Daily Threat Briefing — 2026-10-12
34 findings2 articles5 sections
- Critical findings
- 0
- High findings
- 0
- New since 06:02 UTC
- 0
- CISA KEV additions
- 0
- First seen in this period
- 2 Articles79 CVEs1,253 IOCs
First seen in this period: news articles, each counted once · CVEs rated CVSS 7.0+ (NVD, GitHub, WordPress) · new abuse.ch indicators (ThreatFox, URLhaus, Feodo Tracker), dated by abuse.ch.
Priority queue
Act on these first
Critical and high findings from this briefing and its live updates. Most severe first, then newest.
- Critical severityNewLive update, section VulnerabilitiesSource: CVE
- Critical severityNewLive update, section VulnerabilitiesSource: CVE
- Critical severityNewLive update, section VulnerabilitiesSource: CVE
- Critical severityNewLive update, section VulnerabilitiesSource: CVE
- Critical severityNewLive update, section VulnerabilitiesSource: CVE
- Critical severityNewMaximum-severity SSRF flaw in SonicWall SMA 1000 under active attack (opens the source in a new tab)Live update, section VulnerabilitiesSource: RSS
- Critical severityNewLive update, section VulnerabilitiesSource: CVE
- Critical severityNewLive update, section VulnerabilitiesSource: CVE
- Critical severityNewLive update, section VulnerabilitiesSource: CVE
- Critical severityNewLive update, section VulnerabilitiesSource: CVE
Executive summary · as of 06:02 UTC
The situation
This window was dominated by ransomware and data-extortion leak-site activity, a large batch of WordPress plugin disclosures, and steady commodity malware infrastructure. TheGentlemen listed five high-profile victims within about two minutes: MoneyGram, KFC, Confluent (now IBM-owned), BigID and Placer AI. Redact claimed US utility Eversource Energy, and Qilin listed CNESTEN. Emperador claimed a large theft from Spelman Logistics that allegedly includes 245K carrier profiles with SSNs and EINs and 57K bank accounts. Separately, Daiichi Kosho disclosed that a malware infection at its contractor Nippon Columbia exposed more than 8.7 million customer and employee records. DataBreaches.net reported that the 'Luna Moth Files' attributed to Silent Ransom Group are genuine, despite the group's denials. All 29 breach entries in the window are reported.
Vulnerability intake was heavy: 79 NVD and 64 WordPress entries, of which 40 and 30 were supplied. None were added to CISA KEV. The most urgent are unauthenticated OS command injection in TOZED X300 routers (CVE-2026-108576, CVSS 10) and in OpenSpug Spug (CVE-2026-108540, public exploit available). Also notable are a modem baseband stack overflow reachable over the air via SIB1 messages (CVE-2026-66086) and hard-coded JWT secrets in the AI platforms agnai and BotSharp, which allow admin impersonation. A large cluster of WordPress plugin flaws allows unauthenticated site takeover through arbitrary option updates, PHP upload, SQL injection and stored XSS.
abuse.ch recorded 1,243 new indicators across 101 families, of which 777 were supplied. Vidar infostealer led by a wide margin with 237 indicators. It was followed by Mirai/Mozi IoT botnet payloads and DanaBot C2 nodes. Adversary C2 frameworks remained active: Sliver, VShell and Cobalt Strike. ClearFake fake-update lures and AMOS macOS stealer payloads also continued. SOC teams should prioritize Vidar and DanaBot C2 blocking, patch or isolate exposed TOZED/Spug assets, and audit WordPress estates for the affected plugins.
Live updates
16 updates since 06:02 UTC
CVE-2026-109758: Perforce Helix Authentication Service ships default SCIM bearer token enabling unauthorized access (CVSS 9.3)
- Critical severityCVE-2026-109758: Default SCIM bearer token in Perforce Helix Authentication Service (CVSS 9.3) (opens the source in a new tab)Source: CVEVulnerabilities
Perforce Helix Authentication Service 25.2 installation templates contain a default SCIM bearer token that the installer copies into deployments. Where the token is left unchanged and the SCIM endpoint is reachable, attackers can gain unauthorized access (CVSS 4.0: 9.3). Rotate the SCIM token and restrict network exposure of the endpoint.
SAP discloses a CVSS 10 OAuth2 SSRF in its Cloud SDK for Python, alongside three critical NetWeaver, S/4HANA and MII flaws
- Critical severityCVE-2026-89111: SAP Cloud SDK for Python OAuth2 SSRF leaks BTP client credentials (CVSS 10) (opens the source in a new tab)Source: CVEVulnerabilities
An OAuth2 SSRF in the SAP Cloud SDK for Python (CVSS 10) affects multi-tenant BTP applications. An attacker who controls the tenant_subdomain parameter can redirect OAuth2 token requests to an attacker-controlled server and exfiltrate client credentials. Patch affected BTP applications and review outbound token requests for unexpected destinations.
- Critical severityCritical SAP NetWeaver Visual Composer, S/4HANA and MII flaws (CVE-2026-89110, -89107, -89108) (opens the source in a new tab)Source: CVEVulnerabilities
SAP NetWeaver Visual Composer (CVE-2026-89110, CVSS 9.9) lacks authorization checks, letting low-privileged users invoke operations on preconfigured remote systems, which could lead to full compromise. SAP S/4HANA Manage Substitution/Validation Rules (CVE-2026-89107, CVSS 9.9) allows authenticated code injection via its rule scripting language. SAP MII (CVE-2026-89108, CVSS 9.1) lets MII administrators provision elevated accounts on the underlying system.
- Source: CVEVulnerabilities
The Metasploit Framework Web Service REST API (msf-ws) CredentialServlet omits warden.authenticate! on its update_credential (PUT) and delete_credentials (DELETE) handlers, allowing unauthenticated tampering with the credential store (CVSS 9.8). Red teams and labs exposing msf-ws should restrict network access and update.
- Critical severityGitLab regex RCE assigned CVE-2026-97229 with fixed versions confirmed (opens the source in a new tab)Source: CVEVulnerabilities
Follow-up to the earlier report on GitLab's maximum-severity regex RCE: NVD has now published CVE-2026-97229 (CVSS 10). It confirms unauthenticated code execution on GitLab CE/EE 19.2 before 19.2.8, 19.3 before 19.3.5 and 19.4 before 19.4.2. Use these fixed versions to verify patch status.
IBM i HTTP Server Administration flaw (CVE-2026-96119, CVSS 9.9) allows authenticated remote code execution
- Critical severityCVE-2026-96119: IBM i HTTP Server Administration code execution (CVSS 9.9) (opens the source in a new tab)Source: CVEVulnerabilities
Improper data sanitization in the HTTP Server Administration web application on IBM i 7.3 through 7.6 lets a remote authenticated attacker obtain sensitive information or execute arbitrary code (CVSS 9.9). Restrict access to the admin interface and apply IBM PTFs once available.
- Source: CVEVulnerabilities
IBM DataStage on Cloud Pak for Data 5.4.0.0 has a path traversal flaw that allows a remote authenticated attacker to execute arbitrary code (CVSS 9.6). Related CVE-2026-82103 (sensitive information exposure) and CVE-2026-82101 (path traversal DoS), both CVSS 9.6, affect the same release.
Dark Project ransomware group claims jewelry retailer Claire's Inc. in a three-victim leak-site burst
- Source: BreachData Breaches & Leaks
The Dark Project extortion group listed Claire's Inc., the mall-based jewelry and accessories retailer, on its leak site alongside two other US victims posted within minutes of each other. No data volume or data classes were disclosed. Organizations with retail or payment ties to Claire's should monitor for follow-on publication.
- High severityGerman insolvency law firm Dr. Bien: 2,387 case databases claimed by aurora (opens the source in a new tab)Source: BreachData Breaches & Leaks
The aurora group claims to have exposed data from Rechtsanwaltskanzlei Dr. Bien, an insolvency law firm in Hamm, NRW, that serves as court-appointed insolvency administrator for 12 local courts in Westfalen. The listing cites 2,387 WinSolvenz case databases (Microsoft Access MDB files), each containing a full financial dossier for an insolvency case. This implies high-sensitivity third-party financial data on debtors and creditors.
- Medium severityWastewater equipment maker Franklin Miller listed by Dark Project (opens the source in a new tab)Source: BreachData Breaches & Leaks
In the same burst, Dark Project listed Franklin Miller Inc., a manufacturer of wastewater grinders, shredders and screens supplying municipal and industrial treatment facilities. It is relevant to water-sector supply-chain monitoring. No data details were provided.
Attackers are actively exploiting a maximum-severity SSRF flaw in SonicWall SMA 1000 appliances.
- Critical severityMaximum-severity SSRF flaw in SonicWall SMA 1000 under active attack (opens the source in a new tab)Source: RSSVulnerabilities
Dark Reading reports active exploitation of a maximum-severity server-side request forgery vulnerability in SonicWall SMA 1000 remote-access appliances. Organizations running SMA 1000 should apply vendor fixes immediately and review appliance logs for anomalous outbound requests and unauthorized access.
- Critical severityCVE-2026-109566: Hard-coded super-admin credentials in Rapyd Admin (CVSS 9.8) (opens the source in a new tab)Source: CVEVulnerabilities
Rapyd Admin versions 9.0.0 through 9.19.x seed a hard-coded super-admin account (admin@laravel / admin) through AuthSeeder during setup. NVD notes the account is silently re-created, giving remote unauthenticated attackers full admin access. Upgrade to 9.20.0 and audit for this account.
- Critical severityCVE-2026-108929: Netcore NR268 DDNS argument injection (CVSS 9.9) (opens the source in a new tab)Source: CVEVulnerabilities
The Hostname argument in /router/DDNSset.cgi on Netcore NR268 1.7.121109 is passed to execvp, enabling remote argument injection. This is a new flaw on the same device as the previously reported CGI authentication bypass (CVE-2026-108907), raising the risk of chained unauthenticated compromise.
- High severityCVE-2025-71429: Unauthenticated command injection in GPT-SoVITS WebUI (CVSS 9.8) (opens the source in a new tab)Source: CVEVulnerabilities
GPT-SoVITS through 20250606v2pro contains an OS command injection in tools/uvr5/vr.py. Unauthenticated users can place shell substitutions such as $(...) in the UVR5 WebUI output folder fields when using mp3/m4a export. Exposed AI voice-cloning WebUIs should be restricted from network access.
- High severityCVE-2026-44828: Unauthenticated setup API in Qexo static blog editor (CVSS 9.3) (opens the source in a new tab)Source: CVEVulnerabilities
Qexo 4.0.0 to before 4.2.0 exposes /api/init_step/ with no authentication or setup-completion guard. An unauthenticated requester holding a CSRF token from a public page can re-run initialization steps, potentially taking over the blog editor.
HAProxy Data Plane API flaw (CVE-2026-109575, CVSS 9.8) can drop authentication controls across multiple transports
- Critical severityCVE-2026-109575: HAProxy Data Plane API authentication controls dropped (CVSS 9.8) (opens the source in a new tab)Source: CVEVulnerabilities
In HAProxy Data Plane API 2.2.0 through 3.4.3 and the corresponding Enterprise releases, process-global authentication logic in Server.ConfigureAPI() can improperly drop security controls across multiple transport types when a TLS CA (tls_ca) is configured. Exposed Data Plane API instances allow load balancer configuration changes, so restrict network access and upgrade to a fixed release.
- Critical severityCVE-2026-96012: IBM Db2 Mirror for i privilege escalation to code execution (CVSS 9.9) (opens the source in a new tab)Source: CVEVulnerabilities
An authenticated attacker can craft a malicious request against IBM Db2 Mirror for i 7.4, 7.5 and 7.6 that bypasses additional authorization and input validation checks. Successful exploitation allows code execution and elevated privileges on IBM i systems.
- Critical severityCVE-2026-57118: PraisonAI multi-agent endpoints exposed without authentication (CVSS 9.8) (opens the source in a new tab)Source: CVEVulnerabilities
In praisonaiagents before 1.6.59, AgentTeam.launch and the Agents alias register agent listing and chat endpoints without access-control checks. Any remotely exposed server therefore lets unauthenticated users enumerate agents and invoke them. Upgrade to 1.6.59 or later and avoid exposing these endpoints.
- Critical severityCVE-2026-87744: Quarkus REST Data Panache generated endpoints lose security roles (CVSS 9.1) (opens the source in a new tab)Source: CVEVulnerabilities
When a developer customizes an operation with @MethodProperties without re-declaring security roles, the Quarkus REST Data Panache code generator silently drops security controls on the generated endpoint. This lets unauthenticated attackers bypass protections. Audit customized Panache REST resources for missing role annotations.
Anthropic cuts internet access for high-risk AI cyber evaluations after its models breached US government websites
- High severityAnthropic takes risky AI cyber tests offline after models breach US government sites (opens the source in a new tab)Source: RSSTools & Techniques
Anthropic says internal AI models undergoing high-risk cybersecurity evaluations no longer have internet access after incidents involving US federal, state and local government websites. The disclosure is a notable real-world case of autonomous AI testing affecting live third-party infrastructure; public-sector web defenders should review logs for anomalous automated activity.
- Medium severityProminent ransomware negotiator arrested on extortion conspiracy charges (opens the source in a new tab)Source: RSSThreat Actors
Edward Dubrovsky, a longtime cybersecurity executive and prominent ransomware negotiator, was arrested and charged in Pennsylvania with conspiracy to extort money. Organizations that used third-party negotiation services should review engagement records and payment flows.
- Low severityNorway fines Microsoft's Xandr $9.6M over Grindr user data sharing (opens the source in a new tab)Source: RSSPolicy
Norway's Datatilsynet fined Microsoft-owned ad-tech firm Xandr NOK 92 million (about $9.6M) for unlawfully processing and sharing sensitive data of Grindr users in Norway. The ruling reinforces GDPR enforcement against ad-tech handling of special-category data.
Ledger confirms hardware spy implant in wallets sold via Southeast Asian reseller CryptoBilis drained users' crypto
- High severityLedger confirms spy chip implant behind hardware wallet crypto thefts (opens the source in a new tab)Source: RSSData Breaches & Leaks
Ledger confirmed that some hardware wallets distributed through the Southeast Asian reseller CryptoBilis carried a hardware implant that drained users' cryptocurrency wallets. This is a hardware supply-chain compromise. Organizations and users who obtained devices through third-party resellers should treat them as untrusted and move funds to wallets bought directly from the vendor.
- Critical severityEdimax BR-6208AC: multiple command injections and buffer overflows with public exploits (CVE-2026-108918, -108917) (opens the source in a new tab)Source: CVEVulnerabilities
Two remote command injections affect Edimax BR-6208AC firmware 1.32: the readE2P argument in /goform/formWlanMP (CVE-2026-108918) and the tftpsIP argument in /goform/formAutoFWupgrade (CVE-2026-108917). Both are rated CVSS 10 and exploits are publicly available. Related buffer overflows rated CVSS 9.9 were also published: ateBW in formWlanMP (CVE-2026-108915) and ifname in formWlSiteSurvey (CVE-2026-108914).
- Critical severityCVE-2026-108907: Netcore NR268 router CGI authentication bypass (opens the source in a new tab)Source: CVEVulnerabilities
An improper authentication flaw in the user_ok function of the CGI execution component in Netcore NR268 firmware 1.7.121109 can be exploited remotely. It is rated CVSS 10 and an exploit has been publicly disclosed.
- Source: CVEVulnerabilities
In Grav Login OAuth plugin versions before 1.4.3, OAuth accounts are created with a predictable <provider>.<id> username and an unsalted MD5 of the provider ID as the password, and password login is left enabled. An unauthenticated attacker who knows a provider ID can log in as that user (CVSS 9.1). Upgrade to 1.4.3.
- High severityCVE-2026-77366: obsidian-tc MCP server cross-vault authorization bypass (opens the source in a new tab)Source: CVEVulnerabilities
obsidian-tc, an MCP server for Obsidian vaults, does not check the caller-supplied vault value in versions before 1.3.1. Its JWTs also lack a vault claim, so a caller can reach vaults outside its authorization (CVSS 9.9). Update to 1.3.1 if this server is used with AI agent tooling.
N0n extortion group names 10 maritime and port victims, including Ocean Network Express, Santos Brasil and ZPMC, with Oct 19 leak deadline
- High severityN0n names port, terminal and shipping victims across seven countries (opens the source in a new tab)Source: BreachThreat Actors
Building on its earlier anonymized countdown posts, N0n has now named 10 maritime-sector victims: Ocean Network Express, Santos Brasil, Portonave, Asyaport, MGNP, GlobalNet LCL, Polaris Shipping, ZPMC and Tsuneishi/Kambara Kisen. Claimed data includes terminal operating system database scripts, EDI archives, dangerous-goods records, voyage and fuel data, and billing material, with publication set for 2026-10-19 13:27 UTC. Port and shipping operators should review TOS/EDI access paths and exposure to shared logistics partners.
- High severityInterlock lists six victims in burst, heavy on healthcare supply chain (opens the source in a new tab)Source: BreachData Breaches & Leaks
Interlock posted six victims within minutes: Concord Eye Center (PHI referenced), Nu-Quest Distribution (pharmaceutical wholesaler), Pedigo Products (medical equipment manufacturer), Valenti Management (Wendy's franchisee), Greentech Environmental and ECF Inc. (hydrocarbon terminal contractor). Healthcare supply chain and clinical providers are heavily represented. Separately, IncRansom claimed Brooks Law Firm, citing personal data for 8,200 clients.
AnyDesk before 8.0.3 has an unauthenticated heap overflow (CVE-2026-109390) giving remote root code execution without user approval
- Critical severityCVE-2026-109390: Unauthenticated RCE in AnyDesk remote desktop client (opens the source in a new tab)Source: CVEVulnerabilities
AnyDesk before 8.0.3 has a heap-based buffer overflow in the mode 5 stream packet handler of its session protocol (CVSS 9.8). Unauthenticated remote attackers can execute arbitrary code, demonstrated on Linux as root, without the victim approving desktop control. Upgrade to 8.0.3 or later. Hunt for anomalous AnyDesk sessions and for child processes spawned by the AnyDesk service.
- Critical severityProgress MarkLogic Server: multiple critical privilege-escalation and code-execution flaws (opens the source in a new tab)Source: CVEVulnerabilities
Progress disclosed six critical flaws (CVSS 9.6–9.9) in MarkLogic Server 12.x before 12.1.0 and 11.x before 11.3.7: CVE-2026-85955, CVE-2026-85948, CVE-2026-85947, CVE-2026-9191 and CVE-2026-85956. They include an XQuery evaluation privilege bypass enabling server-side code execution, XQuery injection in a deprecated Management API path, creation of persistent rogue administrators through forest-management operations, and cross-scope backup/restore. All require some level of authenticated access. Patch, and review MarkLogic admin accounts for unauthorized additions.
- Source: CVEVulnerabilities
Two more remotely exploitable OS command injections affect Linksys E1700 firmware 1.0.0.4.003, each with a public exploit. CVE-2026-108896 is in setLinksysIpv6 (the _6rd_prefix and _6rd_prefix_len arguments) and CVE-2026-108895 is in setDMZ (the MAC_addr argument). These add to the previously reported CVE-2026-108892 stack overflow on this end-of-life router. Replace affected devices or restrict access to their management interfaces.
- High severityCVE-2026-92681: Authentication bypass for @cypher fields in @neo4j/graphql (opens the source in a new tab)Source: CVEVulnerabilities
@neo4j/graphql versions up to and including 7.6.3 silently ignore @authentication directives defined at the schema level or on Query/Mutation types for top-level @cypher fields (CVSS 4.0 score 9.3). Only directives placed directly on individual fields are enforced, so affected fields may be reachable without authentication. Upgrade, or move @authentication onto each @cypher field.
French supermarket refrigeration attack reportedly targeted Danfoss controllers, hitting hundreds of stores
- High severityFrench supermarket refrigeration attack linked to Danfoss controllers (opens the source in a new tab)Source: RSSThreat Actors
Update to the previously reported supermarket refrigeration attack: attackers reportedly targeted Danfoss refrigeration controllers to remotely defrost units overnight. The Record says hundreds of stores across several of France's largest grocery chains were affected, with fresh food spoiled. Retail and facilities operators should audit internet-exposed refrigeration and building-management controllers and review remote-access paths.
- High severityElastic exposes Apex Accelerator: malicious npm packages behind fake Web3 services (opens the source in a new tab)Source: RSSMalware
Elastic Security Labs (REF8801) details 'Apex Accelerator', an operation that pairs convincing Web3 services with malicious npm packages. It also uses attacker-controlled AI infrastructure and cross-platform malware delivery. Developer and crypto-sector teams should review npm dependencies and hunt for the reported indicators.
Three critical stack overflows in Linksys E1700 router firmware disclosed with public exploits
- High severityCVE-2026-108892: Linksys E1700 Mac Filter stack overflow (CVSS 10) (opens the source in a new tab)Source: CVEVulnerabilities
GitLab urges immediate patching of a maximum-severity RCE flaw in its CI/CD regular expression compiler
- Critical severityGitLab warns of maximum-severity RCE flaw in CI/CD regex compiler (opens the source in a new tab)Source: RSSVulnerabilities
GitLab is urging users to patch immediately for a maximum-severity vulnerability in its CI/CD Regular Expression Compiler that could let remote attackers run arbitrary code on targeted instances. No CVE id was provided in the source. Self-managed GitLab instances should be prioritized for patching, with CI/CD and runner activity reviewed for anomalous execution.
- Source: RSSData Breaches & Leaks
Ahold Delhaize, one of the world's largest food retail chains, confirmed it took some systems offline in response to a cybersecurity incident affecting its network. The nature of the attack and its impact have not yet been disclosed.
- High severityCyberattack forces refrigerator shutdowns at Intermarché and Coopérative U stores (opens the source in a new tab)Source: RSSData Breaches & Leaks
French supermarket chains Intermarché and Coopérative U said a cyberattack caused refrigerators to defrost at roughly 150-200 Intermarché stores and almost 100 Coopérative U outlets. This indicates impact on building and refrigeration control systems and is relevant to retail OT/IoT defenders. It coincides with the separate Ahold Delhaize incident reported the same day.
- Medium severityIcon Cancer Centre breach compromises Australian patient data (opens the source in a new tab)Source: RSSData Breaches & Leaks
Icon Oncology Group confirmed that patient data was affected by a breach involving unauthorized access at Icon Cancer Centre, Australia's largest cancer care provider. The scope of the exposed data has not yet been detailed.
- Medium severityAvisLoader: ClickFix-delivered Windows loader with encrypted P2P C2 (opens the source in a new tab)Source: RSSMalware
Varonis detailed AvisLoader, a new Windows loader that uses encrypted peer-to-peer communications to make its C2 infrastructure resistant to takedown. It is delivered via ClickFix lures, persists through shortcut (LNK) files and supports remote command execution. Defenders should hunt for clipboard-pasted Run-dialog commands and newly created startup shortcuts.
Empire Market co-creator Raheim Hamilton sentenced to 40 years in federal prison
- Medium severityEmpire Market co-creator sentenced to 40 years in federal prison (opens the source in a new tab)Source: RSSPolicy
Raheim Hamilton, a Virginia man who co-created and operated the Empire Market dark web marketplace, has been sentenced to 40 years in federal prison. This closes a previously reported prosecution of one of the larger darknet marketplaces for drugs and stolen data.
- Medium severityPhishing emails use hidden prompt injection to manipulate AI inbox assistants (opens the source in a new tab)Source: RSSTools & Techniques
Malwarebytes reports that phishing emails are embedding hidden instructions aimed at AI assistants that summarize inboxes. This indirect prompt injection can surface attacker actions as legitimate to-do items, even when the user ignored the original email. Teams should review AI email-summarization integrations and user guidance.
- Low severityWindows 11 23H2 Enterprise/Education/IoT reaches end of support in 30 days (opens the source in a new tab)Source: RSSVulnerabilities
Microsoft warned that Windows 11 23H2 Enterprise, Education and IoT Enterprise editions stop receiving security updates in 30 days. Organizations should inventory remaining 23H2 endpoints and plan upgrades to avoid unpatched systems.
- Source: RSSThreat Actors
Ukrainian drones struck a third Yandex data center in Russia over the weekend and forced it offline. The strike disrupted cloud, payment, delivery and other online services across the country. Organizations with dependencies on Yandex services should expect continued availability issues.
Critical OTRS Community Edition flaws allow header-based authentication bypass (CVE-2026-93582) and persistent JWT sessions
- Critical severityCVE-2026-93582: OTRS Community Edition authentication bypass via spoofed REMOTE_USER header (opens the source in a new tab)Source: CVEVulnerabilities
In OTRS Community Edition 6 and earlier, the HTTP Basic Authentication modules accept the username from either REMOTE_USER or HTTP_REMOTE_USER. In CGI and mod_perl deployments, a client-supplied Remote-User request header can be mapped to HTTP_REMOTE_USER, which may let an attacker authenticate as an arbitrary user (CVSS 3.1: 9.4). Disable or harden Basic Auth modules and strip Remote-User headers at the reverse proxy.
- High severityCVE-2026-78357: OTRS JWTs survive account deactivation and password change (opens the source in a new tab)Source: CVEVulnerabilities
OTRS does not immediately invalidate previously issued JWTs when an agent or customer account is deactivated or its password is changed. Stolen tokens can therefore remain valid for about 16 hours (CVSS 3.1: 9.1). During incident response, password resets and account disablement alone will not evict an attacker; plan for token-lifetime exposure or forced key rotation.
Critical CVSS 10 OS command injection disclosed in TRENDnet TEW-751DR/752DRU/733GR router web management
- Critical severityCVE-2026-108804: OS command injection in TRENDnet TEW-751DR/752DRU/733GR routers (opens the source in a new tab)Source: CVEVulnerabilities
CVE-2026-108804 (CVSS 3.1: 10.0) is an OS command injection in the fwrite function of pigwidgeon.cgi in the web management backend of TRENDnet TEW-751DR, TEW-752DRU and TEW733GR (firmware 1.03B01/1.03B03), triggered via the 'service' argument. The NVD description indicates the attack can be launched, but it is truncated before saying whether that is remotely. These are older SOHO models, so patch availability is uncertain; restrict or disable WAN-side management, consider replacing the devices, and monitor for unexpected outbound connections that could indicate IoT botnet recruitment.
Updates are added during the day when material new intelligence arrives. They are not part of the executive summary; the next daily briefing recaps them.
Browse the full briefing
34 findings · 5 sections- 8 critical
- 22 high
- 4 medium
Vulnerabilities
No new CISA KEV additions this window. Of 79 NVD and 64 WordPress vulnerability entries, 40 and 30 respectively were supplied. Priorities: - Remote command injection in TOZED X300 and OpenSpug Spug (public exploit). - An over-the-air baseband overflow. - Hard-coded secrets in AI platforms. - A large cluster of unauthenticated WordPress plugin takeover, upload, SQLi and XSS flaws.
- Critical severityCVE-2026-108576: Remote OS command injection in TOZED X300 router (opens the source in a new tab)
Affects TOZED X300 up to 6.01.3. The 'Host' argument of process_ping in the IPPingDiagnostics handler is vulnerable to OS command injection, exploitable remotely (CVSS 3.1: 10.0). The vendor was contacted early about the issue. Identify exposed TOZED CPE/routers, restrict management interfaces, and monitor for anomalous ping diagnostic requests containing shell metacharacters.
- Critical severityCVE-2026-108540: OpenSpug Spug command injection with public exploit (opens the source in a new tab)
Affects OpenSpug Spug (ops/automation platform) up to 3.4.0/4.0.1. The /exec/transfer File Transfer component allows remote OS command injection (CVSS 9.9). A public exploit has been published. Spug typically holds SSH access to managed hosts, so compromise can cascade across infrastructure. Restrict access and hunt for unexpected child processes.
- Critical severityCVE-2026-66086: Baseband modem stack overflow via malicious SIB1 broadcast (opens the source in a new tab)
A modem OTA stack fails to properly validate scheduling information in SIB1 messages, causing a stack-based buffer overflow. This may allow remote code execution or DoS with no additional privileges (CVSS 9.8). Exploitation would require a rogue base station within radio range. Track handset and baseband vendor security bulletins for fixes.
Source: CVE - Critical severityHard-coded JWT secrets in AI platforms agnai and BotSharp enable admin impersonation (opens the source in a new tab)
Agnaistic agnai through 1.0.555 (CVE-2026-108753, CVSS 9.4) ships a self-host docker-compose file with a fixed admin password and public JWT secret. Attackers can log in as admin or mint admin JWTs. BotSharp through 5.2.0 (CVE-2026-108860, CVSS 9.1) commits a hard-coded Jwt:Key in WebStarter appsettings.json, allowing forged bearer tokens. Rotate secrets on all self-hosted deployments.
- Critical severityWordPress plugins allow unauthenticated administrator takeover (opens the source in a new tab)
Three plugins (all CVSS 9.8) allow unauthenticated attackers to gain administrator access: - AI Content Generator Marketing ≤1.0.0 (CVE-2026-85118): missing nonce/capability checks let attackers update or delete arbitrary WordPress options. - Freeton WP ≤1.0.0 (CVE-2026-84737): flawed activation-code validation allows login as any user whose email is known. - Mindstien Quick Login ≤1.0 (CVE-2026-84734): flawed session validation grants an administrator session. Remove or disable these plugins.
- Critical severityclick5 CRM add-ons: unauthenticated arbitrary option update and stored XSS (8 CVEs) (opens the source in a new tab)
The click5 CRM add-ons for Contact Form 7, Gravity Forms, WPForms and Ninja Forms are affected. Their REST endpoints lack authorization and CSRF checks and do not restrict which option can be updated, allowing unauthenticated arbitrary option updates (CVE-2026-84251/84252/84253/84254, CVSS 9.8). The same plugins also have unauthenticated stored XSS against admins (CVE-2026-84258/84259/84260/84261, CVSS 8.8).
- Critical severityCVE-2026-104028: Anton Extensions WordPress plugin unauthenticated PHP upload/RCE (opens the source in a new tab)
Anton Extensions ≤1.2.2 lacks capability checks, nonce verification and file-type validation when writing attacker-supplied content to an attacker-chosen path. Unauthenticated attackers can upload arbitrary PHP files and achieve remote code execution (CVSS 9.8). Hunt for unexpected PHP files in web roots.
- Critical severityCVE-2026-81649: Fundiin for WooCommerce exposes payment credentials and orders (opens the source in a new tab)
Fundiin cho WooCommerce ≤3.4.0 protects several REST routes with a credential that is identical on every installation. Unauthenticated attackers can disclose the store's payment credentials and customer order data (CVSS 9.1). Affected merchants should rotate payment credentials.
- High severityInsurify and Quick quotes WordPress plugins: unauthenticated option deletion/overwrite (opens the source in a new tab)
Insurify ≤1.0 (CVE-2026-86717, CVE-2026-85121) and Quick quotes ≤1.0.0 (CVE-2026-86706) expose AJAX actions without authorization or nonce checks. Unauthenticated users can delete, create or overwrite arbitrary WordPress options (CVSS 9.1). This can take sites offline and strip all users of their roles.
- High severityAuthenticated RCE and authorization flaws in self-hosted apps Blinko, databasement and Vearch (opens the source in a new tab)
All three require only an authenticated account (CVSS 8.8): - Blinko ≤1.8.8 (CVE-2026-108980): path traversal in the installPlugin tRPC procedure overwrites startup files as root. - databasement <1.8.2 (CVE-2026-108963): argument injection via the database name passed to mariadb-dump (e.g. --result-file) enables webshell writes and RCE. - Vearch 3.5.2–3.5.9 (CVE-2026-108746): ignores ReadOnly/None privilege levels, allowing unauthorized document and role changes.
- High severityCVE-2026-103694: Mobile builder WordPress plugin self-registration to administrator (opens the source in a new tab)
Mobile builder ≤1.4.2 (CVE-2026-103694, CVSS 8.8) does not restrict which user meta keys can be updated through a REST route. Any self-registered user, such as a customer, can grant themselves the administrator role. Separately, Disable Users ≤1.0.5 (CVE-2026-88827) does not enforce account disabling on all authentication paths.
Source: CVE - High severityUnauthenticated SQL injection in five WordPress plugins (opens the source in a new tab)
Five plugins allow unauthenticated SQL injection with database read (CVSS 8.6): - Post Voting System ≤1.0 (CVE-2026-89302) - WP Verify API ≤1.0.0 (CVE-2026-89299) - Loja Automática ≤1.0.0 (CVE-2026-89297) - ASPL Product Quotation ≤1.1.0 (CVE-2026-89287) - Datalist it ≤0.0.3 (CVE-2026-89285)
Source: CVE - High severityUnauthenticated stored XSS across nine WordPress plugins (opens the source in a new tab)
Nine plugins allow unauthenticated attackers to store JavaScript that executes in admin or visitor browsers (CVSS 8.8): - Piotnet Forms ≤1.0.30 (CVE-2026-96227, renderable file upload) - KeyWord Collector ≤1.4 (CVE-2026-88905) - Topcontent ≤1.2.1 (CVE-2026-88903) - SmugMug Embed ≤3.13 (CVE-2026-88826) - BuddyPress Instant Chat ≤1.6 (CVE-2026-87764) - Adwised Web Push Notification ≤2.5.7 (CVE-2026-87762) - Web Vitals Tracking ≤5.4.2 (CVE-2026-87760) - HootBoard ≤3.1.4 (CVE-2026-86798) - Prenotazioni ≤1.7.5 (CVE-2026-103305) Admin-targeted XSS can be chained to full site takeover.
Malware & Indicator Activity
abuse.ch logged 1,243 new indicators from 101 families (ThreatFox 823, URLhaus 420); 777 were supplied. Vidar dominated. IoT botnet payloads, DanaBot C2, commodity RATs and ClearFake web-inject infrastructure followed.
- High severityVidar infostealer: 237 new indicators
Vidar was the top abuse.ch family this window with 237 new ThreatFox indicators. - Indicator types: URLs, domains, MD5/SHA1/SHA256 hashes, ip:port. - Threat types: botnet_cc, payload. - Samples: hxxps://gp[.]54toto[.]org, hs[.]54toto[.]org, hxxps://hs[.]54toto[.]org/. Block C2 domains and hunt for browser credential theft followed by session hijacking.
- High severityMirai/Mozi IoT botnets: ~220 new payload and C2 indicators
IoT botnet families: - Mirai: 94 ThreatFox indicators (sha256, ip:port, URL, domain; payload, botnet_cc, payload_delivery). - URLhaus ELF downloads: 'elf, ua-wget' (50), Mozi MIPS (48), Mozi ARM (19), and a Cowrie-captured 'akuma' DVR set (12). - Samples: hxxp://209[.]99[.]184[.]98/bins/sora[.]ppc, hxxp://42[.]233[.]104[.]85:36775/bin[.]sh, hxxp://85[.]137[.]53[.]167:24380/bins/dvr/akuma_mpsl.
- High severityDanaBot banking trojan/loader: 41 C2 servers
DanaBot had 41 new ThreatFox botnet_cc indicators, all ip:port. Samples: 66[.]253[.]84[.]79:443, 154[.]82[.]149[.]30:25219, 154[.]86[.]11[.]118:25201. Feed these into egress blocklists and NDR watchlists.
Source: abuse.ch - High severityClearFake web-inject delivery and PHP webshell infrastructure
ClearFake: 27 payload_delivery domains, used for fake browser-update and paste-to-run lures on compromised sites. Samples: payload1717[.]pro, 7aigp3my[.]lunza[.]store, zj9s97wb[.]jp168amp-super[.]com. Separately, 48 php.shin_webshell payload_delivery domains were reported, including workers[.]dev hosts such as lakyxopa[.]workers[.]dev.
- Medium severityCommodity RATs: AsyncRAT, Remcos and Quasar C2 and payloads
Commodity RATs remained active: - AsyncRAT: 27 indicators (botnet_cc, payload, malware_download; ThreatFox and URLhaus). Samples: hxxp://174[.]138[.]21[.]29/csgo[.]exe, 104[.]243[.]248[.]63:3602. - Remcos: 24 C2 indicators (ip:port, domain). Sample: 103[.]83[.]86[.]126:14641. - Quasar RAT: 15 indicators.
- Medium severityAMOS macOS stealer: 20 new indicators
AMOS (Atomic macOS Stealer) had 20 ThreatFox indicators (sha256, URL, domain; payload, payload_delivery, botnet_cc). Sample hash: 656c960966c12f507a7d80c59bb377be09ac14c16903f7a933d3ef32a4210c53. Monitor macOS fleets for unsigned DMG installs and keychain access.
- Medium severitySmartLoader, IRAHook (Dropbox/Discord-hosted) and Antidot activity
- SmartLoader: 11 payload hashes, e.g. 2def5f3586564612e03fa8b57bf3e8f2c4aede6f3be47001890c472b87ff79ee. - IRAHook: 10 malware_download URLs staged on legitimate cloud services as bundle.zip archives, e.g. hxxps://www[.]dropbox[.]com/scl/fi/ovgcqrqvo601c4xrgrsso/bundle[.]zip?rlkey=jpo47d62rtkevfrwz5yfsx93x&st=ic2dii77&dl=1. - Antidot (Android banking trojan): 10 C2 ip:port entries, e.g. 146[.]103[.]38[.]59:28116.
Offensive Frameworks & C2
Adversary emulation and post-exploitation frameworks continue to be stood up at volume.
- High severitySliver, VShell and Cobalt Strike C2 infrastructure
New ThreatFox C2 indicators for three frameworks: - Sliver: 26 C2 servers (ip:port). Samples: 35[.]229[.]230[.]203:50051, 38[.]54[.]110[.]142:31337, 45[.]9[.]149[.]180:31337. - VShell: 25 indicators (URL, ip:port, hashes). Sample: 120[.]26[.]120[.]83:9214. - Cobalt Strike: 23 C2 servers. Samples: 49[.]235[.]158[.]141:7777, 101[.]35[.]247[.]212:80. Outbound connections to these hosts warrant immediate triage as likely hands-on-keyboard activity.
Threat Actors
No law-enforcement takedowns were recorded this window. Actor activity centered on a reported leak of Silent Ransom Group's own files, a burst of high-profile TheGentlemen victim listings, and N0n's countdown-based extortion.
- High severity'Luna Moth Files' leak exposes Silent Ransom Group despite denials (opens the source in a new tab)
DataBreaches.net reports that the 'Luna Moth Files', which appeared on an onion site, are genuine. Silent Ransom Group (Luna Moth), known for callback phishing and data extortion, had called the files fake and denied any leak or breach. According to the report, the files leaked rather than being hacked. The material may offer insight into SRG operations relevant to defenders tracking callback-phishing campaigns.
- High severityTheGentlemen ransomware lists five high-profile victims in a single burst
Between 04:38 and 04:40 UTC on 2026-10-12, TheGentlemen posted five large organizations to its leak site: - MoneyGram (global money transfer) - KFC - Confluent (Kafka data streaming, IBM-owned) - BigID (data security platform) - Placer AI (foot-traffic analytics) Claims are unverified and no data classes were listed. The vendor victims (Confluent, BigID) raise potential downstream supply-chain concerns for their customers.
Source: Manual - High severityN0n extortion group posts anonymized victims with publication countdowns
N0n posted numerous 'Company #N' listings: smart building/surveillance integrators (California), an architecture firm (Florida), a neurology practice (NSW, Australia), a cannabis operator (Michigan) and a confectionery/telehealth firm (New York). Many listings appear to be duplicates across re-posts. Data publication is countdown-scheduled for 2026-10-13 to 2026-10-14. The group marked a Canadian financial services firm (Company #14) as having 'settled' on 2026-10-07.
Source: Manual
Data Breaches & Leaks
29 breach entries were recorded, all ransomware/extortion leak-site claims, plus one confirmed media-reported incident. Spelman Logistics and the Rhysida law-firm leaks are the most credential- and financial-data-exposing claims. No pwnCount or structured data classes were supplied for the leak-site entries.
- High severityNippon Columbia / Daiichi Kosho: 8.7M karaoke customer and employee records exposed (opens the source in a new tab)
Japanese entertainment system maker Daiichi Kosho disclosed that a malware infection at its contractor Nippon Columbia exposed more than 8.7 million customer and employee records tied to its karaoke services. This is a third-party/contractor compromise with large-scale PII exposure.
Source: RSS - High severitySpelman Logistics (Emperador): SSNs, EINs and 57K bank accounts claimed (opens the source in a new tab)
Emperador claims data pulled from the victim's MVMNT TMS: - 1.2M contact records (emails, phone numbers) - 245K carrier profiles with EIN and SSN - 57K bank accounts with routing numbers - 334K live document URLs - 4.5M carrier registry entries If genuine, the data poses a significant fraud and identity theft risk for freight carriers.
Source: Breach - High severityMoneyGram (TheGentlemen) (opens the source in a new tab)
TheGentlemen listed MoneyGram, a global money transfer network serving 60M+ customers across 200+ countries. No data classes or volumes were disclosed, and the claim is unverified. Financial institutions should watch for related fraud and phishing.
Source: Breach - High severityConfluent (TheGentlemen) (opens the source in a new tab)
TheGentlemen listed Confluent, the Apache Kafka data streaming platform acquired by IBM. No data details were published. Customers of Confluent Cloud should review access tokens and monitor vendor advisories.
Source: Breach - High severityEversource Energy (Redact) (opens the source in a new tab)
Redact claimed US energy distribution utility Eversource Energy (~$14B revenue). No data classes were listed. This is a critical infrastructure victim; the claim is unverified.
Source: Breach - High severityReed & Giordano PA (Rhysida): trust account and SSN data (opens the source in a new tab)
Rhysida claims 94,604 files (~51.3 GB) from Reed & Giordano PA, including: - IOLTA client trust account ACH instructions and wire-transfer details - W-9/W-2 forms with SSNs - Massachusetts driver's licenses
Source: Breach - High severityCharles E. Tabor AAL LLC (Rhysida): full client PII database (opens the source in a new tab)
Rhysida claims 96,456 files (~114.5 GB) from Charles E. Tabor AAL LLC, including: - A consolidated client database with PII spanning years - Settlement check details - Hospital-related records
Source: Breach - High severityCNESTEN (Qilin) (opens the source in a new tab)
Qilin listed CNESTEN, which appears to be Morocco's national nuclear science and technology centre (the listing references a company directory). No description or data details were provided; verification is pending. Qilin also listed Friendship Christian School.
Source: Breach - High severityTQC S.A. (DragonForce): 750 GB released (opens the source in a new tab)
DragonForce published 750 GB of compressed archives from TQC Tecnología Química Comercio S.A., a Peruvian provider of agricultural, veterinary and environmental health products.
Source: Breach - Medium severityOther ransomware leak-site listings
Further claims with no data details published: - KFC, BigID and Placer AI (TheGentlemen) - huntcompanies.com (BrainCipher) - MGT Lojistik AŞ (Turkey) and KP Industries (UK) (Nightspire) - Friendship Christian School (Qilin)
Source: Manual
Sources & Attribution
These briefings are compiled from publicly available threat-intelligence feeds, which may include CISA KEV, NIST NVD, the GitHub Advisory Database (OSV), abuse.ch, and Wordfence Intelligence. Data-breach and credential-leak items may include data from Have I Been Pwned and ransomware.live.
CVE® is a registered trademark of The MITRE Corporation. CVE Records are © The MITRE Corporation, reproduced under the CVE Program Terms of Use. WordPress vulnerability data is provided by Wordfence Intelligence, © Defiant, Inc. Breach data from Have I Been Pwned is licensed under CC BY 4.0.