LSN โ€” Local Sovereign Network

๐Ÿ”ด CRITICAL โ€” Active Multi-Purpose Botnet

TI-2026-002 โ€” MikroTik Recon & Telegram Stealer Botnet

Coordinated IoT Botnet ยท 15 Nodes ยท 14 Countries ยท Residential Infrastructure

Botnet MikroTik Targeting Telegram Theft SIM Farm Detection Cryptominer Competition libssh2 15 Zombie Nodes

๐Ÿ“‹ Executive Summary

A precisely synchronized botnet of 15 compromised residential devices spanning 14 countries was caught executing an identical 9-step reconnaissance sequence on our honeypot between March 21 and April 9, 2026. Every node logged in with root:root, used the same libssh2 library, and ran the exact same commands โ€” a textbook indicator of centralized command-and-control (C2) orchestration.

The attack sequence reveals a multi-purpose intelligence collection operation that simultaneously:

  • ๐Ÿ” Identifies MikroTik routers โ€” using RouterOS-specific /ip cloud print
  • โ›๏ธ Checks for existing cryptominers โ€” competitive displacement before deployment
  • ๐Ÿ“ฑ Steals Telegram Desktop sessions โ€” hijacking tdata for account takeover
  • ๐Ÿ“ก Detects SMS/GSM gateway infrastructure โ€” identifying SIM farms for abuse
  • ๐Ÿท๏ธ Searches for a mystery marker โ€” D877F783D5D3EF8Cs (infection tag)

7 of 10 tested nodes are confirmed MALICIOUS by GreyNoise. The IPs belong to residential ISPs (Korea Telecom, Rostelecom, Swisscom, Viettel, Windstream) โ€” indicating these are compromised home routers/IoT devices, not dedicated infrastructure.

๐ŸŽฏ The 9-Step Attack Sequence

Every single botnet node executes this exact sequence in under 90 seconds:

STEP 1 โ”‚ /ip cloud print โ† MikroTik RouterOS detection
STEP 2 โ”‚ ifconfig โ† Network interface recon
STEP 3 โ”‚ uname -a โ† OS fingerprint
STEP 4 โ”‚ cat /proc/cpuinfo โ† CPU capability profiling
STEP 5 โ”‚ ps | grep '[Mm]iner' โ† Cryptominer competition check
STEP 6 โ”‚ ps -ef | grep '[Mm]iner' โ† Double-check (different flags)
STEP 7 โ”‚ ls -la ~/.local/share/TelegramDesktop/tdata ...
         โ† Telegram session + GSM/SIM farm detection
STEP 8 โ”‚ locate D877F783D5D3EF8Cs โ† Infection/control marker
STEP 9 โ”‚ echo Hi | cat -n โ† Connectivity verification

๐Ÿ”ฌ Step 7 โ€” Full Target List

The single most revealing command checks for 5 distinct asset types:

# Telegram Desktop session data (hijack accounts)
~/.local/share/TelegramDesktop/tdata
/home/*/.local/share/TelegramDesktop/tdata
 
# GSM hardware modems
/dev/ttyGSM*
/dev/ttyUSB-mod*
/dev/modem*
 
# SMS gateway software (gammu-smsd, etc.)
/var/spool/sms/*
/var/log/smsd.log
/etc/smsd.conf*
/var/config/sms/*
 
# Qualcomm QMI modem interface
/usr/bin/qmuxd
/var/qmux_connect_socket
 
# SIM card management
/etc/config/simman

This reveals the botnet operator is specifically looking for: (a) Telegram accounts to hijack, (b) physical SIM farm hardware for SMS fraud/2FA bypass, and (c) Qualcomm cellular modem interfaces. This is hallmark SIM-swap and account takeover infrastructure.

๐ŸŒ Global Botnet Map โ€” 15 Nodes, 14 Countries

NodeIP AddressCountryISP / ASNDate HitGreyNoise
๐ŸŸข 1178.160.209.67 ๐Ÿ‡ฆ๐Ÿ‡ฒ ArmeniaTelecom Armenia ยท AS12297 Mar 21โ€”
๐Ÿ”ด 285.30.212.24 ๐Ÿ‡ท๐Ÿ‡บ RussiaRostelecom ยท AS42610 Mar 23MALICIOUS
๐Ÿ”ด 3121.165.204.105 ๐Ÿ‡ฐ๐Ÿ‡ท South KoreaKorea Telecom ยท AS4766 Mar 24MALICIOUS
๐ŸŸข 436.89.252.58 ๐Ÿ‡ฎ๐Ÿ‡ฉ IndonesiaTelkom Indonesia ยท AS7713 Mar 25โ€”
๐Ÿ”ด 5222.108.39.109 ๐Ÿ‡ฐ๐Ÿ‡ท South KoreaKorea Telecom ยท AS4766 Mar 28โ€”
๐Ÿ”ด 65.187.97.40 ๐Ÿ‡ฌ๐Ÿ‡ต GuadeloupeCanal Telecom ยท AS21351 Mar 31MALICIOUS
๐Ÿ”ด 7144.2.91.96 ๐Ÿ‡จ๐Ÿ‡ญ SwitzerlandSwisscom ยท AS3303 Apr 1MALICIOUS
๐Ÿ”ด 8173.185.74.18 ๐Ÿ‡บ๐Ÿ‡ธ USAWindstream ยท AS7029 Apr 4MALICIOUS
๐Ÿ”ด 9183.98.76.106 ๐Ÿ‡ฐ๐Ÿ‡ท South KoreaKorea Telecom ยท AS4766 Apr 5MALICIOUS
๐ŸŸข 10218.145.181.48 ๐Ÿ‡ฐ๐Ÿ‡ท South KoreaKorea Telecom ยท AS4766 Apr 6โ€”
๐Ÿ”ด 11173.185.74.18 ๐Ÿ‡บ๐Ÿ‡ธ USAWindstream ยท AS7029 Apr 7MALICIOUS
๐ŸŸข 1281.18.37.142 ๐Ÿ‡ฐ๐Ÿ‡ฟ KazakhstanRadiobaylanys ยท AS29046 Apr 8โ€”
๐Ÿ”ด 13171.241.43.241 ๐Ÿ‡ป๐Ÿ‡ณ VietnamViettel ยท AS7552 Apr 9MALICIOUS
๐ŸŸข 1487.249.188.12 ๐Ÿ‡ธ๐Ÿ‡ช SwedenLjusnet ยท AS25417 Apr 10โ€”
๐ŸŸข 15202.44.227.204 ๐Ÿ‡น๐Ÿ‡ญ ThailandInternet Thailand ยท AS4618 Apr 11โ€”

๐Ÿงฌ Behavioral Fingerprint

๐Ÿ”‘ Credential
root:root
100% of nodes โ€” default IoT credential
๐Ÿ”ง SSH Library
libssh2 1.11.x
C library โ€” automated tool, not human
โฑ๏ธ Session Duration
~90 seconds
Connect โ†’ 9 commands โ†’ disconnect. Machine speed.
๐Ÿ“Š Session Pattern
1 per IP
Single recon pass. One IP per day โ€” rotating proxy.

๐Ÿ“… Activity Timeline (Mar 21 โ€” Apr 11)

One new zombie IP hits the honeypot every 1-3 days โ€” steady C2 tasking rhythm:

Mar 21 โ–ˆ 178.160.209.67 Armenia ยท Telecom Armenia
Mar 23 โ–ˆ 85.30.212.24 Russia ยท Rostelecom
Mar 24 โ–ˆ 121.165.204.105 South Korea ยท KT
Mar 25 โ–ˆ 36.89.252.58 Indonesia ยท Telkom
Mar 28 โ–ˆ 222.108.39.109 South Korea ยท KT
Mar 31 โ–ˆ 5.187.97.40 Guadeloupe ยท Canal Telecom
Apr 1 โ–ˆ 144.2.91.96 Switzerland ยท Swisscom
Apr 4 โ–ˆ 173.185.74.18 USA ยท Windstream
Apr 5 โ–ˆ 183.98.76.106 South Korea ยท KT
Apr 6 โ–ˆ 218.145.181.48 South Korea ยท KT
Apr 7 โ–ˆ 173.185.74.18 USA ยท Windstream (repeat)
Apr 8 โ–ˆ 81.18.37.142 Kazakhstan ยท Radiobaylanys
Apr 9 โ–ˆ 171.241.43.241 Vietnam ยท Viettel
Apr 10 โ–ˆ 87.249.188.12 Sweden ยท Ljusnet
Apr 11 โ–ˆ 202.44.227.204 Thailand ยท Internet Thailand
๐Ÿ”ด = GreyNoise MALICIOUS   ๐ŸŸ  = Not yet flagged

๐Ÿ” Mystery Marker: D877F783D5D3EF8Cs

Every node runs locate D877F783D5D3EF8Cs โ€” a search for a specific file with a hexadecimal-like name. This pattern is consistent with:

  • Infection marker โ€” a file dropped by prior malware to signal "already compromised"
  • C2 beacon file โ€” presence indicates active C2 communication channel
  • Competition check โ€” if file exists, another operator already controls this device

The marker was not found in VirusTotal, ThreatFox, or AlienVault OTX, suggesting it's either a custom botnet-specific artifact or recently deployed. Google shows some references โ€” likely from other honeypots catching the same botnet.

โš–๏ธ Criminal Activities Assessment

ActivityEvidenceSeverityImpact
๐Ÿ  Residential Device Compromise 15 IPs across residential ISPs, all with root:root โ— Critical Homeowners unknowingly host malware
๐Ÿ“ก MikroTik Router Hijacking /ip cloud print โ€” RouterOS-specific command โ— Critical DNS hijacking, traffic interception, VPN tunneling
๐Ÿ“ฑ Telegram Session Theft Checks tdata directory for all users โ— Critical Account takeover, impersonation, crypto scams
๐Ÿ“Ÿ SIM Farm Detection Checks for GSM modems, SMS daemons, QMI interfaces โ— High 2FA bypass, SMS fraud, SIM-swap attacks
โ›๏ธ Cryptominer Preparation ps | grep '[Mm]iner' โ€” checking for competitors โ— High Resource theft, electricity costs for victims
๐Ÿค– Botnet Infrastructure Identical commands across 15 IPs = centralized C2 โ— Critical Platform for DDoS, spam, further attacks

๐ŸŽฏ Why MikroTik Routers Are the Prize

๐Ÿ”“ DNS Hijacking
Change DNS servers to redirect all traffic through attacker-controlled resolvers. Victims see fake banking/crypto sites.
๐Ÿ•ณ๏ธ VPN Tunneling
Create L2TP/PPTP tunnels through the router. Use the victim's clean IP as exit node for attacks.
๐Ÿ“Š Traffic Sniffing
MikroTik's packet sniffer tool captures all unencrypted traffic. Credentials, emails, browsing history.
๐Ÿ›ก๏ธ Persistent Access
RouterOS survives reboots. Add SSH keys, scheduler scripts, and SOCKS proxy. Invisible to users.

๐Ÿ—บ๏ธ Geographic Distribution Analysis

๐Ÿ‡ฐ๐Ÿ‡ท South Korea ร— 4
All Korea Telecom (AS4766) โ€” largest cluster
๐Ÿ‡บ๐Ÿ‡ธ USA ร— 2
Windstream + Cablevision (residential)
๐Ÿ‡ท๐Ÿ‡บ Russia ร— 1
Rostelecom (state telco)
๐Ÿ‡ฆ๐Ÿ‡ฒ Armenia ร— 1
Telecom Armenia (ex-Soviet ISP)
๐Ÿ‡จ๐Ÿ‡ญ Switzerland ร— 1
Swisscom (tagged "scanner" by Shodan)
๐Ÿ‡ฌ๐Ÿ‡ต Guadeloupe ร— 1
French Caribbean island โ€” Canal Telecom
๐Ÿ‡ฐ๐Ÿ‡ฟ Kazakhstan ร— 1
Radiobaylanys โ€” Ural region ISP
๐Ÿ‡ป๐Ÿ‡ณ Vietnam ร— 1
Viettel (military-owned telco, ICS-tagged)
๐Ÿ‡ฎ๐Ÿ‡ฉ Indonesia ร— 1
Telkom Indonesia (state telco)
๐Ÿ‡ธ๐Ÿ‡ช Sweden ร— 1
Ljusnet โ€” rural xDSL provider
๐Ÿ‡น๐Ÿ‡ญ Thailand ร— 1
Internet Thailand โ€” retail ISP
Pattern: Overwhelmingly residential ISPs with no hosting infrastructure. 4 nodes on Korea Telecom (AS4766) suggests either a regional infection vector or KT's large install base of vulnerable IoT devices. The Viettel IP is tagged ICS (Industrial Control Systems) by Shodan โ€” potentially compromised critical infrastructure.

๐Ÿ›ก๏ธ Our Defense Response

๐Ÿฏ
Cowrie Honeypot
Captured full 9-step sequence. Returned fake data for /ip cloud print.
๐Ÿ›ก๏ธ
CrowdSec
All 15 IPs submitted to community blocklist. Auto-blocked across 49K+ decisions.
๐Ÿ”’
MikroTik CCR1009
SSH hardened, non-default credentials, API access restricted to K3S CIDR only.

๐Ÿง  Final Assessment

This is a professionally operated reconnaissance botnet with a sophisticated multi-layer target assessment strategy. The operator is not simply looking for machines to infect โ€” they are categorizing assets:

If MikroTik routerโ†’ Hijack for DNS poisoning, VPN tunnel, persistent proxy
If Telegram presentโ†’ Steal session for account takeover, impersonation, crypto scams
If SIM farm/GSM modemโ†’ Abuse for 2FA bypass, SMS fraud, SIM-swap attacks
If computing power availableโ†’ Deploy cryptominer (after removing competition)
If marker presentโ†’ Skip โ€” already controlled by this or allied operator

The 1-IP-per-day rotation suggests the C2 has a large pool of compromised devices and tasks each one against a batch of targets before moving to the next. Our honeypot is likely one of thousands of IPs being profiled. The South Korean cluster (4 IPs on Korea Telecom) may indicate a regional exploitation campaign targeting KT's residential router infrastructure.

๐Ÿ“Ž Investigation Metadata

Investigation IDTI-2026-002
Date Conducted2026-04-18
Data SourcesCowrie honeypot logs, GreyNoise Community API, Shodan InternetDB, IPWhois, WHOIS, ThreatFox, AlienVault OTX
Analysis ToolsLSN SOC Dashboard, 25 MCP Tools, Python correlation scripts, botnet fingerprinting
Monitoring Period2026-03-21 โ†’ 2026-04-11 (22 days)
Botnet Nodes15 confirmed (10 primary + 5 additional via command correlation)
GreyNoise Hits7/10 tested = MALICIOUS
Confidence LevelVERY HIGH โ€” identical command sequences across 15 IPs is definitive botnet evidence
โš  Personal capacity. Research published independently โ€” not reflecting employer views. Derived from passive observation of attacks against personal infrastructure. Full disclaimer โ†’
โ† Previous Honeypot Field Reports โ€” 2 / 20 Next โ†’