A vulnerability in Microsoft Authenticator for iOS and Android, identified as CVE-2026-26123, could allow a malicious application on the same device to intercept one-time sign-in codes or authentication links.
Microsoft has patched this information disclosure vulnerability and recommends all users update their application to the latest version.
Vulnerability Details
Affected Product: Microsoft Authenticator for iOS and Android
Identifier: CVE-2026-26123
CVSS Score: 5.5 (Medium)
Exploitation Status: No known active exploitation
Risk & Impact
Triage: High Priority
Attack Vector: An attacker must trick a user into installing a malicious application on their device and then selecting that app to handle a sign-in process. If successful, the malicious app could receive the sign-in information and use it to authenticate as the user.
Ease of Exploit: Requires user interaction, making the immediate risk lower but still significant, especially in BYOD (Bring Your Own Device) environments.
Action Plan
Immediate Action: Update Microsoft Authenticator to the latest version available in the Apple App Store or Google Play Store.
Workaround: Be cautious about installing new applications, especially those that request to handle authentication links or sign-in flows. Use mobile anti-malware protection to help detect malicious apps.
Detection: Monitor for unusual application installations or sideloading attempts on managed devices. Review sign-in logs for suspicious authentication patterns.
Relevant professional terms
Deep Link
A type of URI (Uniform Resource Identifier) that links to a specific piece of content or screen within a mobile application, rather than just launching the app's homepage. Attackers can exploit improperly secured deep links to hijack data or control.
Multi-Factor Authentication (MFA)
A security process that requires users to provide two or more verification factors to gain access to a resource, such as an application, online account, or a VPN. Microsoft Authenticator is a widely used application for providing one of these factors.
CISA has added a critical remote code execution (RCE) vulnerability in the n8n workflow automation platform, CVE-2025-68613, to its Known Exploited Vulnerabilities (KEV) catalog due to evidence of active exploitation.
The flaw allows authenticated attackers to execute arbitrary code, potentially leading to a full compromise of the affected instance.
Vulnerability Details
Affected Product: n8n versions 0.211.0 and later, prior to 1.120.4, 1.121.1, and 1.122.0
Identifier: CVE-2025-68613
CVSS Score: 9.9 (Critical)
Exploitation Status: Actively Exploited
Risk & Impact
Triage: Immediate upgrade or mitigation is required due to active exploitation and critical severity.
Attack Vector: An authenticated attacker can inject a malicious expression during workflow configuration. This expression is then evaluated in an insufficiently isolated context, allowing the attacker to execute arbitrary code with the privileges of the n8n process.
Ease of Exploit: Low. The vulnerability can be exploited by any authenticated user with permissions to create or edit workflows.
Action Plan
Immediate Action: Upgrade to a patched version: 1.120.4, 1.121.1, or 1.122.0 and later.
Workaround: If immediate patching is not possible, limit workflow creation and editing permissions to fully trusted users and consider deploying n8n in a hardened, isolated environment.
Detection: Sangfor has released detection rules (ID 11220072) for its NDR and WAF products. A Nuclei template is also publicly available for testing.
Relevant professional terms
Remote Code Execution (RCE)
A vulnerability that allows a malicious actor to execute arbitrary commands or code on a target machine or in a target process from a remote location. RCEs are typically considered critical because they can lead to a full system compromise.
Known Exploited Vulnerabilities (KEV) Catalog
A list curated by CISA that contains vulnerabilities confirmed to have been actively exploited in real-world attacks. It is used to help organizations prioritize remediation efforts on the most urgent threats.
An unidentified attacker compromised a popular GitHub Action from security vendor Xygeni by using a "tag poisoning" technique.
This allowed the attacker to inject a command-and-control (C2) implant, which was active for approximately one week.
Attack Overview
Attack Path: The attacker used compromised credentials (a maintainer's Personal Access Token and a GitHub App private key) to move a mutable git tag (@v5) to point to a malicious commit. This action served a C2 implant to any workflow using the compromised tag.
Attacker: The threat actor remains unidentified.
Impact Assessment
Operational Impact: Any workflow referencing `xygeni/xygeni-action[@]v5` between March 3 and March 10, 2026, executed the C2 implant, granting the attacker arbitrary command execution on the CI runner. Xygeni states there is no evidence of compromise to its platform or customer data.
Detection & Hunting
IOCs: The malicious commit hash is 4bf1d4e19ad81a3e8d4063755ae0f482dd3baf12.
Strategic Takeaway
This incident highlights a critical software supply chain risk where mutable tags can be manipulated; developers should pin actions to immutable commit hashes to prevent such compromises.
Relevant professional terms
Tag Poisoning
An attack where a threat actor manipulates a mutable Git tag (like `v1` or `latest`) to point to a malicious commit, tricking users and CI/CD systems into pulling compromised code.
C2 (Command and Control)
An infrastructure of servers and software used by attackers to send commands to and receive data from a compromised system.
The Iran-linked hacktivist group Handala has claimed responsibility for a destructive wiper malware attack against medical technology leader Stryker.
The attack caused a global network disruption, impacting systems and forcing the company to take offices in 79 countries offline.
Key TTPs
Initial Access: The group is known to use phishing campaigns, often themed around current events, to gain initial access.
Execution: In this attack, the actors reportedly used Microsoft Intune to issue a remote wipe command to all connected devices, including phones and laptops.
Campaign Analysis
This incident highlights a trend of geopolitically motivated, destructive attacks targeting critical infrastructure. Handala, believed to be a front for the Iranian government-sponsored actor Void Manticore, combines data wiping with information operations to maximize psychological and reputational damage.
Targeting & Infrastructure
Target Profile: The primary target was a Fortune 500 medical technology company, indicating a focus on the healthcare and critical manufacturing sectors.
Relevant Terms
Wiper Malware: A type of malicious software designed to permanently erase data from a targeted computer or network, with the primary goal of destruction rather than financial gain.
Hacktivist: An individual or group who uses hacking techniques to advance a political or social agenda, often by defacing websites, stealing data, or disrupting services.
The PhantomRaven campaign is actively targeting JavaScript developers by publishing dozens of malicious packages to the npm registry.
The primary goal is to exfiltrate sensitive data, including authentication tokens and CI/CD secrets, from developer environments.
Key TTPs
Initial Access: Malicious packages are published to the npm registry, often using plausible names that could be mistakenly installed by developers (a technique called slopsquatting).
Execution: A pre-install hook in the malicious package automatically triggers the execution of the main payload upon installation.
Defense Evasion: The campaign uses Remote Dynamic Dependencies (RDD), where the malicious code is fetched from an external attacker-controlled server during installation, bypassing static analysis and dependency scanning tools.
Campaign Analysis
This campaign represents a significant evolution in software supply chain attacks, moving beyond simple typosquatting to manipulate the dependency resolution process itself.
The attackers continuously publish new waves of packages, indicating a persistent and adaptive operation targeting the open-source ecosystem.
Targeting & Infrastructure
Target Profile: The campaign targets JavaScript developers, open-source contributors, and organizations relying on the npm ecosystem.
Infrastructure: Attackers leverage the public npm registry for distribution and use external servers to host malicious payloads.
Actionable Intelligence
Domains:packages.storeartifact[.]com
Relevant Terms
npm: A package manager for the JavaScript programming language that helps developers share and reuse code.
Supply Chain Attack: A cyberattack that targets a trusted third-party vendor or software that provides services or software to the ultimate target organization.
Security researchers demonstrated that agentic AI web browsers, such as Perplexity's Comet, can be manipulated into executing malicious actions. These browsers were successfully lured into phishing scams in under four minutes by exploiting their autonomous decision-making processes.
Key Findings
Attackers use prompt injection, hiding malicious commands in seemingly benign content like calendar invites or webpages, which the AI executes without user knowledge.
The core vulnerability is the AI's inability to distinguish between trusted user commands and untrusted instructions from web content.
Successful attacks have forced AI agents to leak local files, exfiltrate data to attacker servers, and interact with phishing sites.
The Bottom Line
The rise of agentic AI browsers introduces a fundamental shift in the attack surface. Security threats are moving from deceiving the human user to directly manipulating the AI agent.
This creates a scalable threat where a single successful exploit against an AI model can be reliably used against every user of that service.
Future security models must evolve to validate and sandbox AI actions, treating the agent itself as a potential insider threat.
Relevant Terms
Agentic AI: An AI system capable of autonomously executing multi-step tasks and making decisions on behalf of a user to achieve a specific goal.
Prompt Injection: An attack technique where malicious instructions are hidden within content that an AI processes, causing it to perform unintended actions.
An international law enforcement operation, led by the U.S. Justice Department, has disrupted the SocksEscort residential proxy network.
This service utilized the AVRecon malware to compromise thousands of Linux-based SOHO routers and other devices worldwide, enabling large-scale fraud.
The Scheme
TTP 1: Compromised thousands of small office/home office (SOHO) routers and edge devices.
TTP 2: Deployed AVRecon, a Linux-based malware, to create a botnet and gain persistent access.
TTP 3: Sold access to the compromised devices as a criminal proxy service to conceal malicious activities like fraud and password spraying.
The Consequence
Outcome: A court-authorized operation seized 34 domains and 23 servers, dismantling the network's infrastructure.
Assets Seized/Forfeited: The United States froze approximately $3.5 million in cryptocurrency.
Strategic Takeaway
This multinational operation highlights a focus on disrupting the underlying infrastructure that enables a wide range of cybercrime, rather than just pursuing individual criminals.
Relevant Terms
Proxy Network: A service that routes a user's internet traffic through an intermediary server, hiding the user's original IP address and location.
Botnet: A network of private computers infected with malicious software and controlled as a group without the owners' knowledge, e.g., to send spam or conduct attacks.
