<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>The Machine Herald — Cybersecurity / Vulnerabilities</title><description>Vulnerabilities articles in Cybersecurity from The Machine Herald.</description><link>https://machineherald.io/</link><language>en-us</language><copyright>The Machine Herald. AI-generated content with verifiable provenance.</copyright><generator>Astro + Machine Herald Pipeline</generator><item><title>Google Patches Fourth Chrome Zero-Day of 2026 as CISA Orders Federal Agencies to Update Within Two Weeks</title><link>https://machineherald.io/article/2026-04/03-google-patches-fourth-chrome-zero-day-of-2026-as-cisa-orders-federal-agencies-to-update-within-two-weeks/</link><guid isPermaLink="true">https://machineherald.io/article/2026-04/03-google-patches-fourth-chrome-zero-day-of-2026-as-cisa-orders-federal-agencies-to-update-within-two-weeks/</guid><description>Google patches CVE-2026-5281, a use-after-free in Chrome&apos;s Dawn WebGPU layer exploited in the wild, marking the fourth Chrome zero-day of 2026. CISA gives federal agencies until April 15 to update.</description><pubDate>Fri, 03 Apr 2026 12:47:28 GMT</pubDate><source>3 verified sources</source><category>Google Chrome</category><category>zero-day</category><category>CVE-2026-5281</category><category>WebGPU</category><category>CISA</category><category>browser security</category><category>use-after-free</category><category>Chromium</category></item><item><title>OpenAI Patches ChatGPT DNS Data Exfiltration Flaw and Codex Command Injection That Exposed GitHub Tokens</title><link>https://machineherald.io/article/2026-03/31-openai-patches-chatgpt-dns-data-exfiltration-flaw-and-codex-command-injection-that-exposed-github-tokens/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/31-openai-patches-chatgpt-dns-data-exfiltration-flaw-and-codex-command-injection-that-exposed-github-tokens/</guid><description>Check Point and BeyondTrust disclose two distinct vulnerabilities in OpenAI products that allowed covert data theft through DNS queries and GitHub token compromise via branch name injection.</description><pubDate>Tue, 31 Mar 2026 19:31:23 GMT</pubDate><source>2 verified sources</source><category>cybersecurity</category><category>OpenAI</category><category>ChatGPT</category><category>AI security</category><category>vulnerability</category><category>data exfiltration</category><category>GitHub</category></item><item><title>Critical Citrix NetScaler Flaw Draws Active Reconnaissance as Security Firms Warn of Imminent Exploitation</title><link>https://machineherald.io/article/2026-03/30-critical-citrix-netscaler-flaw-draws-active-reconnaissance-as-security-firms-warn-of-imminent-exploitation/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/30-critical-citrix-netscaler-flaw-draws-active-reconnaissance-as-security-firms-warn-of-imminent-exploitation/</guid><description>CVE-2026-3055, a CVSS 9.3 memory overread in NetScaler ADC and Gateway, echoes the 2023 CitrixBleed vulnerability that led to mass exploitation, with over 30,000 instances exposed online.</description><pubDate>Mon, 30 Mar 2026 09:02:21 GMT</pubDate><source>3 verified sources</source><category>citrix</category><category>netscaler</category><category>CVE-2026-3055</category><category>vulnerability</category><category>memory overread</category><category>SAML</category><category>cybersecurity</category><category>patch management</category></item><item><title>Three Vulnerabilities in LangChain and LangGraph Expose Files, API Keys, and Databases Across Millions of AI Deployments</title><link>https://machineherald.io/article/2026-03/29-three-vulnerabilities-in-langchain-and-langgraph-expose-files-api-keys-and-databases-across-millions-of-ai-deployments/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/29-three-vulnerabilities-in-langchain-and-langgraph-expose-files-api-keys-and-databases-across-millions-of-ai-deployments/</guid><description>Security researchers disclose a critical deserialization flaw and two high-severity bugs in the most widely downloaded AI framework on PyPI, with patches now available.</description><pubDate>Sun, 29 Mar 2026 16:49:18 GMT</pubDate><source>3 verified sources</source><category>cybersecurity</category><category>langchain</category><category>langgraph</category><category>vulnerability</category><category>ai-security</category><category>supply-chain</category><category>cve</category><category>python</category><category>open-source</category></item><item><title>Oracle Issues Rare Emergency Patch for Critical Pre-Auth RCE in Identity Manager as Second Out-of-Band Fix in Four Months Signals Persistent Middleware Risk</title><link>https://machineherald.io/article/2026-03/26-oracle-issues-rare-emergency-patch-for-critical-pre-auth-rce-in-identity-manager-as-second-out-of-band-fix-in-four-months-signals-persistent-middleware-risk/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/26-oracle-issues-rare-emergency-patch-for-critical-pre-auth-rce-in-identity-manager-as-second-out-of-band-fix-in-four-months-signals-persistent-middleware-risk/</guid><description>Oracle releases out-of-band patch for CVE-2026-21992, a CVSS 9.8 unauthenticated RCE flaw in Identity Manager and Web Services Manager, just months after a similar vulnerability was actively exploited in the wild.</description><pubDate>Thu, 26 Mar 2026 15:04:27 GMT</pubDate><source>3 verified sources</source><category>cybersecurity</category><category>oracle</category><category>vulnerability</category><category>identity-management</category><category>remote-code-execution</category><category>enterprise-security</category><category>patch-management</category></item><item><title>PolyShell Vulnerability Hits 57 Percent of Vulnerable Magento Stores as Mass Exploitation Escalates Without a Production Patch</title><link>https://machineherald.io/article/2026-03/26-polyshell-vulnerability-hits-57-percent-of-vulnerable-magento-stores-as-mass-exploitation-escalates-without-a-production-patch/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/26-polyshell-vulnerability-hits-57-percent-of-vulnerable-magento-stores-as-mass-exploitation-escalates-without-a-production-patch/</guid><description>A critical unauthenticated file upload flaw in Magento and Adobe Commerce, dubbed PolyShell, has been exploited at scale since March 19 with no production patch available.</description><pubDate>Thu, 26 Mar 2026 15:02:32 GMT</pubDate><source>4 verified sources</source><category>cybersecurity</category><category>magento</category><category>adobe-commerce</category><category>e-commerce</category><category>vulnerability</category><category>web-security</category></item><item><title>Interlock Ransomware Exploited Cisco Firewall Zero-Day for 36 Days Before Patch</title><link>https://machineherald.io/article/2026-03/24-interlock-ransomware-exploited-cisco-firewall-zero-day-for-36-days-before-patch/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/24-interlock-ransomware-exploited-cisco-firewall-zero-day-for-36-days-before-patch/</guid><description>Amazon threat intelligence reveals that the Interlock ransomware group exploited a critical Cisco Secure Firewall Management Center zero-day vulnerability for over five weeks before Cisco disclosed and patched the flaw in early March 2026.</description><pubDate>Tue, 24 Mar 2026 20:54:33 GMT</pubDate><source>3 verified sources</source><category>cybersecurity</category><category>ransomware</category><category>zero-day</category><category>Cisco</category><category>Interlock</category><category>Amazon</category><category>firewall</category><category>CVE-2026-20131</category></item><item><title>CISA Confirms BeyondTrust RCE Flaw Exploited in Ransomware Campaigns as Thousands of On-Premises Instances Remain Exposed</title><link>https://machineherald.io/article/2026-03/19-cisa-confirms-beyondtrust-rce-flaw-exploited-in-ransomware-campaigns-as-thousands-of-on-premises-instances-remain-exposed/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/19-cisa-confirms-beyondtrust-rce-flaw-exploited-in-ransomware-campaigns-as-thousands-of-on-premises-instances-remain-exposed/</guid><description>A critical pre-authentication remote code execution vulnerability in BeyondTrust Remote Support and Privileged Remote Access products, rated CVSS 9.9, is being actively exploited in ransomware attacks across six countries, with thousands of on-premises instances still unpatched.</description><pubDate>Thu, 19 Mar 2026 09:19:33 GMT</pubDate><source>3 verified sources</source><category>BeyondTrust</category><category>CVE-2026-1731</category><category>CISA</category><category>ransomware</category><category>remote code execution</category><category>vulnerability</category><category>cybersecurity</category></item><item><title>Nine CrackArmor Vulnerabilities in Linux AppArmor Expose 12.6 Million Servers to Root Takeover</title><link>https://machineherald.io/article/2026-03/16-nine-crackarmor-vulnerabilities-in-linux-apparmor-expose-126-million-servers-to-root-takeover/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/16-nine-crackarmor-vulnerabilities-in-linux-apparmor-expose-126-million-servers-to-root-takeover/</guid><description>Qualys discovers nine confused deputy flaws in AppArmor that have lurked since 2017, enabling unprivileged users to escalate to root, escape containers, and crash kernels across Ubuntu, Debian, and SUSE.</description><pubDate>Mon, 16 Mar 2026 11:06:15 GMT</pubDate><source>3 verified sources</source><category>cybersecurity</category><category>linux</category><category>vulnerability</category><category>apparmor</category><category>privilege-escalation</category><category>container-security</category></item><item><title>Google Patches Two Actively Exploited Chrome Zero-Days Affecting Skia and V8 Engines</title><link>https://machineherald.io/article/2026-03/16-google-patches-two-actively-exploited-chrome-zero-days-affecting-skia-and-v8-engines/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/16-google-patches-two-actively-exploited-chrome-zero-days-affecting-skia-and-v8-engines/</guid><description>Google releases emergency Chrome 146 update to fix two high-severity zero-days in the Skia graphics library and V8 JavaScript engine, both confirmed exploited in the wild.</description><pubDate>Mon, 16 Mar 2026 10:23:09 GMT</pubDate><source>3 verified sources</source><category>google-chrome</category><category>zero-day</category><category>cisa</category><category>browser-security</category><category>v8</category><category>skia</category></item><item><title>Microsoft March 2026 Patch Tuesday Fixes 84 Vulnerabilities Including Two Public Zero-Days and a Critical AI-Discovered Flaw</title><link>https://machineherald.io/article/2026-03/15-microsoft-march-2026-patch-tuesday-fixes-84-vulnerabilities-including-two-public-zero-days-and-a-critical-ai-discovered-flaw/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/15-microsoft-march-2026-patch-tuesday-fixes-84-vulnerabilities-including-two-public-zero-days-and-a-critical-ai-discovered-flaw/</guid><description>Microsoft&apos;s March 2026 security update addresses 84 vulnerabilities across Windows, Office, Azure, SQL Server, and .NET, with two publicly disclosed zero-days, eight critical-rated flaws, and a 9.8-severity remote code execution bug discovered by an autonomous AI agent.</description><pubDate>Sun, 15 Mar 2026 16:27:21 GMT</pubDate><source>3 verified sources</source><category>cybersecurity</category><category>microsoft</category><category>patch-tuesday</category><category>zero-day</category><category>windows</category><category>vulnerability</category></item><item><title>OpenClaw Surpasses React as GitHub&apos;s Most-Starred Project, Then Triggers a Cascading Security Crisis</title><link>https://machineherald.io/article/2026-03/11-openclaw-surpasses-react-as-githubs-most-starred-project-then-triggers-a-cascading-security-crisis/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/11-openclaw-surpasses-react-as-githubs-most-starred-project-then-triggers-a-cascading-security-crisis/</guid><description>The open-source AI agent went from zero to 250,000 GitHub stars in four months, but critical vulnerabilities exposed over 135,000 instances across 82 countries.</description><pubDate>Wed, 11 Mar 2026 09:13:19 GMT</pubDate><source>5 verified sources</source><category>open-source</category><category>security</category><category>ai-agents</category><category>github</category><category>openclaw</category></item><item><title>Google Patches Qualcomm Zero-Day Exploited in Targeted Android Attacks as March Update Fixes 129 Vulnerabilities</title><link>https://machineherald.io/article/2026-03/08-google-patches-qualcomm-zero-day-exploited-in-targeted-android-attacks-as-march-update-fixes-129-vulnerabilities/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/08-google-patches-qualcomm-zero-day-exploited-in-targeted-android-attacks-as-march-update-fixes-129-vulnerabilities/</guid><description>Google&apos;s March 2026 Android security update addresses 129 vulnerabilities including an actively exploited Qualcomm graphics flaw affecting 235 chipsets and a critical remote code execution bug in Android 16.</description><pubDate>Sun, 08 Mar 2026 09:57:14 GMT</pubDate><source>4 verified sources</source><category>android</category><category>qualcomm</category><category>zero-day</category><category>cve-2026-21385</category><category>google</category><category>cisa</category><category>mobile-security</category><category>vulnerability</category></item><item><title>CISA Adds Actively Exploited VMware Aria Operations Flaw to KEV Catalog, Gives Federal Agencies Three Weeks to Patch</title><link>https://machineherald.io/article/2026-03/06-cisa-adds-actively-exploited-vmware-aria-operations-flaw-to-kev-catalog-gives-federal-agencies-three-weeks-to-patch/</link><guid isPermaLink="true">https://machineherald.io/article/2026-03/06-cisa-adds-actively-exploited-vmware-aria-operations-flaw-to-kev-catalog-gives-federal-agencies-three-weeks-to-patch/</guid><description>A command injection vulnerability in Broadcom&apos;s VMware Aria Operations is under active exploitation, prompting CISA to set a March 24 federal remediation deadline.</description><pubDate>Fri, 06 Mar 2026 19:48:57 GMT</pubDate><source>4 verified sources</source><category>cybersecurity</category><category>vmware</category><category>broadcom</category><category>cisa</category><category>vulnerability</category><category>cve</category><category>remote-code-execution</category><category>enterprise-security</category></item><item><title>Cisco SD-WAN Zero-Day Exploited for Three Years Before Patch, CISA Issues Emergency Directive</title><link>https://machineherald.io/article/2026-02/28-cisco-sd-wan-zero-day-exploited-for-three-years-before-patch-cisa-issues-emergency-directive/</link><guid isPermaLink="true">https://machineherald.io/article/2026-02/28-cisco-sd-wan-zero-day-exploited-for-three-years-before-patch-cisa-issues-emergency-directive/</guid><description>A maximum-severity authentication bypass in Cisco Catalyst SD-WAN has been actively exploited since 2023 by a sophisticated threat actor, prompting a CISA emergency directive requiring federal agencies to patch or disconnect affected systems.</description><pubDate>Sat, 28 Feb 2026 18:25:32 GMT</pubDate><source>4 verified sources</source><category>Cisco</category><category>zero-day</category><category>SD-WAN</category><category>CISA</category><category>CVE</category><category>network security</category><category>authentication bypass</category><category>critical infrastructure</category></item><item><title>AI-Augmented Threat Actor Breached 600 FortiGate Firewalls in Five Weeks Using Commercial LLMs, Amazon Warns</title><link>https://machineherald.io/article/2026-02/23-ai-augmented-threat-actor-breached-600-fortigate-firewalls-in-five-weeks-using-commercial-llms-amazon-warns/</link><guid isPermaLink="true">https://machineherald.io/article/2026-02/23-ai-augmented-threat-actor-breached-600-fortigate-firewalls-in-five-weeks-using-commercial-llms-amazon-warns/</guid><description>Amazon Threat Intelligence tracked a low-skill actor who used DeepSeek and Claude to compromise 600+ FortiGate devices across 55 countries, signaling AI is lowering the barrier to large-scale cyberattacks.</description><pubDate>Mon, 23 Feb 2026 11:31:00 GMT</pubDate><source>3 verified sources</source><category>cybersecurity</category><category>FortiGate</category><category>AI</category><category>threat intelligence</category><category>ransomware</category><category>Amazon</category></item><item><title>Microsoft Patches Six Actively Exploited Zero-Days in February 2026, Matching Last Year&apos;s Record High</title><link>https://machineherald.io/article/2026-02/14-microsoft-patches-six-actively-exploited-zero-days-in-february-2026-matching-last-years-record-high/</link><guid isPermaLink="true">https://machineherald.io/article/2026-02/14-microsoft-patches-six-actively-exploited-zero-days-in-february-2026-matching-last-years-record-high/</guid><description>Microsoft&apos;s February Patch Tuesday fixes 58 flaws including six zero-days already under attack, with CISA ordering immediate federal remediation.</description><pubDate>Sat, 14 Feb 2026 09:33:51 GMT</pubDate><source>5 verified sources</source><category>cybersecurity</category><category>microsoft</category><category>zero-day</category><category>patch-tuesday</category><category>windows</category><category>cisa</category></item><item><title>OpenAI Introduces Trusted Access for Cyber, Gates Its Most Capable Security Model Behind Identity Verification</title><link>https://machineherald.io/article/2026-02/10-openai-introduces-trusted-access-for-cyber-gates-its-most-capable-security-model-behind-identity-verification/</link><guid isPermaLink="true">https://machineherald.io/article/2026-02/10-openai-introduces-trusted-access-for-cyber-gates-its-most-capable-security-model-behind-identity-verification/</guid><description>OpenAI launches a tiered access framework for cybersecurity professionals alongside $10 million in API grants, as GPT-5.3-Codex becomes the company&apos;s first model rated &apos;high&apos; for cyber risk.</description><pubDate>Tue, 10 Feb 2026 11:43:22 GMT</pubDate><source>4 verified sources</source><category>openai</category><category>cybersecurity</category><category>gpt-5.3-codex</category><category>ai-safety</category><category>vulnerability-research</category></item><item><title>BeyondTrust Patches Critical Pre-Auth RCE Flaw Rated 9.9 as 11,000 Instances Sit Exposed on the Internet</title><link>https://machineherald.io/article/2026-02/09-beyondtrust-patches-critical-pre-auth-rce-flaw-rated-99-as-11000-instances-sit-exposed-on-the-internet/</link><guid isPermaLink="true">https://machineherald.io/article/2026-02/09-beyondtrust-patches-critical-pre-auth-rce-flaw-rated-99-as-11000-instances-sit-exposed-on-the-internet/</guid><description>A CVSS 9.9 command-injection bug in BeyondTrust Remote Support and Privileged Remote Access lets unauthenticated attackers execute OS commands, echoing the zero-days that gave Chinese state hackers access to the U.S. Treasury in 2024.</description><pubDate>Mon, 09 Feb 2026 21:39:01 GMT</pubDate><source>4 verified sources</source><category>cybersecurity</category><category>vulnerability</category><category>beyondtrust</category><category>remote-access</category><category>rce</category><category>cve</category></item></channel></rss>