Why Should I Care? β€” 2026-10-28 | πŸ”΄ 3 HIGH Β· 🟑 0 MEDIUM Β· πŸ”΅ 25 RADAR Β· βšͺ 187 FILTERED

πŸ“‹ Briefing β€” 2026-10-28

28 vendor intel items scanned  |  πŸ”΄ 3 HIGH  |  🟑 0 MEDIUM  |  πŸ”΅ 25 RADAR  |  βšͺ 187 FILTERED

πŸ”΄ Critical β€” action required:

  1. CISA Adds Two Known Exploited Vulnerabilities to Catalog (CVE-2026-48356, CVE-2026-111497) β€” Yes, if you run Adobe Commerce/Magento or use Google Brotli: these are actively exploited vulnerabilities that allow remote code execution or denial of service, requiring immediate patching.
  2. Mitsubishi Electric GENESIS, GENESIS64, ICONICS Suite, Hyper Historian, AnalytiX, MobileHMI, IoTWorX, MC Works64, GENESIS32 and BizViz (CVE-2023-4863, CVE-2026-16436, CVE-2026-16437, CVE-2026-16438) β€” Yes, if you run Mitsubishi Electric GENESIS64, ICONICS Suite, or related HMI/SCADA software: an attacker with local admin access can execute arbitrary code via a crafted file, and unauthenticated remote attackers can bypass authentication to tamper with or destroy industrial process data.
  3. EPICS Base (CVE-2026-114595) β€” Yes, if you run EPICS Base version 7.0.10 or earlier: unauthenticated remote attackers can crash your control system or potentially execute arbitrary code via crafted network messages.

Everything else can wait.

πŸ”΅ 15 items on the radar β€” see below ↓


πŸ”΄ HIGH β€” Handle Now


CISA Adds Two Known Exploited Vulnerabilities to Catalog

CISA Advisories [CISA KEV] | CVE-2026-48356, CVE-2026-111497

❓ Why Should I Care?
Yes, if you run Adobe Commerce/Magento or use Google Brotli: these are actively exploited vulnerabilities that allow remote code execution or denial of service, requiring immediate patching.

🎯 Affected versions: Adobe Commerce and Magento (unrestricted file upload), Google Brotli (unchecked return value)

🎭 In plain English:
These are two separate bugs: one lets attackers upload malicious files to your Adobe/Magento store to take over the server, and the other causes Google's Brotli compression library to crash or behave unpredictably. An attacker could use the Adobe bug to install malware on your e-commerce site, stealing customer data or defacing the site, or use the Brotli bug to crash services relying on it.

πŸ”§ Prerequisites:

  • Running a vulnerable version of Adobe Commerce or Magento
  • Running a vulnerable version of Google Brotli
  • Internet exposure of the affected services

⏱ Urgency: Critical: CISA has confirmed active exploitation in the wild, meaning attackers are currently using these flaws to compromise systems.

πŸ’‘ Context: The Adobe vulnerability is an unrestricted file upload flaw, while the Brotli issue is an unchecked return value leading to potential memory corruption or crashes.


Mitsubishi Electric GENESIS, GENESIS64, ICONICS Suite, Hyper Historian, AnalytiX, MobileHMI, IoTWorX, MC Works64, GENESIS32 and BizViz

CISA Advisories [CISA KEV] | CVSS 9.1 | CVE-2023-4863, CVE-2026-16436, CVE-2026-16437, CVE-2026-16438

❓ Why Should I Care?
Yes, if you run Mitsubishi Electric GENESIS64, ICONICS Suite, or related HMI/SCADA software: an attacker with local admin access can execute arbitrary code via a crafted file, and unauthenticated remote attackers can bypass authentication to tamper with or destroy industrial process data.

🎯 Affected versions: GENESIS64 10.97 through 10.97.2, ICONICS Suite 10.97 through 10.97.2, GENESIS <=11.03, GENESIS64 <=10.97.3, ICONICS Suite <=10.97.3, Hyper Historian <=10.97.3, AnalytiX <=10.97.3, MobileHMI <=10.97.3, IoTWorX <=10.97.3, MC Works64 all versions, GENESIS32 all versions, BizViz all versions
Not affected: GENESIS64 and ICONICS Suite versions 10.97.3 or later (for CVE-2023-4863); Note: CVE-2026-16436/16437/16438 affect GENESIS64/ICONICS Suite <=10.97.3, implying 10.97.4+ may be fixed for those, but advisory explicitly lists <=10.97.3 as affected for the new CVEs.

🎭 In plain English:
These vulnerabilities allow an attacker to either crash the system, steal sensitive industrial data, or manipulate process controls (like opening/closing valves) without proper authorization. For example, a malicious actor could upload a specially crafted file to the HMI server, triggering a memory error that lets them run their own code, effectively taking over the control system.

πŸ”§ Prerequisites:

  • Local administrator access to the host running GENESIS64/ICONICS Suite for CVE-2023-4863
  • Network access to the HMI/SCADA server for CVE-2026-16436/16437/16438
  • Ability to import a crafted file (for CVE-2023-4863)
  • Ability to interact with the authentication mechanism (for CVE-2026-16436/16437)

⏱ Urgency: Immediate patching is required for any internet-facing or network-accessible HMI/SCADA systems due to the combination of remote authentication bypass and local code execution risks.

βœ… Fixed in: 10.97.3

πŸ’‘ Context: The root cause for CVE-2023-4863 is an out-of-bounds write in the libwebp library, while the 2026 CVEs stem from weak password recovery mechanisms and incorrect privilege assignment in the application logic.


EPICS Base

CISA Advisories | CVSS 9.8 | CVE-2026-114595

❓ Why Should I Care?
Yes, if you run EPICS Base version 7.0.10 or earlier: unauthenticated remote attackers can crash your control system or potentially execute arbitrary code via crafted network messages.

🎯 Affected versions: EPICS Base release <=7.0.10
Not affected: EPICS Base release 7.0.11 and later

🎭 In plain English:
The software that manages your industrial equipment has a flaw in how it handles incoming network data. An attacker on the network can send a specially crafted message that causes the software to crash (denial of service) or, in the worst case, run malicious commands on the server. For example, an attacker could send a malformed packet to the EPICS channel access port, causing the control system to freeze or allowing them to install malware.

πŸ”§ Prerequisites:

  • Running EPICS Base version 7.0.10 or earlier
  • Network access to the EPICS Base services (e.g., Channel Access)
  • No authentication required for exploitation

⏱ Urgency: Immediate patching is required because the vulnerability allows unauthenticated remote code execution with a critical CVSS score of 9.8.

βœ… Fixed in: 7.0.11

πŸ’‘ Context: The root cause is insufficient validation of integer values and length parameters in network message processing, leading to memory corruption.


🟑 MEDIUM (0)

None.


πŸ”΅ On the Radar (25)

  • Apple Backports Fix for Exploited CoreGraphics Flaw to Older iPhones and iPads (The Hacker News) β€” Apple has patched a high-severity flaw (CVE-2026-86950) in CoreGraphics that allows attackers to execute arbitrary code by tricking a device into processing a malicious file. This bug is being used in real-world, sophisticated attacks against specific individuals. If you have devices running iOS 18, iPadOS 18, macOS 15, or macOS 26, they are currently vulnerable unless updated to the latest backported versions (iOS 18.7.11, iPadOS 18.7.11, macOS 15.8.1, or macOS 26.7.1).
  • CISA Confirms Active Exploitation of Two PHP2Shell Flaws Enabling Remote Code Execution (The Hacker News) β€” CISA has confirmed that attackers are successfully using two critical bugs in PHP to take over servers without needing a login. This is not a theoretical risk; it is happening now. If your web server runs an older version of PHP, an attacker can send a specific request that crashes the memory management in a way that lets them run any command they want on your system, potentially leading to data theft, ransomware, or botnet recruitment.
  • Hackers hijack Convertr analytics script to steal payment card data (BleepingComputer) β€” Hackers compromised a shared analytics script (Plausible) on Convertr's infrastructure to inject a skimmer that steals credit card numbers, CVVs, and CPF IDs during checkout. This is a supply-chain attack affecting dozens of major Brazilian fashion brands. If you or your customers used Convertr checkout in the last few days, assume the card data is compromised.
  • UNM Health ransomware attack disrupts patient care, surgeries (BleepingComputer) β€” A ransomware group called Byod attacked UNM Health, causing outages that forced the hospital to divert ambulances and cancel surgeries. The attackers claim to have compromised 36,000 systems, including Active Directory and the Microsoft Entra tenant, and are threatening to leak patient and employee data. Even though UNM says core systems were not penetrated, the incident highlights how quickly ransomware can disrupt life-critical services and expose sensitive records.
  • Japan's GMO urges customers to shut down PCs amid ransomware threat (BleepingComputer) β€” A threat actor breached GMO's Onamae Desktop Cloud, potentially installing malware on 26,739 customer servers. The attacker stole data, demanded $50 million, and threatened to trigger ransomware. GMO confirmed data was leaked but stated ransomware was not activated. If you run trading bots (MT4/MT5) or services on this platform, your credentials and data are compromised.
  • Warp discloses breach exposing AI agent logs, user telemetry data (BleepingComputer) β€” A sophisticated attack on a third-party vendor used by Warp exposed the 'brain' of AI agents for roughly 1% of users. This means attackers may have seen what your AI agents were doing, including the code they generated and the tasks they performed, which can leak proprietary logic or internal system structures.
  • Microsoft releases Windows OOB updates to distrust eMudhra certificates (BleepingComputer) β€” Microsoft is removing trust in specific eMudhra root certificates because eight intermediate CA keys were generated with extremely weak values (under 256), allowing anyone to brute-force them and forge valid certificates. This means any system still trusting these roots could be tricked into accepting fraudulent TLS connections or code signatures. For example, an attacker could intercept HTTPS traffic to a legitimate site by presenting a forged certificate signed by one of these compromised intermediates, which would otherwise appear valid to your browser and applications.
  • Lazarus Abuses South Korean Security Software to Deploy Malware Using Post-Quantum Encryption (The Hacker News) β€” Lazarus Group is actively hacking South Korean companies by exploiting vulnerabilities in mandatory banking and government security software (CrossWeb EX). They are using advanced techniques like in-memory execution and post-quantum encryption to hide their tracks and steal data, which means standard antivirus and network monitoring may miss the initial compromise.
  • Keystone to Destroy 18,835 Hardware Wallets After Hackers Steal Device Verification Key (The Hacker News) β€” Hackers stole the secret key Keystone uses to prove its hardware wallets are genuine. This means counterfeit devices could now pass the official verification check, potentially allowing attackers to steal crypto from users who trust the fake device. While your private keys and recovery phrases were not stolen, the integrity of the device itself is compromised for units shipped after October 20. For infrastructure professionals, this is a supply-chain and trust-anchor failure: the mechanism designed to prevent counterfeit hardware has been broken, forcing a mass destruction of unsold stock and creating a window for social engineering attacks.
  • Microsoft Authenticator Flaw Let Permissionless Android Apps Steal Work Account Tokens (The Hacker News) β€” A flaw in the Android version of Microsoft Authenticator allowed any app on the device to impersonate trusted Microsoft apps like Teams and request sign-in tokens. If successful, the attacker gains full access to the user's work email, OneDrive files, and calendar. This is not a remote exploit; it requires a malicious app to be installed on the phone, but the impact is total account compromise.
  • Tailscale warns of SSH flaws allowing root privilege escalation (BleepingComputer) β€” Tailscale has disclosed 11 critical vulnerabilities (Tailworm) that allow attackers with valid access to escalate to root privileges, steal cloud credentials, or crash the control server. For example, an attacker with a standard SSH session could use socket forwarding to execute arbitrary commands as root on your servers, bypassing local user restrictions.
  • Attackers Hijack Dahua IP Cameras in Ukraine Using P2P Access and Auth Weaknesses (The Hacker News) β€” Attackers are hijacking Dahua cameras by guessing their serial numbers (which are sequential and predictable) to access them through the cloud P2P relay, bypassing the need for direct internet exposure. Once in, they use known authentication bypass flaws (CVE-2021-33044/33045) to create persistent backdoor accounts that survive factory resets. For example, a warehouse with Dahua cameras for inventory tracking could have its video feed stolen and its configuration altered to hide evidence or track personnel, with the attacker retaining access even after a password change.
  • GoAhead TLS Flaws Allow Revoked Client Certificates and Enable Session Decryption (The Hacker News) β€” GoAhead, a web server embedded in hundreds of millions of devices, had a bug where it loaded a list of banned (revoked) certificates but never actually checked it. This means a stolen or retired device with a revoked certificate could still log in and access the management interface as if it were valid. Additionally, a separate flaw in some builds allowed passive attackers to decrypt session traffic. The fix is in version 6.1.1, but you must wait for your device manufacturer to ship a firmware update.
  • CyberStrikeAI Uses 44 AI Agents to Automate Hacking Operations Against Japan Targets (The Hacker News) β€” Threat actors are deploying multi-agent AI frameworks (like CyberStrikeAI) to automate reconnaissance, exploitation, and data exfiltration at scale. This means attackers can test thousands of targets simultaneously, adapt to local languages (e.g., Japanese phishing forms), and maintain persistence without human fatigue. For example, a single campaign compromised 19 organizations in days by automating vulnerability validation and C2 setup, turning what used to be a weeks-long manual process into a rapid, industrialized attack.
  • OpenAI's GPT-6.1 Terra is safer overall, but shows more destructive-action flags in tests (BleepingComputer) β€” General news about AI model safety testing results.

βšͺ 187 low-priority items filtered.


πŸ¦… Aggregated and triaged by Donna AI  |  Sources: 8 vendor feeds  |  CISA KEV

Subscribe to Why Should I Care?

Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.
jamie@example.com
Subscribe