPs, sibling IPs uncorroborated); alert on vendor_confirmed. Vendor IP 95.181.213.21 was offline at research time, so domains are the more durable signal.

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 and Gold in full, 2 calls a day on a free account)
  • Bazuka Dropper Artifacts - TikTok18 Dropper Filename and Published APK Hashes in Defender Telemetry · unlock with TL Creds
  • Bazuka Modules - Keylogger, Overlay, SMS, Clipper and ADB Module Evidence in Alerts · unlock with TL Creds
  • Bazuka C2 - Defender Network and DNS Events to [indicator], [indicator] and Known IPs · unlock with TL Creds
  • Bazuka Dropper Install Behaviour - Stager Installed by TikTok Premium Package · unlock with TL Creds
  • Bazuka Self-Protection - Play Protect, ADB and Uninstall Shield Abuse on Android · unlock with TL Creds
  • Bazuka C2 - DNS Queries for [indicator] and [indicator] · unlock with TL Creds
  • Bazuka Dropper - Sideloaded TikTok Premium Package or Known Hash on Android Fleet · unlock with TL Creds
  • Bazuka Core - Play Protect Disabled, Wireless Debugging Enabled or ADB Pairing PIN Interception · unlock with TL Creds

Indicators of compromise

As of 2026-11-20, Threadlinqs Intelligence tracks 26 indicators of compromise for TL-2026-6604.

Sample indicators (3 of 26)

  • IP address: 95.181.213[.]21
  • Domain: brokerok[.]org
  • File hash: f046b50166c8229b49839eeebcc7fd14bf3f198e4ec1bfde8a5b271e1407dec5

Defanged, TLP:CLEAR. Full set of 26 indicators: unlock with TL Creds.

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

Community OSINT corroboration

2 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

Bazuka is a previously undescribed Android fraud framework documented by ThreatFabric on 2026-11-18 and covered by The Hacker News on 2026-11-20. ThreatFabric encountered it in late October 2026 while investigating an NFC card-relay application and initially judged it a simple wrapper before finding a full framework: a dropper or downloader, a core component and 23 modules. The panel is operated in an affiliate (traffer) model: traffers create landing pages (for example IPTV-themed pages) and lure victims with social-media ads or bogus sites. The sample dropper masquerades as a 'premium, adult-friendly version of TikTok' (package com.tiktok.premium; VirusTotal filename TikTok18-Dropper-tiktok18-premium-...apk). The panel's dropper builder can set minimal manifest permissions (without Accessibility), which ThreatFabric says shows intent to pass Google Play review; ThreatFabric has not observed this actually happening.

Infection chain: the dropper asks the victim to allow installation from unknown sources and installs a stager (package com.tiktok.premium.pro) via PackageManager sessions; a variant 'downloader' directly downloads the core. The stager/downloader fetches the core and campaign-specific modules. The core and modules are DEX files loaded in memory with InMemoryDexClassLoader (not written to disk) and are not obfuscated. The core orchestrates C2, loads modules, dispatches commands, collects device information, and manages permissions and updates. It runs a threat-level check (Build class, sensors, debugger/security-app presence, battery), grants itself permissions through Accessibility or ADB shell commands, disables the Google Play Protect component through the accessibility API, can enable Developer options and Wireless Debugging and intercept the pairing PIN to establish a TLS connection to the ADB daemon, and supports a self-destruct command (clears storage, deletes the database, disables the app). Commands are queued by priority, with enableUninstallProtection highest. C2 uses two WebSocket channels, Control and Data, with frames encrypted using ChaCha20-Poly1305 and rotating keys. Infrastructure includes Cloudflare-fronted brokerok.org subdomains plus raw IPs embedded in certificates bundled with the droppers.

The 23 modules cover: call_forwarding (USSD), clip_monitor (swaps BTC, ETH, TRC20 and SOL wallet addresses), contacts, device_control (ADB shell commands), file_manager, input_recorder (keylogger with regex for card numbers, wallet addresses and passwords), location, media (photos, camera, microphone), oem_power, optics (remote control, screencasting, phishing overlays), preflight, proxy (reverse TCP proxy), push, screen_control (lock-screen PIN, password and pattern capture), self_adb (embedded ADB client), session_ui (HTML phishing overlays), smsmessage (read/send SMS, extract verification codes, attempt deletion), system_checker (anti-analysis), tapread-nfc_module (contactless bank card interaction), toast, twofa (one-time codes), uninstall_shield (blocks uninstall/force-stop) and ussd. A novel overlay technique screenshots the home or lock screen and displays it on command to conceal background activity. ThreatFabric states the framework does not include automated payments.

The operator control panel integrates LLMs (locally hosted or API-accessed) for chat, victim 'score' (how attractive a victim is as a target), a permission-grant click-through strategy and AI agents that identify victims of interest. The panel UI mixes English and Russian; Russian-language slang, the domain brokerok.org and a Moscow-hosted server (95.181.213.21, ONE DASH VPS) are noted by ThreatFabric as indicators, but no actor or country attribution is made. Targeted countries, banks and apps are not stated in the sources. ThreatFabric notes the panel frontend was reachable without authentication and the first in-the-wild attempts suggest the product has found a buyer. Severity HIGH is an analyst judgment from the described capabilities, not a vendor rating.

MITRE ATT&CK techniques by tactic:

Defense Evasion

T1407 Download New Code at Runtime; T1629.001 Impair Defenses: Prevent Application Removal; T1629.003 Impair Defenses: Disable or Modify Tools; T1633.001 Virtualization/Sandbox Evasion: System Checks; T1655.001 Masquerading: Match Legitimate Name or Location

Credential Access

T1417.001 Input Capture: Keylogging; T1417.002 Input Capture: GUI Input Capture; T1453 Abuse Accessibility Features

Collection

T1429 Audio Capture; T1513 Screen Capture; T1517 Access Notifications; T1636.004 Protected User Data: SMS Messages

Command and Control

T1437.001 Application Layer Protocol: Web Protocols; T1521.001 Encrypted Channel: Symmetric Cryptography

Impact

T1582 SMS Control; T1641.001 Transmitted Data Manipulation

defense-evasion

T1604 Proxy Through Victim

Execution

T1623.001 Command and Scripting Interpreter: Unix Shell

Exfiltration

T1646 Exfiltration Over C2 Channel

Initial Access

T1660 Phishing

Remediation

Immediate actions

  • Block the Bazuka network IOCs (brokerok.org subdomains, gargona.cash, 95.181.213.21, 170.168.61.117) at DNS and egress
  • Hunt managed Android fleets for packages com.tiktok.premium and com.tiktok.premium.pro and the published SHA-256 hashes
  • Isolate and factory-reset devices where the packages are present, then rotate banking, crypto, email and 2FA credentials and review SMS-based authentication
  • Alert on apps that enable Wireless Debugging/Developer options or disable Play Protect without administrator action

Workarounds

  • Keep Google Play Protect enabled and disable Wireless Debugging and Developer options on devices that do not need them
  • Do not grant Accessibility, 'install unknown apps' or overlay permissions to apps not from a trusted source

Longer-term hardening

  • Block installation from unknown sources and restrict Accessibility service grants via MDM/mobile threat defense
  • Deploy mobile threat defense capable of detecting in-memory DEX loading and overlay/accessibility abuse
  • Prefer phishing-resistant authentication (passkeys, hardware tokens) over SMS one-time codes
  • Educate users that 'premium' or adult-themed versions of popular apps outside official stores are malicious

Affected and fixed versions

  • Google — Android (devices permitting sideloaded apps and Accessibility/Wireless Debugging grants)
    Vulnerable versions: Not version-specific; behavior-based malware, no specific Android versions stated in sources

Timeline

  • Earlier NFC-relay Android case (RatOn) reported by The Hacker News (month precision from article URL); ThreatFabric cites NFC relay via a RAT-downloaded app as precedent for the investigation that led to Bazuka.
  • The Hacker News reports the WindRelay NFC-relay Android banking trojan (month precision from article URL), the class of malware whose investigation later exposed Bazuka.
  • ThreatFabric encounters Bazuka in late October 2026 (exact day not stated; date approximate) while investigating an NFC card-relay application whose core and tapread-nfc_module were fetched automatically.
  • Zimperium zLabs publishes ArcticRAT research, a separate LLM-using Android RAT also using Wireless Debugging; covered by The Hacker News alongside Bazuka.
  • ThreatFabric publishes 'Bazuka: the fraud framework that can do it all', describing the dropper, core, 23 modules, ADB bridging, LLM-integrated control panel, and infrastructure IOCs.
  • The Hacker News (Ravie Lakshmanan) reports on Bazuka, highlighting the home/lock-screen screenshot overlay and Wireless Debugging pairing-PIN interception.

Sources cited

Drafted by the Threadlinqs research pipeline from 7 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