🛡️ CISO Intel — Tuesday, 16-06-2026

Blog generation failed — API error. Presenting Discord briefing as fallback.

Alright, team. Tuesday. Not a Patch Tuesday in the traditional sense for Microsoft, but that doesn’t mean the bad guys took the day off. The wire isn’t exactly screaming, but there are a few things that caught my eye, especially in the AI and supply chain space. Let’s get into it.

🔴 Critical Threats & Active Exploitation

WordPress Plugins — Supply Chain Attack (via CDN) — Attackers injected malicious JavaScript into CDN-served files for OptinMonster, TrustPulse, and PushEngage plugins, creating unauthorized admin accounts and backdoors on compromised WordPress sites. The attack was active between June 12-14. 📌 Source: Sansec / Security Affairs — https://securityaffairs.com/164506/malware/awesome-motive-cdn-supply-chain-attack.html

Oracle PeopleSoft — CVE-2026-35273 (CVSS not yet public, but described as zero-day) — ShinyHunters group exploited a zero-day to compromise over 100 organizations and 300 instances, many in higher education. This was patched on June 10, but active exploitation was confirmed. 📌 Source: Rod's Blog — https://rods.blog/splunk-rce-arch-linux-supply-chain-peoplesoft-vpn-flaws/

🛡️ CVEs Worth Your Attention

CVE-2026-XXXXX | Splunk Enterprise < 10.2.4 / 10.0.7 | CVSS 9.8 | Unauthenticated RCE via exposed PostgreSQL sidecar service endpoint. Allows arbitrary file creation/truncation. PoC: Yes (implied by description of exploitability). This is a nasty one. CVSS 9.8 for a reason. Patch and restrict network access immediately. 📌 Source: Rod's Blog — https://rods.blog/splunk-rce-arch-linux-supply-chain-peoplesoft-vpn-flaws/

⚡ New TTPs & Attack Research

Arch Linux AUR Supply Chain Attack — Attackers hijacked or adopted orphaned packages in the Arch User Repository (AUR) to inject malicious post-install scripts. These scripts deployed a credential-stealing ELF binary (targeting GitHub tokens, SSH keys, browser data) and an optional eBPF-based rootkit. This is a classic supply chain move, but the eBPF rootkit is a nice touch. T1195.002 (Compromise Software Supply Chain: Compromise Software Dependencies and Development Tools). 📌 Source: Rod's Blog / Risky Business Media — https://rods.blog/splunk-rce-arch-linux-supply-chain-peoplesoft-vpn-flaws/

🏗️ DevSecOps & Cloud Security

WordPress Plugin Supply Chain via CDN — As noted in Critical Threats, this isn’t just a WordPress issue, it’s a CDN compromise. Attackers gained CDN API keys and modified JavaScript files served to millions of sites. This highlights the risk of third-party content delivery. T1195.001 (Compromise Software Supply Chain: Compromise Software Dependencies and Development Tools). 📌 Source: Sansec / Security Affairs — https://securityaffairs.com/164506/malware/awesome-motive-cdn-supply-chain-attack.html

🔧 Patches & Vendor Releases

Splunk Enterprise — Patches are available for versions below 10.2.4 and 10.0.7 for the CVSS 9.8 RCE. 🟢 solid fix 📌 Source: Rod's Blog — https://rods.blog/splunk-rce-arch-linux-supply-chain-peoplesoft-vpn-flaws/

🧪 Threat Intel & Malware

BlueRabbit Backdoor — A new Golang-based backdoor with ransomware capabilities, believed to be operated by an Iranian cyber espionage group. It uses RabbitMQ for command operations, Redis for victim state tracking, and S3-compatible MinIO for data exfiltration. This group is doing double extortion, integrating data theft with file encryption. T1071.001 (Application Layer Protocol: Web Protocols), T1567.002 (Exfiltration Over Web Service: Exfiltration to Cloud Storage), T1486 (Data Encrypted for Impact). 📌 Source: Broadcom Inc. — https://symantec-enterprise-blogs.broadcom.com/blogs/threat-intelligence/bluerabbit-golang-backdoor-ransomware

The Gentlemen Ransomware — This group, tracked by Microsoft as Storm-2697, continues to operate “Ransomware-as-a-Service” or a highly coordinated direct attack model. They’ve named Mackay Sugar (Australia’s second-largest sugar producer) as a victim on their leak site. They are escalating, branching into MSP-focused attacks and exploiting edge devices (Ivanti, Fortinet, SonicWall) and SimpleHelp RMM. 📌 Source: SecurityWeek · FortiGuard Labs — https://www.securityweek.com/ransomware-attack-shuts-down-mills-of-australias-second-largest-sugar-producer/

🌐 Industry, Brand & Internet Security

Typosquatting & BGP Hijacking — Ongoing discussions highlight the persistent threat of typosquatting for phishing and brand abuse, and BGP hijacking for traffic redirection and DDoS. Attackers are registering lookalike domains, often exploiting common typos or TLD swaps. BGP hijacking remains a concern due to the protocol’s inherent trust model. 📌 Source: SCRYPEX · Директ Лайн · Ebuilder Security — direct link unavailable

Maine Data Breach Portal Closure — The state of Maine’s public data breach reporting portal was closed after two fake breach notices were posted for VRChat and Discord. This impacts public access to breach information. 📌 Source: The Record (Recorded Future News) — https://therecord.media/maine-closes-data-breach-portal-fake-reports

FIFA World Cup 2026 Cyber Threats — FortiGuard Labs identified a sharp increase in FIFA-themed domain registrations, reused hosting infrastructure, and naming patterns tied to phishing, typosquatting, and malware delivery. FIFA-related logins are already appearing in criminal data streams. 📌 Source: Security Buzz — https://securitybuzz.co/cybercriminals-targeted-the-2026-fifa-world-cup-before-play-even-began/

Columbia Pacific Advisors Data Breach — An investment management firm disclosed a data breach from November 2025, with a review concluding on June 2, 2026, that sensitive PII (SSNs, dates of birth, financial info, health info) may have been exposed. 📌 Source: Class Action Lawsuits — https://classactionlawsuits.com/columbia-pacific-advisors-data-breach/

Infinite Campus Data Breach — ShinyHunters stole personal information from over 137,000 school staff accounts in a Salesforce data theft attack targeting the Infinite Campus K-12 student information system. Data includes names, emails, job titles, and phone numbers. 📌 Source: BleepingComputer — https://www.bleepingcomputer.com/news/security/infinite-campus-data-breach-affects-137-000-school-staff-accounts/

June 2026 Stealer Logs Data Breach — Have I Been Pwned added a collection of 56 million unique email addresses and 124 million unique passwords from various stealer logs to its database. 📌 Source: Have I Been Pwned — https://haveibeenpwned.com/PwnedWebsites/June2026StealerLogs

🤖 AI & LLM Security

Prompt Injection: A Structural Flaw, Not a Patchable Bug — OWASP’s June 2026 report (summarized by Tech Times and DEV Community) argues that prompt injection is an architectural limitation of LLMs, not a bug that can be conventionally patched. It remains LLM01 on OWASP’s Top 10. LLMs cannot reliably distinguish between instructions and data. This is a fundamental challenge for agentic AI. 📌 Source: Tech Times · DEV Community — https://www.techtimes.com/articles/305886/20260614/ai-agent-security-hits-its-reckoning-prompt-injection-may-be-permanent-flaw.htm

AI Agent Vulnerability Discovery — A cybercriminal group used an LLM to discover a semantic logic flaw in an open-source admin tool’s authentication, bypassing 2FA in specific request sequences. This highlights AI’s capability for advanced vulnerability research. 📌 Source: testRigor — https://testrigor.com/blog/ai-cybersecurity-exploits-the-security-flaw-traditional-testing-failed-to-catch/

AI Coding Assistant Security Risks — Prompt injection against AI coding assistants has led to critical CVEs (e.g., CVE-2025-53773 for RCE in a major assistant and Visual Studio). These tools amplify dependency confusion risks as they autonomously select packages. Proprietary code and credentials are leaking through AI prompts to external models. 📌 Source: Witness AI — https://witness.ai/blog/8-security-risks-of-ai-coding-assistants/

Detecting Compromised AI Agents in Google Cloud — Google Cloud is promoting a 4-layer defense architecture using Model Armor, Cloud Logging, Google SecOps, and SOAR to detect anomalous agent behavior and contain threats. Focus is on prompt injection hijacking GCP service accounts. 📌 Source: Google Security Operations — https://www.googlecloudcommunity.com/gc/Security-Blog/Prompt-Injection-to-Playbook-Detecting-Compromised-AI-Agents-in/ba-p/682910

Open Source AI Security Tools Maturing — Tools like Garak (LLM vulnerability scanner) and DeepTeam (agentic AI red teaming framework) are gaining traction. Garak v0.15.0 introduced multi-turn GOAT probes and an Agent-breaker probe for tool-use behaviors. 📌 Source: Free and Open Source AI Security Tools · GitHub (praetorian-inc/augustus) — https://www.freecodecamp.org/news/free-and-open-source-ai-security-tools-for-2026/

📋 Compliance & Regulatory

RBI Cybersecurity Framework Deadline for Indian Banks — The Reserve Bank of India (RBI) has given Indian banks until the end of June to address cybersecurity gaps. Non-compliance for fintechs can lead to license revocation and penalties. The framework mandates real-time threat monitoring, incident reporting within six hours, regular vulnerability assessments, and board-level accountability. 📌 Source: Hunto AI · AITechNews — https://www.hunto.ai/post/rbi-cybersecurity-compliance-for-fintechs

💡 Marcus’s Take

Another Tuesday, another reminder that the perimeter is dead. The WordPress CDN compromise and the Arch Linux supply chain attack are both classic T1195 plays – attackers hitting upstream to own downstream. It’s about trust, and how quickly it can be weaponized.

But the real shift is in AI. Prompt injection being an architectural flaw rather than a bug? That’s a fundamental game-changer. It means we can’t patch our way out of it; we have to design around it. Assume compromise, enforce least privilege, and get visibility into agent actions. That’s the blast radius we’re now managing.