PSA and RMM

Solve any challenge with one platform

Operate more efficiently, reduce complexity, improve EBITDA, and much more with the purpose-built platform for MSPs.

Cybersecurity and Data Protection

Ensure security and business continuity, 24/7

Protect and defend what matters most to your clients and stakeholders with ConnectWise's best-in-class cybersecurity and BCDR solutions.

Automation and Integrations

Integrate and automate to unlock cost savings

Leverage generative AI and RPA workflows to simplify and streamline the most time-consuming parts of IT.

University

University Log-In

Check out our online learning platform, designed to help IT service providers get the most out of ConnectWise products and services.

About Us

Experience the ConnectWise Way

Join hundreds of thousands of IT professionals benefiting from and contributing to a legacy of industry leadership when you become a part of the ConnectWise community.

News and Press

Experience the ConnectWise Way

Join hundreds of thousands of IT professionals benefiting from and contributing to a legacy of industry leadership when you become a part of the ConnectWise community.

ConnectWise

9/22/2025 | 15 Minute Read

Monthly Threat Brief: August 2025

Topics:

Contents

    Ready to dive even deeper?

    Check out the 2025 MSP Threat Report for an in-depth analysis of overall trends.

    Welcome to the August 2025 edition of the ConnectWise Cyber Research Unit™ (CRU) Monthly Threat Brief. This report provides a comprehensive look at the most significant cybersecurity developments impacting the MSP and SMB landscape. We break down the month’s top stories, critical vulnerabilities, and prevalent malware families to help you stay informed and prepared in an evolving threat environment. We have also begun adding descriptions of all the new detection signatures added to the ConnectWise SIEM™.

    Top stories for August 2025

    Akira Ransomware SSLVPN Activity

    In the past month, we’ve tracked a significant uptick in Akira ransomware incidents tied to SonicWall SSLVPN appliances, with activity first observed in late July. Early reporting suggested a possible zero-day vulnerability, but SonicWall has since clarified that these attacks are strongly linked to CVE-2024-40766 and poor credential management practices, such as reused passwords during firewall migrations. Research from Arctic Wolf, Huntress, and Google Threat Intelligence Group (GTIG) indicates that the threat actor UNC6148 is leveraging previously stolen credentials and OTP seeds to regain access even after security updates are applied. ConnectWise SOC observations reinforce this assessment, with 15 confirmed Akira incidents over the last 30 days, many involving direct logins through SSLVPN with no failures, pointing toward credential compromise rather than exploitation.

    Once inside, the attackers have relied heavily on living-off-the-land techniques, complemented by tools such as NetExec and pypykatz for network and credential harvesting. We’ve also seen deliberate efforts to evade endpoint detection and response (EDR) tools through custom DLLs, drivers, and registry checks, as well as persistence established with AnyDesk, TeamViewer, and MeshAgent. Prior to launching encryption payloads, typically disguised under changing names such as w.exe or Akira.exe, actors deployed utilities such as WinRAR, FileZilla, Rclone, and Cloudflare tunneling for data theft and lateral movement. This campaign continues the broader trend of edge appliance exploitation fueling ransomware activity, underscoring the need for timely patching, enforced password resets, multi-factor authentication (MFA) adoption, and vigilant monitoring of both firewall and endpoint logs across customer environments.

    Critical Privilege Escalation Risk Targets Exchange Hybrid Deployments

    Over the past month, one of the most important advisories came from Microsoft regarding CVE-2025-53786, a critical flaw in hybrid Exchange environments. The vulnerability stems from the long-standing shared service principal architecture, which allows on-premises Exchange servers and Exchange Online to authenticate using the same credentials. If attackers gain admin access to an Exchange server, they can extract OAuth certificates and request service tokens that impersonate hybrid users for up to 24 hours, bypassing conditional access and MFA controls. Although no active exploitation has been reported, Microsoft and CISA view the risk as severe, with CISA issuing Emergency Directive 25-02 mandating immediate mitigations for federal agencies by August 11, 2025.

      To address the design flaw, Microsoft is forcing a shift away from shared service principals and toward dedicated hybrid applications in Entra ID, giving each organization its own isolated authentication entity. The company has laid out an aggressive enforcement timeline beginning with temporary service disruptions in August and culminating in a permanent block of legacy authentication on October 31, 2025. MSPs should prioritize patching on-premises Exchange servers, deploying the dedicated hybrid apps, and running health checks to ensure configurations are correct. Beyond the technical risk of domain-wide compromise, the looming service disruptions create business pressure for timely migration, as clients relying on hybrid features, such as free/busy lookups and MailTips, will see outages if they don’t move quickly. For service providers, this is both a pressing security matter and a client communication challenge; one where deadlines, business continuity, and compliance all converge.  

    WinRAR Zero-Day Used in Targeted Phishing Campaigns

    In July, a Russian-aligned group, RomCom, was observed exploiting a newly disclosed zero-day in WinRAR (CVE-2025-8088) through targeted phishing campaigns. The flaw, a path traversal vulnerability in how WinRAR handles alternate data streams (ADS), allows attackers to silently extract malicious files into attacker-controlled directories instead of user-specified paths. This technique enabled the delivery of malicious DLLs, LNK files, and executables that established persistence and command-and-control access. ESET researchers documented three distinct execution chains, leveraging tools such as a Mythic agent, a modified PuTTY CAC variant dubbed SnipBot, and RustyClaw/MeltingClaw, each incorporating anti-analysis checks and invalidly signed binaries. Campaigns between July 18 and 21 targeted financial, manufacturing, defense, and logistics organizations in Europe and Canada, though no successful compromises were reported.  

    RomCom’s exploitation of CVE-2025-8088 marks the group’s third zero-day in two years, following prior use of Microsoft Word and Firefox vulnerabilities. The attack underscores both the sophistication and persistence of the threat actor, who routinely pairs zero-day exploitation with tailored phishing lures. For MSPs, the key concern is that WinRAR lacks automatic updates, requiring manual deployment of the patched version 7.13 across all endpoints. The attack surface extends beyond standalone WinRAR installations to any software using UnRAR.dll, meaning client applications must also be audited for exposure. While this campaign did not result in compromise, the pattern of zero-day use against sectors common to MSP clients reinforces the importance of patch management, monitoring for archive-based attacks, and enhancing email security controls around RAR attachments.

    Browser Extension Clickjacking Technique Targets Password Managers and Other DOM-Injecting Tools

    This month, researchers disclosed a new attack technique called **DOM-based Extension Clickjacking**, which abuses how browser extensions inject interface elements into web pages. By making these injected elements invisible, often by setting their opacity to zero, attackers can trick users into interacting with hidden password manager popups or other extension UI elements when clicking on seemingly harmless items like cookie banners or CAPTCHA prompts. Security researcher Marek Toth tested 11 major password managers and found all were vulnerable to at least one variant of the attack, with six leaking complete credit card details and eight exposing personal information under test conditions. The issue extends beyond password managers to any extension that manipulates the DOM, including crypto wallets, note-taking apps, and productivity tools. Some passkey authentication implementations were also shown to be vulnerable due to improper session binding, allowing attackers to hijack authentication flows.  

      For MSPs, this disclosure underscores the risk posed by browser extensions as an overlooked attack surface. As of August 2025, vendors such as Dashlane, Keeper, NordPass, ProtonPass, and RoboForm have released fixes, while Bitwarden, 1Password, LastPass, Enpass, iCloud Passwords, and others remain exposed, affecting tens of millions of users. Since these attacks operate entirely within the browser, they generate minimal forensic evidence and bypass traditional SIEM and SOC detection. Recommended mitigations include configuring extensions to activate only “on click” rather than persistently, auditing all deployed extensions for UI injection behaviors, and tightening browser extension policies for high-value accounts. This research highlights the need to treat extensions as high-risk software with privileged access, balancing convenience against the security trade-offs they introduce.  

    Top vulnerability in August 2025  

    CVE‑2025‑8088

    CVE‑2025‑8088 is a high‑severity path traversal zero‑day in the Windows version of WinRAR (and associated components like UnRAR.dll) that allows arbitrary code execution by extracting crafted archives into sensitive directories such as the Startup folder, establishing persistence, and often delivering backdoors automatically. Researchers discovered exploitation beginning July 18, 2025, with the RomCom (also Storm‑0978/Tropical Scorpius/UNC2596) group deploying malware, including SnipBot, RustyClaw, and Mythic Agent via spear‑phishing archives disguised as job applications targeting financial, manufacturing, defense, and logistics firms across Europe, Canada, and organizations linked to humanitarian efforts. WinRAR released a patch in version 7.13 (July 30, 2025), but since the software lacks auto‑update capabilities, users must upgrade manually, which is especially urgent as CISA added this CVE to its Known Exploited Vulnerabilities (KEV) catalog on August 12, 2025, mandating federal remediation by September 2, 2025.

    Top malware in August 2025

    The Diamond Model

    This section leverages the Diamond Model of Intrusion Analysis to structure the examination of recent malware activity, providing a clear analytical framework that links adversary, capabilities, infrastructure, and victimology. By applying the Diamond Model, we can better contextualize malicious behavior, identify patterns across campaigns, and highlight the relationships between threat actors, tools, and targeted entities.

    Akira

    Akira is a ransomware strain that first emerged in March 2023. It operates under a ransomware‑as‑a‑service (RaaS) model, typically targeting both Windows and Linux systems within small and mid-sized organizations across sectors such as manufacturing, IT, healthcare, finance, professional services, and education. Notable victims include BHI Energy, Nissan Australia, Tietoevry, Stanford University, and potentially the Toronto Zoo. The group employs double‑extortion tactics, exfiltrating sensitive data before encrypting systems and threatening to leak the stolen information.

    Starting in July 2025, security researchers observed an uptick in Akira deployments exploiting SonicWall SSL VPN appliances as the initial attack vector. Early speculation pointed to a zero‑day vulnerability, as some breaches occurred on fully patched SonicWall devices, even those protected by MFA and recently rotated passwords. However, further analysis, including official statements from SonicWall, ThreatLocker, S‑RM, and others, suggests that the attacks are tied to a previously disclosed vulnerability, CVE‑2024‑40766, rather than an unknown flaw. Mitigation guidance emphasizes updating to SonicOS7.3.0, resetting migrated local account credentials, disabling or restricting SSLVPN exposure, enforcing MFA (though it may not fully prevent attacks in this case), enabling Botnet Protection and Geo‑IP filtering, and applying strict password hygiene.

    Additionally, recent incident response analysis (late July–early August 2025) has uncovered bring‑your‑own‑vulnerable‑driver (BYOVD) techniques used by Akira affiliates. Specifically, two Windows drivers, “rwdrv.sys” (a legitimate ThrottleStop driver) and “hlpdrv.sys” (a malicious helper driver), have been deployed to disable Microsoft Defender and evade antivirus/EDR protections.

    Aliases

    • Akira, Megazord, Redbike

    Infrastructure

    Recent IOCs

    Related File(s)
    SHA256 Filename(s)

    d0db094355ac9727a280be3466e5fa113ac88fa9108c4c8ef541e405f6b3ec0a

    Win.exe

    Victimology

    • Recently targeted business sectors: Construction, attorneys, healthcare

    Capabilities  

    MITRE ATT&CK Techniques

    Tactic Technique ID Technique Name
    Initial Access   T1078   Valid Accounts
    Initial Access   T1190   Exploit Public-Facing Application
    Initial Access   T1133   External Remote Services
    Initial Access   T1566.001   Phishing: Spearphishing Attachment
    Initial Access   T1566.002   Phishing: Spearphishing Link
    Credential Access   T1003   OS Credential Dumping
    Credential Access   T1003.001   OS Credential Dumping: LSASS Memory
    Discovery   T1016   System Network Configuration Discovery
    Discovery   T1082   System Information Discovery
    Discovery   T1482   Domain Trust Discovery
    Discovery   T1057   Process Discovery
    Discovery   T1069.001   Permission Groups Discovery: Local Groups
    Discovery   T1069.002   Permission Groups Discovery: Domain Groups
    Discovery   T1018   Remote System Discovery
    Persistence   T1136.002   Create Account: Domain Account
    Defense Evasion   T1562.001   Impair Defenses: Disable or Modify Tools
    Command and Control   T1219   Remote Access Software
    Command and Control   T1090   Proxy
    Collection   T1560.001   Archive Collected Data: Archive via Utility
    Exfiltration   T1048   Exfiltration Over Alternative Protocol
    Exfiltration   T1537   Transfer Data to Cloud Account
    Exfiltration   T1567.002   Exfiltration Over Web Service: Exfiltration to Cloud Storage
    Impact   T1486   Data Encrypted for Impact
    Impact   T1490   Inhibit System Recovery
    Impact   T1657   Financial Theft
    FakeUpdates

    SocGholish, also known as FakeUpdates, is a JavaScript-based malware framework attributed to the threat actor TA569, also tracked as Mustard Tempest, UNC1543, and others. Active since at least 2017, SocGholish is typically distributed through compromised legitimate websites, often leveraging outdated CMS platforms. Victims are presented with fake browser or software update prompts, which, if clicked, initiate the download of malicious JavaScript loaders. These loaders serve as a foothold for deploying secondary payloads such as remote access tools, Cobalt Strike, and ransomware variants such as RansomHub. The group primarily targets organizations in North America and Europe across sectors, including government, education, legal, and healthcare, using SocGholish as an initial access broker for other threat actors, including ransomware groups like Evil Corp and LockBit.

    In August 2025, new intelligence highlighted SocGholish’s expanded use of traffic distribution systems (TDS), such as Keitaro and Parrot TDS, to more precisely funnel users from compromised sites to malware payloads while evading detection by security researchers and sandboxes. The actors now actively monitor victims throughout the infection chain and selectively deliver payloads only to targets deemed high-value or legitimate, aborting on virtual machines or analysis environments. Researchers at Silent Push and The Hacker News reported increased use of domain shadowing and dynamic JavaScript obfuscation techniques, along with a MaaS delivery model where TA569 sells access to corporate networks to ransomware operators. This evolution reinforces SocGholish’s role as a persistent and stealthy threat with wide-reaching impact in the cybercrime ecosystem.

    Aliases

    • SocGholish

    Infrastructure

    Recent IOCs

    Related File(s)
    SHA256 Filename(s)

    c938fcdf21466e093c155ebb2d68e0326866f0854369d65bd9cafc75e10df5fc

    installer1.js

    e82d841558a1be1e237b9537ecd80f0cac0ba4ea2a987f5909106535356f0753

    140.8662.84.js

    6bf58141283a15fa912e24c7cbaefac5313d1a9f07714bb6706a95bac7ee73df

    Chrome.ad6cddcd.js

    3bebad0f3e20134c397e73e42e37bdade93940a1849b0e1e8b96bb50251bc113

    Chrome.3d34d1e1.js

    954621306ccb67b6dd5b22ac11cf615523ee9d653e5d9e2dd9fa2485f1d6eab8

    Chrome.ae5520cb.js
    Related Domain Name(s)
    app.makemoremoneychallenge[.]vip
    aysrcz2333[.]top
    hr.lexnational[.]com
    hr.lexnational[.]com
    Related IP Address(es)
    209.141.51[.]11
    173.44.141[.]44
    209.141.51[.]11

    Victimology

    • Recently targeted business sectors: Healthcare, non-profits, IT services

    Capabilities

    MITRE ATT&CK Techniques

    Tactic Technique ID Technique Name
    Resource Development   T1608.006   Stage Capabilities: SEO Poisoning
    Initial Access   T1189   Drive-by Compromise
    Execution   T1059.001   Command and Scripting Interpreter: PowerShell
    Execution   T1059.007   Command and Scripting Interpreter: JavaScript
    Execution   T1204.002   User Execution: Malicious File
    Persistence   T1547.001   Boot or Logon Autostart Execution: Registry Run Keys / Startup Folder
    Discovery   T1482   Domain Trust Discovery
    Discovery   T1087.001   Account Discovery: Local Account
    Discovery   T1087.002   Account Discovery: Domain Account
    Discovery   T1069.002   Permission Groups Discovery: Domain Groups
    Discovery   T1082   System Information Discovery
    Discovery   T1018   Remote System Discovery
    Discovery   T1069.001   Permission Groups Discovery: Local Groups
    Discovery   T1033   System Owner/User Discover
    Command and Control   T1105   Ingress Tool Transfer
    Credential Access   T1555   Credentials from Password Stores
    GuLoader

    GuLoader is a highly evasive shellcode-based downloader first spotted in late 2019, also known as CloudEyE, that’s typically distributed via phishing campaigns, often through ZIP or Visual Basic Script attachments, or embedded in NSIS installers sent via malspam. It uses sophisticated anti‑analysis and obfuscation techniques, such as string encryption, control flow obfuscation, and delayed execution, to evade antivirus and sandbox detection. Its primary role is to fetch and execute secondary payloads, such as information stealers, remote access trojans (RATs), and other malware. Researchers have dissected its loader behavior, unpacking malicious shellcode via junk data offsets and NSIS plugins, and extracting its encrypted configuration for deeper analysis.

    Aliases

    • Vbdropper, CloudEyE

    Infrastructure

    Recent IOCs

    Related File(s)
    SHA256 Filename(s)

    cec1a7d8159b98f062a9499286fbf66de6891fce3e2bfc8fd0a22459c39e92ce

    generated_script (3).vbs

    Related Domain Name(s)

    dethintonsx[.]com

    bbrlxwtkbgvpnpfykexu.supabase[.]co

      Victimology

      • Recently targeted business sectors: IT services, retail, attorneys

      Capabilities  

      MITRE ATT&CK Techniques

      Tactic Technique ID Technique Name
      Initial Access   T1566.002   Phishing: Spearphishing Link
      Defense Evasion   T1070.004   Indicator Removal: File Deletion
      Defense Evasion   T1055   Process Injection
      Defense Evasion   T1497.001   Virtualization/Sandbox Evasion: System Checks
      Defense Evasion   T1497.003   Virtualization/Sandbox Evasion: Time Based Evasion
      Command and Control   T1105   Ingress Tool Transfer
      Command and Control   T1071.001   Application Layer Protocol: Web Protocols
      Command and Control   T1102   Web Service
      Execution   T1204.002   User Execution: Malicious File
      Execution   T1204.001   User Execution: Malicious Link
      Execution   T1106   Native API
      Execution   T1547.001   Boot or Logon Autostart Execution: Registry Run Keys / Startup Folder

      New detections

      The following is a list of new detection signatures added to the ConnectWise SIEM in August 2025.

      [O365] User Compromised or Compromised SignIn event

      Detects Microsoft 365 (Entra) “UserCompromised” or “CompromisedSignIn” operations. We recommend investigating this activity. To correlate with other user events, match values from “UserKey” to the fields “userId” or “userStates.aadUserId.” To remediate, sign the user out of all sessions, reset the user’s password, and ensure MFA is enabled. Please note that offline detection timing types may cause delayed reporting from Microsoft.

      [Azure] Possible credential leak: Successful sign-in blocked by Conditional Access

      Detects a user sign-in using a valid password which is subsequently blocked by a conditional access policy. This can indicate possible usage of valid credentials by an attacker. If the sign-in is blocked by a GeoBlocks policy, an attacker may be able to use a VPN or otherwise spoof locations or other parameters of the policy to bypass the CAP block. We recommend investigating this activity.

      [Windows] Windows Sandbox installation registry event

      Detects when a registry value is set for the Windows Sandbox (WSB), indicating that this feature has been installed on the host. Windows Defender does not run within Windows Sandbox, so this feature can be used by attackers to covertly execute malware or gain persistence. Although this feature requires local administrator privileges to enable, we recommend verifying that this is an expected activity in your environment.

      [Windows] Windows Sandbox config file creation

      Detects the creation of a “wsb” file type, indicating a Windows Sandbox configuration file has been created. Windows Defender does not run within the Windows Sandbox, so this feature can be used by attackers to evade detection and establish persistence via covert malware execution or C2 from within the sandbox. Although this feature requires local administrator privileges to enable, we recommend verifying that this is expected activity in your environment.

      [Windows] Windows Sandbox command line activity

      Detects command-line usage of Windows Sandbox. Windows sandbox configurations can be run as commands in the CLI, evading creation of a “.wsb” config file. Windows Defender does not run within the Windows Sandbox, so this feature can be used by attackers to evade detection and establish persistence via covert malware execution or C2 from within the sandbox. Although this feature requires local administrator privileges to enable, we recommend verifying that this is expected activity in your environment.

      [Windows] Windows Sandbox scheduled task creation

      Detects when a scheduled task is created for Windows Sandbox to run. Windows Defender does not run within the Windows Sandbox, so this feature can be used by attackers to evade detection and establish persistence via covert malware execution or C2 from within the sandbox. Although this feature requires local administrator privileges to enable, we recommend verifying that this is expected activity in your environment.

      Related Articles