DanSec

Cybersecurity Brief

Friday 23 October 2026

In this brief Capcom, LiteLLM, Arista, VeloCloud, Railway, GitHub, +16 more
Vendors
Capcom, LiteLLM, Arista, VeloCloud, Railway, GitHub, Quest, State Bank of India, PHP, Pillow, Node.js, Linux, Atlassian, ConnectWise, Meta, eMudhra
Threat actors
Mirai, Sandworm, Storm-0940, Prince of Persia
CVEs
CVE-2026-115560, CVE-2026-112900

Major Incidents or Breaches

  • Capcom confirmed a data breach affecting over 11.7 million CAPCOM ID accounts, exposing customer email addresses and associated information. Attackers also changed the login passwords for 196 accounts [17][32].
  • Cloud platform Railway issued an urgent warning to users to rotate secrets after detecting suspicious activity on systems storing customer environment variables, indicating a potential breach [16].
  • The LiteLLM project’s PyPI package and Docker images were compromised in a supply chain attack. Hackers used stolen GitHub CLI OAuth tokens to upload and approve malicious packages containing info-stealer malware targeting credentials, secrets, and environment variables [18][34].
  • Quest Diagnostics reported unauthorized access to its Quanum Lab Services Manager, affecting certain patient records [33].
  • A Crime Stoppers chapter suspended tip submissions after a breach exposed information about anonymous tipsters, breaking the promise of anonymity [39].
  • State Bank of India suffered a customer PII data leak, with concerns about the potential for misuse in the underground market [24].

Newly Discovered Vulnerabilities

  • PHP disclosed a new unpatched remote code execution vulnerability (CVE-2026-115560) two days after releasing emergency PHP2Shell patches. Attackers began exploiting PHP2Shell flaws four days prior to patch release, with millions of exploit attempts observed, primarily aiming to deploy Mirai malware variants [3][8][22][12].
  • Pillow, the widely used Python image library, was found to have a critical code execution flaw (CVE-2026-112900) that allows arbitrary OS command execution via crafted image files [5].
  • Researchers demonstrated a remote code execution exploit (“Node2Shell”) on Node.js using malicious Brotli data, allowing attacker-supplied code execution on official builds [6][12].
  • Two CVSS 9.0-rated remote code execution flaws in Linux KSMBD were publicly disclosed before kernel fixes became available, exposing Linux servers running the built-in SMB server [13][12].
  • Arista Networks patched a critical zero-day SQL injection vulnerability in VeloCloud Orchestrator (VCO) that was being actively exploited to gain operator privileges. This is the third VCO zero-day exploited in four months [10][19][31].
  • Atlassian released patches for 75 vulnerabilities in Data Center and Server products, including 14 critical-severity issues in third-party components [35].
  • A critical flaw in the x402 SDK could allow malicious APIs to drain the full token balances of AI agents by manipulating payment requests [7].

Notable Threat Actor Activity

  • The Russian state-sponsored group Sandworm, in collaboration with another actor, carried out a destructive wiping attack on a Ukrainian logistics company using a new mesh network backdoor [4].
  • Microsoft attributed a supply chain attack on ConnectWise to the China-linked group Storm-0940. The attackers used an AI-enabled penetration-testing tool to automate attacks on over 2,200 tenants, gaining privileged access in some environments. The tool, Artex, was used to accelerate exploitation [25][30].
  • Iran’s Prince of Persia APT hosted a command and control server in the US for the first time since 2016, following a three-month operational lull [36].
  • Threat actors attempted to exploit PHP2Shell vulnerabilities to deploy the Potassium variant of the Mirai botnet, though these attempts were largely unsuccessful [8].
  • Hackers compromised the LiteLLM release pipeline via stolen GitHub tokens to distribute info-stealer malware, highlighting ongoing risks in the software supply chain [18][34].
  • AI agents are increasingly being identified as insider threats. They often store sensitive credentials and operate autonomously on enterprise networks, raising concerns about unmanaged access and risk [9][23].
  • Attackers are leveraging AI-enabled tools to automate penetration testing and accelerate the exploitation process, as observed in recent supply chain attacks [25][30].
  • Phishing campaigns are targeting job seekers with spoofed Calendly pages to steal Facebook and Google credentials. The use of AI enables attackers to scale these campaigns through automated job application processes [1][2].
  • Machine-speed attacks, such as those involving stolen GitHub tokens and malicious AI integrations, can result in credential theft and data exfiltration within minutes, outpacing traditional detection and response capabilities [21].
  • Hundreds of EU supermarkets suffered significant losses due to outdated OT firmware in refrigeration systems managed by third parties, underscoring risks associated with unpatched industrial control systems [27].
  • Insurers are divided on covering AI-related risks, with some offering specialized agentic AI coverage and others excluding it, leaving enterprises potentially exposed [28].

Regulatory or Policy Developments Affecting the Security Industry

  • Meta was ordered by a New Mexico court to pay $25 billion in penalties for 43 million privacy violations related to misleading users about Facebook’s data privacy practices in the wake of the Cambridge Analytica scandal [14][40].
  • eMudhra, an Indian certificate authority, will exit the global SSL/TLS certificate business after researchers found eight intermediate CA keys with easily guessable values, raising concerns about certificate integrity [11][12].

Sources

  1. Capcom confirms data breach affecting 11.7 million accounts - bleepingcomputer.com
  2. 11 Million Impacted by Capcom Data Breach - securityweek.com
  3. Railway warns devs to rotate secrets after potential breach - bleepingcomputer.com
  4. LiteLLM release pipeline hacked to distribute info-stealer malware - bleepingcomputer.com
  5. LiteLLM Compromised in Fresh Supply Chain Attack - securityweek.com
  6. Quest Diagnostics Reports Cyber Breach Affecting ‘Certain Patient Records’ - securityweek.com
  7. Crime Stoppers chapter freezes tips after breach breaks anonymity promise - malwarebytes.com
  8. State Bank of India Suffers a Customer PII Data Leak - darkreading.com
  9. URGENT - PHP Warns of Another Unpatched RCE Two Days After Emergency PHP2Shell Patches - thehackernews.com
  10. PHP2Shell Exploit Attempts Hit Millions as Attackers Try to Deploy Mirai Malware - thehackernews.com
  11. Amazon says PHP2Shell exploitation began 4 days before patches - bleepingcomputer.com
  12. ThreatsDay: AI Secrets Leak, PingCAP Source Theft, Signal Hang-Up Flaw, Laravel Auth Bypass + 21 More - thehackernews.com
  13. Critical Flaw in Python’s Popular Pillow Image Library Allows Code Execution via a Crafted Image - thehackernews.com
  14. Researchers Demonstrate Node2Shell RCE Exploit on Node.js Build Using Crafted Brotli Data - thehackernews.com
  15. Two CVSS 9.0 Linux KSMBD RCE Flaws Disclosed on Public Mailing Lists Before Kernel Fixes - thehackernews.com
  16. Arista Warns of Critical VeloCloud SQL Injection Zero-Day Exploited in Attacks - thehackernews.com
  17. Arista patches new VeloCloud Orchestrator zero-day exploited in attacks - bleepingcomputer.com
  18. Attackers Exploit Third VeloCloud Orchestrator Zero-Day in Four Months - securityweek.com
  19. Atlassian Patches 75 Flaws in Data Center and Server Products - securityweek.com
  20. Critical x402 SDK Flaw Could Let Malicious APIs Drain AI Agents’ Token Balances - thehackernews.com
  21. Russian-Linked Hackers Used New Mesh Network Backdoor in Destructive Ukraine Attack - thehackernews.com
  22. Microsoft Links Storm-0940 to ConnectWise Supply Chain Attack - darkreading.com
  23. China APT Uses AI Pentest Tool Artex in ConnectWise Attacks - securityweek.com
  24. Iran’s Prince of Persia APT Hosts C&C in US for First Time Since 2016 - securityweek.com
  25. The AI Agent Nobody Owns: Watch Expert Webinar On Your Newest Insider Threat - thehackernews.com
  26. AI Agents as Insider Threats: Inside Microsoft's Bet on a Unified Security Stack - darkreading.com
  27. Job Seekers Who Use AI for Applications Need to be on the Lookout for Scams - blog.knowbe4.com
  28. Warning: Calendly Phishing Pages Are Targeting Job Seekers - blog.knowbe4.com
  29. Machine speed attacks require machine speed detection and response - bleepingcomputer.com
  30. Les supermarchés de l'UE paient d'un firmware de réfrigérateur non mis à jour - darkreading.com
  31. As AI Risks Grow, Insurers Split on Covering the Damage - darkreading.com
  32. Meta Ordered to Pay $25 Billion After Court Finds 43 Million Privacy Violations - thehackernews.com
  33. Facebook’s misleading privacy claims land Meta a $25 billion penalty - malwarebytes.com
  34. eMudhra to Exit Global SSL/TLS Business After 8 Intermediate CA Keys Found Easily Guessable - thehackernews.com