Threat reportMalwareTL-2026-2889
TIKTOUK Toolkit Harvests AWS, SMTP and Database Credentials from Exposed WordPress Backups and Config Files (CVE-2026-60137, CVE-2026-63030)
LevelBlue SpiderLabs analyzed TIKTOUK, a toolkit that probes WordPress sites for exposed backup/config files (wp-config.php.bak, .env, .git/config, backup.sql, debug.log) and recovers database, AWS, SMTP and API credentials. A leaked control panel held ~50,000 server-side credentials across ~37,000 domains, including hundreds of actor-validated live AWS keys.
Key facts
| Identifiers | TL-2026-2889, CVE-2026-60137, CVE-2026-63030 |
|---|---|
| Also known as | TIKTOUK, wp2shell (associated CVE chain) |
| Severity | HIGH |
| CVSS | 9.8 (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H) |
| Exploitation status | Linked CVE exploited in the wild; Listed in CISA KEV: CVE-2026-63030 (added ) |
| EPSS | CVE-2026-63030: 38.5% (98.4th percentile) |
| Affected products | WordPress WordPress core; WP Mail SMTP WP Mail SMTP plugin (stored credentials targeted); Easy WP SMTP Easy WP SMTP plugin (stored credentials targeted); FluentSMTP FluentSMTP plugin (stored credentials targeted) |
| Weaknesses (CWE) | CWE-89, CWE-436 |
| Malware | Go botnet (SHA-1 9903f4576980ff7cfd560ca57c665a4b59b3c30d), TIKTOUK |
| Tools | TIKTOUK HTTP hub |
| MITRE ATT&CK techniques | T1005 Data from Local System, T1059.006 Python, T1071.001 Web Protocols, T1078.004 Cloud Accounts, T1119 Automated Collection, T1140 Deobfuscate/Decode Files or Information, T1190 Exploit Public-Facing Application, T1552.001 Credentials In Files, T1589.001 Credentials, T1595.002 Vulnerability Scanning |
| Motivation | UNKNOWN |
| Target sectors | technology, ecommerce, news - media |
| Target regions | Global |
| Detection rules | 9 (1 shown free) |
| Sources | 10 cited sources |
Detection coverage
As of 2026-10-04, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-2889 across Splunk SPL, Microsoft KQL and Sigma, covering 23 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: Network Traffic To or From TIKTOUK Panel and Payload Hosts (31.56.58.59, 193.32.162.134, 195.178.110.209)
Detects network connections to or from the TIKTOUK payload host/controller (31.56.58.59) and control panels (193.32.162.134, 195.178.110.209). Outbound traffic from web servers indicates tool download (jscrawl-amd64, wp2s scripts, Go botnet) or stolen-credential submissions to the HTTP hub.
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1071.001, T1105, T1041
- Data sources: network_traffic_datamodel, firewall_logs, defender_for_endpoint_devicenetworkevents, commonsecuritylog
- Severity: high
- Confidence: high
Splunk SPL
| tstats summariesonly=true allow_old_summaries=true count AS connections sum(All_Traffic.bytes_out) AS bytes_out min(_time) AS first_seen max(_time) AS last_seen FROM datamodel=Network_Traffic.All_Traffic WHERE (All_Traffic.dest IN ("31.56.58.59","193.32.162.134","195.178.110.209") OR All_Traffic.src IN ("31.56.58.59","193.32.162.134","195.178.110.209")) BY All_Traffic.src All_Traffic.dest All_Traffic.dest_port All_Traffic.action
| rename "All_Traffic.*" AS *
| eval ioc_role=case(dest="31.56.58.59" OR src="31.56.58.59","payload_host_controller", true(),"tiktouk_control_panel")
| convert ctime(first_seen) ctime(last_seen)
| table first_seen last_seen src dest dest_port action ioc_role connections bytes_out
| sort - bytes_outMicrosoft KQL
let iocIps = dynamic(["31.56.58.59", "193.32.162.134", "195.178.110.209"]);
union isfuzzy=true
(DeviceNetworkEvents
| where Timestamp > ago(24h)
| where RemoteIP in (iocIps)
| project Timestamp, Device = DeviceName, SrcIP = LocalIP, DstIP = RemoteIP, DstPort = RemotePort, Process = InitiatingProcessFileName, CommandLine = InitiatingProcessCommandLine, Source = "MDE"),
(CommonSecurityLog
| where TimeGenerated > ago(24h)
| where DestinationIP in (iocIps) or SourceIP in (iocIps)
| project Timestamp = TimeGenerated, Device = DeviceName, SrcIP = SourceIP, DstIP = DestinationIP, DstPort = DestinationPort, Process = ApplicationProtocol, CommandLine = RequestURL, Source = "CEF")
| order by Timestamp descSigma
title: Network Connection To TIKTOUK Panel Or Payload Infrastructure
id: c27f5e81-b9a3-4d60-a8e4-7d1c2f9b3e06
status: experimental
description: Detects firewall-logged traffic to the IP addresses used by TIKTOUK as payload host/controller and control panels. Hits from web servers suggest tool download or stolen-credential submission to the TIKTOUK HTTP hub.
references:
- https://www.levelblue.com/blogs/spiderlabs-blog/tiktouk-tracing-a-wordpress-credential-collection-toolkit
author: Threadlinqs Team
date: 2026-10-04
tags:
- attack.command-and-control
- attack.t1071.001
- attack.t1105
- attack.exfiltration
- attack.t1041
logsource:
category: firewall
detection:
selection_dst:
dst_ip:
- '31.56.58.59'
- '193.32.162.134'
- '195.178.110.209'
selection_src:
src_ip:
- '31.56.58.59'
- '193.32.162.134'
- '195.178.110.209'
condition: 1 of selection_*
falsepositives:
- Threat-intel sandboxes or researchers deliberately contacting the published IOCs
- IP reassignment, since these addresses may be reused by unrelated hosts over time
level: highKnown false positives:
- Threat-intel sandboxes or analysts deliberately contacting the published IOCs
- Later reassignment of these IPs to unrelated services
Tuning notes: Prioritize hits where the source is a WordPress or web server. Check the initiating process (python3, jscrawl-amd64, curl, wget) to confirm toolkit activity. Inbound hits from these IPs against internet-facing WordPress hosts show scanning rather than compromise. The tstats form needs an accelerated Network_Traffic data model; swap to an index search if not accelerated. Re-check IOC freshness before relying on a block, since the IPs may be rotated.
- Burst Probing of WordPress Backup and Config Files ([indicator], [indicator], [indicator], [indicator], debug.log) · unlock with TL Creds
- TIKTOUK Toolkit Execution ([indicator], [indicator], [indicator]) and Known Binary Hashes · unlock with TL Creds
- Stolen-Credential Submission to TIKTOUK Hub (POST [indicator] and [indicator]) · unlock with TL Creds
- WordPress REST Batch Request With Nested UNION ALL SELECT (wp2shell / wp2s_crack Options-Table Query) · unlock with TL Creds
- AWS Key Abuse From TIKTOUK Infrastructure or SES/Bedrock Validation Calls (CloudTrail) · unlock with TL Creds
- TIKTOUK Sample Hashes and Tool Download From TIKTOUK Payload Host · unlock with TL Creds
- TIKTOUK WordPress REST Batch Fingerprinting (Malformed http:// Path + categories/0 + block-renderer) · unlock with TL Creds
- Successful Disclosure of WordPress Backup/Config File (HTTP 200/206 on [indicator], [indicator], [indicator], [indicator], debug.log) · unlock with TL Creds
Indicators of compromise
As of 2026-10-04, Threadlinqs Intelligence tracks 23 indicators of compromise for TL-2026-2889.
Sample indicators (3 of 23)
- IP address:
193.32.162[.]134 - File hash:
c6b8d0cdb53da98a5d15e79b7bb9e9f4c272c4f9592acdc291089f126f892f45 - IP address:
195.178.110[.]209
Defanged, TLP:CLEAR. Full set of 23 indicators: unlock with TL Creds.
Technical analysis
TIKTOUK is a three-part credential collection toolkit documented by LevelBlue SpiderLabs (analyst Maor Gabay) on 2026-10-01. The probing component (wp2s_poll.py) fetches target lists and pages to identify WordPress installations, then sends REST batch requests that combine a malformed 'http://' path with DELETE operations against /wp/v2/categories/0 and POST operations against /wp/v2/block-renderer/core/paragraph to fingerprint the site. When a JSON request is rejected (HTTP 403), it retries with multipart encoding, which can flip the response to HTTP 200. It also collects secret-pattern matches from page content alongside target classifications and submits them to the hub.
The credential collection component (wp2s_crack.py) retrieves wp-config.php.bak and parses database credentials and WordPress key/salt material. It sends nested REST batch requests containing author_exclude and UNION ALL SELECT expressions, decodes hexadecimal option values, and queries the options table for plugin settings. It also requests .env, .git/config, backup.sql and wp-content/debug.log. It recovers SMTP passwords from WP Mail SMTP (XSalsa20-Poly1305 secretbox), Easy WP SMTP (AES-256-CTR with a SHA-256-derived key) and FluentSMTP (AES-256-CTR keyed from LOGGED_IN_KEY, with the LOGGED_IN_SALT suffix removed), and derives Amazon SES SMTP passwords from AWS secrets. Results are submitted per target to /api/crack/report on the TIKTOUK panel.
A stripped Linux Go binary, jscrawl-amd64, fetches pages and their referenced JavaScript, scans for secrets (SendGrid, Anthropic, Bedrock tokens, AWS credential pairs) and submits findings to /v1/ingest. A separate Go botnet binary with remote command execution capability is associated with the same infrastructure. Payloads are served from 31.56.58.59, and TIKTOUK control panels were seen on 193.32.162.134 and 195.178.110.209. The whole system runs around a central HTTP hub that distributes tasks and aggregates stolen data.
A leaked panel exposed roughly 50,000 real server-side credentials across ~37,000 domains, including hundreds of actor-validated live AWS keys with SES, EC2 and Bedrock abuse potential. The report links the request structures to the WordPress core 'wp2shell' chain patched on 2026-07-17: CVE-2026-60137 (SQL injection via the author__not_in WP_Query parameter, CWE-89; NVD CVSS 5.9 primary / 9.1 secondary) and CVE-2026-63030 (REST API batch-route confusion at /wp-json/batch/v1, CWE-436, CVSS 9.8), which together enable pre-authentication SQL injection leading to RCE on WordPress 6.9.x before 6.9.5 and 7.0.x before 7.0.2. Both CVEs were added to the CISA KEV catalog on 2026-07-21 (due dates 2026-07-24 for CVE-2026-63030 and 2026-08-04 for CVE-2026-60137), and a public PoC appeared on GitHub on 2026-07-18; this KEV listing concerns exploitation of the wp2shell chain generally and is not tied to TIKTOUK. LevelBlue's analysis used synthetic target data and an analyst-controlled hub: the link between TIKTOUK and the CVEs rests on request structure only, and exploitation of the CVEs by TIKTOUK against production WordPress was not demonstrated. No threat actor attribution is made.
MITRE ATT&CK techniques by tactic:
Collection
T1005 Data from Local System; T1119 Automated Collection
Execution
Command and Control
Defense Evasion
T1078.004 Cloud Accounts; T1140 Deobfuscate/Decode Files or Information
Initial Access
T1190 Exploit Public-Facing Application
Credential Access
T1552.001 Credentials In Files
Reconnaissance
Remediation
Patches
- Update WordPress to 6.9.5, 7.0.2 or later (6.8.6 fixes CVE-2026-60137 only)
Immediate actions
- Remove wp-config.php.bak, .env, .git/, backup.sql and wp-content/debug.log from web-accessible paths
- Rotate every credential that may have been in exposed files: database, AWS access keys, SMTP/SES, SendGrid and other API keys, and WordPress keys/salts
- Review AWS CloudTrail for unexpected SES, EC2 and Bedrock activity from rotated keys
- Block 193.32.162.134, 195.178.110.209 and 31.56.58.59 at the perimeter
- Search web logs for REST batch requests containing 'http://:' paths, /wp/v2/categories/0 DELETE and block-renderer/core/paragraph POST sub-requests
Workarounds
- If patching is delayed, block both /wp-json/batch/v1 and rest_route=/batch/v1 at the WAF, and restrict unauthenticated REST access
Longer-term hardening
- Block dotfile, VCS and backup-extension requests at the web server or WAF
- Keep secrets out of web roots and out of client-side JavaScript; use a secrets manager
- Alert on REST batch requests with http:// paths and nested author_exclude/UNION expressions
- Alert on multipart retries following rejected JSON requests and on sensitive-file access followed by outbound result submissions
- Disable WP_DEBUG_LOG or write debug logs outside the web root
Affected and fixed versions
- WordPress — WordPress core
Vulnerable versions: 6.8.0-6.8.5 (CVE-2026-60137 only); 6.9.0-6.9.4; 7.0.0-7.0.1
Fixed in: 6.8.6; 6.9.5; 7.0.2; 7.1 beta2 - WP Mail SMTP — WP Mail SMTP plugin (stored credentials targeted)
- Easy WP SMTP — Easy WP SMTP plugin (stored credentials targeted)
- FluentSMTP — FluentSMTP plugin (stored credentials targeted)
Timeline
- WordPress.org enables forced auto-updates for affected versions due to the severity of the chain.
- NVD publishes CVE-2026-60137 (CWE-89) and CVE-2026-63030 (CWE-436, CVSS 9.8); the RCE chain was reported by Adam Kues of Assetnote/Searchlight Cyber.
- WordPress releases 7.0.2, 6.9.5 and 6.8.6 fixing the author__not_in SQL injection (CVE-2026-60137) and the REST batch route confusion (CVE-2026-63030); 6.8.6 fixes only the former.
- Full mechanism of the wp2shell chain is published and a working proof of concept appears on GitHub, making exploitation details public.
- Rapid7 authenticated vulnerability checks become available for the chain, and Picus publishes a technical breakdown.
- CISA adds CVE-2026-63030 (due 2026-07-24) and CVE-2026-60137 (due 2026-08-04) to the KEV catalog after widespread active exploitation of the wp2shell chain.
- LevelBlue SpiderLabs publishes its analysis of the TIKTOUK toolkit (locked-ioc-9, locked-ioc-8, jscrawl-amd64) and the ~50k-credential leaked control panel.
- Cyber Security News reports the findings, listing IOCs and detection guidance.
- Threadlinqs opens a tracking record for TIKTOUK; LevelBlue notes CVE linkage rests on request structure only and was lab-tested, not shown against production sites.
Sources cited
- NVD CVE-2026-60137 NVD (NIST) · Advisory · 2026-07-17
- NVD CVE-2026-63030 NVD (NIST) · Advisory · 2026-07-17
- CISA Adds Four Known Exploited Vulnerabilities to Catalog (2026-07-21) aviatrix.ai · Government advisory · 2026-07-21
- TIKTOUK: Tracing a WordPress Credential Collection Toolkit (LevelBlue SpiderLabs) levelblue.com · Research · 2026-10-01
- WordPress 7.0.2 Security Release wordpress.org · Vendor report · 2026-07-17
- GHSA-fpp7-x2x2-2mjf (CVE-2026-60137) GitHub · Code repository · 2026-07-17
- GHSA-ff9f-jf42-662q (CVE-2026-63030) GitHub · Code repository · 2026-07-17
- CVE-2026-63030 and CVE-2026-60137 (wp2shell): WordPress RCE Explained (Picus Security) picussecurity.com · Blog · 2026-07-20
- Exposed WordPress Backups Became a Gold Mine of AWS and Email Credentials cybersecuritynews.com · News · 2026-10-02
- New wp2shell WordPress Core Flaw Lets Attackers Take Over Sites (The Hacker News) The Hacker News · News · 2026-07-18
Other reports on CVE-2026-60137
- CTO at NCSC Weekly Summary (week ending July 26, 2026): Cl0p PTC Windchill/FlexPLM exploitation (CVE-2026-12569), SonicWall SMA 1000 zero-days (CVE-2026-15409/CVE-2026-15410), and multi-actor state-linked activity
- September 2026 CVE Landscape: 94 high-impact vulnerabilities, NetScaler zero-days CVE-2026-88771/CVE-2026-88772, AI-agent-assisted exploitation
- CyberXero: AI-Augmented Initial Access Broker Targeting Ukrainian Critical Infrastructure and Global WordPress/E-commerce Platforms
- Chinese-Speaking 'Kapibala' Actor (Red Heron-Linked) Chains WordPress wp2shell, Zyxel GS1900, and Ubiquiti UniFi OS Flaws to Steal Government Data
- AI-Agent-Driven Offensive Operation: Mass Cryptocurrency Wallet and Credential Compromise via Autonomous AI Coding Agents
- wp2shell: WordPress Core REST API Batch-Route Confusion Chained with author__not_in SQL Injection (CVE-2026-63030 / CVE-2026-60137) Yields Unauthenticated Pre-Auth RCE
Previous and next in malware
Drafted by the Threadlinqs research pipeline from 10 sources; published ; last updated —.