The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has added a command injection vulnerability in VMware Aria Operations, identified as CVE-2026-22719, to its Known Exploited Vulnerabilities (KEV) catalog, confirming its active exploitation in attacks.
This high-severity flaw allows an unauthenticated attacker to achieve remote code execution.
Vulnerability Details
Affected Product: VMware Aria Operations versions 8.x prior to 8.18.6. Also impacts VMware Cloud Foundation and vSphere Foundation.
Identifier: CVE-2026-22719
CVSS Score: 8.1 (High)
Exploitation Status: Actively Exploited
Risk & Impact
Triage: Urgent. Federal agencies are required by CISA to patch this vulnerability by March 24, 2026.
Attack Vector: An unauthenticated attacker with network access can inject and execute arbitrary commands during a support-assisted product migration process, leading to remote code execution.
Ease of Exploit: Low. The vulnerability can be exploited remotely without authentication, although it is limited to the timeframe when a product migration is in progress.
Action Plan
Immediate Action: Upgrade VMware Aria Operations to version 8.18.6.
Workaround: For those unable to patch immediately, Broadcom has provided a shell script ("aria-ops-rce-workaround.sh") that can be run as root on each Aria Operations node to mitigate the flaw.
Detection: Monitor network traffic and system logs on VMware Aria appliances for any unusual command execution or unauthorized access, particularly during migration activities.
Relevant professional terms
Remote Code Execution (RCE)
A type of vulnerability that allows an attacker to execute arbitrary code or commands on a target machine from a remote location, often leading to full system compromise.
CISA KEV Catalog
The Known Exploited Vulnerabilities (KEV) catalog is a list maintained by the U.S. Cybersecurity and Infrastructure Security Agency (CISA) that identifies vulnerabilities known to be actively exploited in the wild. Federal agencies are required to remediate these vulnerabilities within specified timeframes.
A high-severity vulnerability, tracked as CVE-2026-21385, in a Qualcomm graphics component is under limited, targeted exploitation against Android devices.
Google has released a patch as part of the March 2026 Android Security Bulletin to address the flaw.
Triage: Urgent; immediate patching is recommended due to active, targeted exploitation.
Attack Vector: The vulnerability is described as an integer overflow or buffer over-read in the graphics component. A local attacker could exploit this to trigger memory corruption and access sensitive memory data.
Ease of Exploit: Requires a local attack vector, suggesting a malicious application or user interaction may be necessary for initial access.
Action Plan
Immediate Action: Apply the March 2026 Android security updates, specifically patch level 2026-03-05 or later.
Workaround: Avoid installing applications from untrusted or unknown sources to mitigate the risk of local exploitation.
Detection: Monitor devices for signs of memory corruption, unexpected reboots, or unauthorized application behavior.
Relevant professional terms
Zero-Day Vulnerability
A security flaw in software, hardware, or firmware that is known to the vendor but does not have an official patch or fix available. This vulnerability was considered a zero-day until it was patched.
Integer Overflow
A type of arithmetic overflow error that occurs when the result of an integer operation exceeds the maximum value that the allocated memory space can store, potentially leading to memory corruption or other unexpected behavior.
Researchers discovered a suite of vulnerabilities, dubbed "PleaseFix," in agentic AI browsers that allow attackers to hijack the AI agent, access local files, and exfiltrate data using a malicious calendar invite.
The vulnerabilities, which do not have a CVE identifier, were reported to the vendor Perplexity and have been patched as of February 2026.
Vulnerability Details
Affected Product: Perplexity Comet on macOS, Windows, and Android. Other agentic browsers may also be at risk.
Exploitation Status: Proof-of-Concept. Not reported as actively exploited in the wild.
Risk & Impact
Triage: High. The vulnerability allows for zero-click data exfiltration and potential credential theft from password managers.
Attack Vector: The attack is initiated by sending a specially crafted calendar invitation to a user. When the user instructs the AI browser to accept or process the invite, hidden malicious instructions (an indirect prompt injection) are executed, allowing the agent to access local files, browse directories, and exfiltrate data.
Ease of Exploit: Simple. The attack does not require traditional malware or exploiting a software bug; it leverages the intended functionality of the AI agent.
Action Plan
Immediate Action: Users of Perplexity Comet should ensure their browser is updated to the latest version, which includes the final patch released in February 2026.
Workaround: Be cautious when processing external content, such as calendar invites or emails, with agentic AI browsers. Avoid instructing the AI to handle content from untrusted sources.
Detection: Monitor for unusual file access or network traffic originating from the AI browser process, especially after processing calendar invites or other external documents.
Relevant professional terms
Agentic AI Browser
A web browser with an integrated AI agent that can go beyond simply displaying content to autonomously interpret instructions and execute tasks across different applications and services on behalf of the user.
Indirect Prompt Injection
An attack technique where malicious instructions are hidden within external content (like a webpage or calendar invite) that an AI model processes. The AI cannot distinguish between the user's original prompt and the attacker's hidden instructions, causing it to perform unintended and potentially harmful actions.
Google's Threat Intelligence Group identified "Coruna," a powerful exploit kit targeting iPhones with iOS versions 13.0 to 17.2.1. The kit, also known as CryptoWaters, bundles 23 exploits across five chains to deliver payloads designed to steal financial information and sensitive data.
Key TTPs
Initial Access: Watering hole attacks on compromised websites, including Ukrainian news sites and fake Chinese cryptocurrency platforms.
Execution: Exploits vulnerabilities in Safari for Remote Code Execution (RCE), followed by Local Privilege Escalation (LPE) to gain control.
Defense Evasion: The framework checks for and terminates execution if the device is in Lockdown Mode or private browsing. Payloads are encrypted with ChaCha20.
Campaign Analysis
The Coruna kit exemplifies the proliferation of sophisticated, nation-state-grade tools to financially motivated criminal groups. Its use by a surveillance vendor customer, a Russian espionage group, and a Chinese criminal actor shows a clear escalation and broadening of the mobile threat landscape.
Targeting & Infrastructure
Target Profile: Initially targeted attacks, then Ukrainian iPhone users via geolocation, and later broad-scale attacks on users of cryptocurrency and pornography sites.
Infrastructure: Hosted on compromised Ukrainian websites and a large network of fake Chinese financial sites. One observed domain was mxbc-v2[.]tjbjdod[.]cn.
Relevant Terms
Exploit Kit: A software toolkit designed to identify and exploit vulnerabilities on a target device, often to deliver a malicious payload.
Exploit Chain: A sequence of multiple exploits used together to compromise a system, where each step enables the next to achieve full control.
Threat actors are distributing the XWorm Remote Access Trojan (RAT) using multi-stage infection chains and deceptive techniques to evade detection. These campaigns leverage phishing and malicious scripts to gain initial access for data theft, surveillance, and deploying additional malware.
Key TTPs
Initial Access: Phishing emails with malicious attachments like ZIP archives, LNK files, or weaponized documents.
Execution: Utilizes PowerShell and VBScript to download and run payloads, often with obfuscated or encoded commands.
Defense Evasion: Employs process hollowing to inject its code into legitimate processes like `aspnet_compiler.exe` and checks for virtual environments to avoid analysis.
Campaign Analysis
The continuous evolution of XWorm's infection chain marks a strategic shift towards more intricate methods to bypass security controls.
Its modular design allows attackers to load plugins for various tasks, from keylogging and data theft to launching DDoS attacks, making it a versatile threat.
Targeting & Infrastructure
Target Profile: Opportunistic, targeting a wide range of industries including healthcare, finance, and government.
Infrastructure: Command-and-control (C2) communication occurs over TCP to dynamic DNS domains or direct IPs, often using non-standard ports.
Relevant Terms
Remote Access Trojan (RAT): Malware that provides an attacker with full remote control over an infected computer, enabling data theft, surveillance, and execution of commands.
Process Hollowing: A defense evasion technique where an attacker creates a new process in a suspended state, replaces its legitimate code with malicious code, and then resumes the process.
Threat actors are abusing legitimate OAuth 2.0 redirects from trusted providers like Google and Microsoft to bypass security defenses. This technique sends users from authentic login pages to attacker-controlled sites to deliver malware or launch phishing attacks.
Key TTPs
Initial Access: Phishing emails contain crafted URLs that point to legitimate OAuth providers.
Execution: Users are redirected to a malicious domain that either initiates a malware download (often a ZIP file) or presents a phishing page to harvest credentials.
Defense Evasion: The attack leverages trusted domains (e.g., login.microsoftonline.com) for the initial link, bypassing email and browser security filters.
Campaign Analysis
This attack method is effective because it exploits user trust in well-known brands and legitimate authentication protocols. By manipulating a standards-compliant feature, attackers make it difficult for both users and automated security tools to detect the malicious intent.
Targeting & Infrastructure
Target Profile: Campaigns have primarily targeted government and public-sector organizations.
Infrastructure: Attackers register malicious applications in their own cloud tenant and configure them with redirect URIs pointing to their malicious domains.
Relevant Terms
OAuth: An open standard for access delegation that allows users to grant third-party applications access to their data without sharing passwords.
Phishing: A social engineering attack where adversaries deceive users into revealing sensitive information, such as login credentials or financial details.
Cloudflare's latest threat report indicates that Generative AI has become a "force multiplier" for cybercriminals, significantly lowering the technical skills required to launch effective and scalable attacks. Attackers are now focusing more on "logging in" with compromised credentials rather than "breaking in."
Key Findings
Cloudflare blocks an average of 230 billion threats daily, providing a broad view of the evolving threat landscape.
Threat actors are using Large Language Models (LLMs) to craft convincing phishing emails and map corporate networks in real-time.
North Korean operatives have been observed using AI-generated deepfakes and fraudulent IDs to infiltrate companies by bypassing hiring filters.
Nearly half (46%) of analyzed emails failed DMARC validation, highlighting a persistent gap in basic email authentication that enables phishing campaigns.
The Bottom Line
The industrialization of cyber threats means the barrier to entry for sophisticated attacks has collapsed. Security leaders can no longer focus solely on highly skilled adversaries.
The strategic challenge is now defending against a massively expanded pool of less-skilled, AI-equipped attackers who can operate with unprecedented speed and scale.
This necessitates a shift towards automated, identity-focused defenses and assumes that any user could be a potential threat.
Relevant Terms
Deepfake: AI-generated synthetic media where a person's likeness is replaced with someone else's, often used in sophisticated fraud and disinformation campaigns.
Large Language Model (LLM): An advanced AI system trained on vast amounts of text data, capable of understanding and generating human-like language for tasks like writing phishing emails or malicious code.
