Miasma Worm Hits 73 Microsoft GitHub Repos via AI Coding Agents

Microsoft Defender Zero-Day Exploited, Apple AI Patches & Insurance Mega-Breaches

(00:00:00) Microsoft Defender Zero-Day Exploited, Apple AI Patches & Insurance Mega-Breaches (00:01:08) Malicious Perplexity Chrome Extension (00:01:55) Apple WebKit Patches and AI Bug Discovery (00:02:37) FUXA SCADA Authentication Bypass (00:03:18) Insurance Sector Breaches: NAIC and Aflac (00:04:07) Watchpoints for the Next Twenty-Four Hours Ransomware operators are actively exploiting CVE-2026-33825, a Microsoft Defender privilege escalation flaw that enables SYSTEM-level access on unpatched Windows endpoints. CISA has added it to the Known Exploited Vulnerabilities catalog, confirming real-world attacks are underway. If your organization hasn't applied the April 14th patch cycle, the risk window is open right now. Also in today's briefing: Apple pushed updates across iOS, macOS, and Safari addressing more than thirty vulnerabilities — four WebKit flaws, including CVE-2026-43707, were discovered using AI tools from Anthropic and OpenAI, signalling that AI-assisted vulnerability research is now a mainstream part of the patch cycle on both sides of the security divide. Microsoft identified a malicious Chrome extension impersonating Perplexity AI that silently routed search queries and browsing behavior to an attacker-controlled server. The Chrome Web Store missed it. The incident highlights a persistent and widening gap in browser extension vetting, especially for AI-branded tools. CISA issued its first critical advisory for the open-source FUXA SCADA and HMI platform, covering an authentication bypass flaw — CVE-2026-13207, CVSS 8.6 — affecting manufacturing, energy, and water treatment environments. Patch 1.3.2 is available. Finally, two insurance-sector breaches surfaced within 72 hours: Aflac Life Insurance Japan confirmed 4.38 million records compromised, including 230,000 bank account numbers, while ShinyHunters published 3.1 terabytes of data from the National Association of Insurance Commissioners via a PeopleSoft zero-day. The vendor patch timeline remains unresolved. This podcast was built using AI technology. A YesWee production. This episode includes AI-generated content.

PoC Exploits, Anonymous Dump & Tata iPhone IP Leak

(00:00:00) PoC Exploits, Anonymous Dump & Tata iPhone IP Leak (00:01:14) Anonymous Exploit Dump — 15 Products (00:02:00) PTC Windchill KEV Listing (00:02:29) Tata Electronics Breach — iPhone 18 Pro IP (00:03:03) Weedhack and CountLoader — Malware at Scale (00:03:45) Amazon Q Developer Credential Risk (00:04:09) Key Watchpoints — What Comes Next A proof-of-concept exploit for CVE-2026-55200 — a CVSS 9.2 integer overflow in libssh2 — is now public, and the attack surface is enormous. Because libssh2 is statically linked into curl, Git, PHP, firmware updaters, and embedded appliances, distro patches won't reach most affected deployments. The same class of bug hit libssh2 in 2019. Seven years later, the exposure is wider than ever. A researcher known as "bikini" compounded the problem by dropping an unvetted exploit archive targeting 15 products — including Gitea, Splunk, RustDesk, VLC, and OpenVPN — with zero vendor notice. Two entries are confirmed high-impact: libssh2 and Gitea (CVE-2026-20896), the latter already exploited in the wild. The coordinated disclosure model is under pressure. CISA added CVE-2026-12569 in PTC Windchill to its Known Exploited Vulnerabilities catalog. The unauthenticated RCE flaw, used to deploy JSP webshells, has had a patch available since June 18 — making the exploitation gap the headline, not the vulnerability itself. The World Leaks ransomware group leaked over 200,000 files from Tata Electronics, including component maps, supplier data, and prototype photographs tied to the iPhone 18 Pro. Apple-specific IP is confirmed on the dark web, with potential overlap into TSMC and Qualcomm files. Also covered: Weedhack malware-as-a-service targeting Minecraft players across 116,000 endpoints, the CountLoader JavaScript campaign infecting 86,000 devices across three continents, and CVE-2026-12957 in Amazon Q Developer — a supply chain risk that can exfiltrate cloud credentials from untrusted repositories. This episode includes AI-generated content.

Tata-Apple IP Theft, Stryker Wiper & Cisco Unified CM Zero-Day

(00:00:00) Tata-Apple IP Theft, Stryker Wiper & Cisco Unified CM Zero-Day (00:01:08) Iranian Wiper Malware, Stryker Hit (00:01:55) Cisco Unified CM Zero-Day Exploited (00:02:21) Telus, LastPass, and OAuth Chain Risk (00:03:11) Patch Wave and FortiGate Exposure (00:03:45) What to Watch Next Six hundred and thirty gigabytes of Apple manufacturing data — engineering schematics, process documentation, and fifty thousand employee records — is now in attacker hands after a breach at Tata Electronics, Apple's primary manufacturing partner in India. The vector was an unpatched VPN vulnerability. This is intellectual property theft at the core of Apple's hardware supply chain, and it carries regulatory exposure under India's data protection framework with fines of up to four percent of annual turnover. The Stryker breach takes a different shape entirely. Handala, a hacktivist group linked to Iranian state-aligned actors, deployed wiper malware against the medical device company, claiming fifty terabytes exfiltrated and reportedly shutting down offices across seventy-nine countries. Wiper attacks don't offer a recovery payment path — they destroy. The downstream risk to healthcare systems is real. On the vulnerability front, CVE-2026-20230, an SSRF flaw in Cisco Unified Communications Manager, is being actively exploited in the wild to achieve remote code execution via webshell deployment. If you're running Unified CM unpatched, that is the immediate priority. Elsewhere, ShinyHunters claims nearly one petabyte stolen from Telus Digital with a sixty-five million dollar ransom attached, while a Klue supply chain breach enabled attackers to pivot through OAuth tokens into LastPass customer data held in Salesforce — a textbook third-party SaaS trust-chain attack. The patch wave this cycle is heavy: emergency RCE fixes for Nginx, a PostgreSQL privilege escalation, and the FortiGate Fortibleed credential exposure all demand immediate action. The common thread across this entire cycle is vendor infrastructure as the primary attack surface. This episode includes AI-generated content.

Klue's Double Extortion, Dialog Leak & $10M US Breach Costs

(00:00:00) Klue's Double Extortion, Dialog Leak & $10M US Breach Costs (00:00:46) Icarus Gets Hit Back (00:01:37) Dialog Misconfiguration, Not Crime (00:02:17) US Breach Costs Hit $10.22 Million (00:03:01) The $1.9 Million AI Security Divide (00:03:27) Third-Party Risk Now Systemic A supply chain attack on market intelligence platform Klue has exposed roughly 195 enterprise customers after attackers stole OAuth tokens tied to Salesforce, Gong, Deel, and other integrations — bypassing MFA entirely. In a rare twist, the original threat actor, Icarus, was itself compromised by a second criminal group, leaving victims navigating simultaneous extortion demands from two separate actors over the same stolen dataset. Meanwhile, a data exposure at the Dialog Group — a private network linked to Peter Thiel — turned out to stem from a website misconfiguration rather than criminal intrusion. The practical outcome was the same: member records, including details linked to a White House intelligence official and a special operations officer, were publicly accessible to anyone who looked. New IBM Cost of a Data Breach data sharpens the financial picture. The average US breach now costs $10.22 million — an all-time high and more than double the global average of $4.44 million. The US recorded 3,322 breaches in 2024, driven by a complex regulatory environment spanning fifty-state notification laws, HIPAA, and SEC disclosure requirements. Two metrics stand out for security leaders. Organizations using AI and automation in security operations saved $1.9 million per breach compared to those without — a gap wide enough to reframe AI adoption as cost control rather than efficiency. Third-party breaches now account for 30% of all incidents, double the prior-year rate, with the Klue case illustrating exactly how a single compromised credential can extend a blast radius across hundreds of downstream customers. A YesWee production. Built using AI technology. This episode includes AI-generated content.

AI Dev Tool Backdoors, Europe's Ransomware Surge & Dark Web AI Explosion

(00:00:00) AI Dev Tool Backdoors, Europe's Ransomware Surge & Dark Web AI Explosion (00:00:38) MCP Implicit Trust Problem (00:01:22) European Ransomware Supply Chain Surge (00:02:12) Dark Web AI Tool Explosion (00:03:07) SIP Telephony Industrialized Exploitation (00:03:34) Watchpoints and Closing A critical vulnerability in AI developer tooling is rewriting the threat model for software teams worldwide. CVE-2026-12957 in Amazon Q Developer allows a malicious config file to execute arbitrary code using the developer's live AWS credentials — silently, with no prompt. But the story is bigger than one vendor: Claude Code, Cursor, and Windsurf carry structurally identical flaws, all rooted in the Model Context Protocol's implicit trust of project-level config files. Patches are available for Amazon Q Developer; the open question is how many other MCP-compatible tools share the same dangerous assumption. In Europe, ransomware disclosures jumped 55% in the first four months of 2026 versus the same period in 2025. The dominant vector is supply chain compromise: a single third-party breach chain hit 64 organisations and exposed over one million personal records. Qilin is now active across 26 of 31 European countries, putting NIS2 and DORA compliance programs under real operational pressure. On the threat democratisation front, dark web posts referencing AI hacking tools surged from 38 in December 2025 to roughly 1,500 by February 2026 — a 40-fold increase. WormGPT is now freemium. Voice cloning from three seconds of audio succeeds in over 90% of social engineering attempts. The floor for capable attacks has dropped sharply. Finally, a honeypot monitoring SIP telephony systems recorded 1.86 million credential attempts in just 18 days alongside 90,000 toll-fraud call attempts — evidence that enterprise phone infrastructure is being monetised at industrial scale. Today's through-line: implicit trust, in config files, supplier relationships, and telephony auth, is being exploited methodically and at volume. This episode includes AI-generated content.