DJBSEC's CyberNews 2026-06-03

Today’s daily news covers the following categories: Vulnerability Threat Intelligence Ransomware Nation-State/APT Malware

Claude Code GitHub Actions Vulnerability Exposes CI/CD Pipelines

Vulnerability Researchers have disclosed a vulnerability affecting Claude Code integrations with GitHub Actions that could expose software development pipelines to unauthorized actions. The flaw reportedly allows attackers to manipulate workflow behavior under specific conditions, potentially impacting source code, secrets, and deployment processes. CI/CD environments remain attractive targets because they provide direct access to software supply chains. Security experts warn that compromised automation workflows can have downstream effects across multiple projects and organizations. Developers are urged to review GitHub Actions configurations and apply available mitigations. Read More

Anthropic Expands Project Glasswing to Protect Critical Infrastructure

Threat Intelligence Anthropic is expanding its Project Glasswing initiative, bringing its advanced AI-driven security capabilities to critical infrastructure sectors. The program leverages the company’s Claude Mythos cybersecurity models to identify vulnerabilities, analyze threats, and strengthen defensive operations. Officials say the expansion is intended to help organizations in energy, transportation, healthcare, and other essential industries improve cyber resilience. The move reflects growing interest in using AI to address increasingly complex security challenges. Experts note that protecting critical infrastructure is becoming a major focus for both governments and private industry. Read More

Anthropic Broadens Reach of Project Glasswing Security Program

Threat Intelligence Anthropic has announced a broader rollout of Project Glasswing, its AI-assisted cybersecurity initiative focused on vulnerability discovery and threat analysis. The company says the program has already demonstrated success in identifying large numbers of software flaws and security weaknesses. By expanding access, Anthropic hopes to accelerate defensive security efforts across both public and private sectors. The announcement highlights the growing role of AI in proactive cybersecurity operations. Industry observers expect similar AI-driven security programs to become increasingly common in the coming years. Read More

Cisco Praises AI-Powered Bug Hunting Efforts

Threat Intelligence Cisco has highlighted the success of AI-assisted vulnerability research efforts, noting that automated systems are helping uncover security flaws more efficiently than traditional methods. While the company praised the results, it declined to disclose the total number of vulnerabilities identified through the initiative. Security researchers say AI is becoming a powerful force multiplier for bug discovery and code analysis. However, organizations still face challenges in validating, prioritizing, and remediating the growing volume of findings. The development underscores how AI is reshaping vulnerability management and software security practices. Read More

Oracle WebLogic Vulnerability Added to Known Exploited Vulnerabilities Catalog

Vulnerability A critical Oracle WebLogic vulnerability, tracked as CVE-2024-21182, has been added to CISA’s Known Exploited Vulnerabilities catalog after evidence of active exploitation. The flaw can potentially allow attackers to compromise vulnerable enterprise application servers. Because WebLogic is widely used in corporate environments, successful attacks could have significant operational consequences. Security agencies are urging organizations to prioritize remediation efforts immediately. The inclusion in the KEV catalog signals that real-world attackers are already weaponizing the vulnerability. Read More

AI-Built Ransomware Toolkit Automates EDR Evasion and Network Discovery

Ransomware Researchers have uncovered an AI-assisted ransomware toolkit capable of automating tasks such as endpoint detection evasion, Active Directory discovery, and attack planning. The toolkit uses AI to streamline activities that traditionally required significant operator expertise and manual effort. Security experts warn that such automation could make ransomware operations faster, more scalable, and more accessible to less-skilled attackers. While human operators still oversee campaigns, AI is increasingly becoming part of the offensive toolkit. The discovery highlights the growing convergence of artificial intelligence and cybercrime. Read More

China Uses Dual-Track Cyber Campaign Against Czech and Taiwan Organizations

Nation-State/APT Researchers report that Chinese state-linked threat actors are conducting coordinated cyber campaigns against organizations in the Czech Republic and Taiwan. The operation combines traditional espionage techniques with broader influence and information-gathering activities. Analysts say the campaign targets government, technology, and strategic sectors to collect intelligence and maintain geopolitical advantages. The dual-track approach demonstrates the increasing sophistication of nation-state cyber operations. Organizations operating in politically sensitive sectors are advised to strengthen monitoring and threat hunting capabilities. Read More

DriveSurge Campaign Hijacks Thousands of Websites

Malware A malware campaign known as DriveSurge has compromised thousands of websites and is using ClickFix and FakeUpdate-style tactics to infect visitors. Victims are tricked into downloading malicious software disguised as browser updates or troubleshooting fixes. Researchers say the campaign leverages trusted websites to increase credibility and improve infection rates. Once systems are compromised, attackers can deploy additional malware, steal credentials, or establish persistence. The activity highlights the continued effectiveness of social engineering combined with website compromise. Read More

OpenAI Upgrades GPT-5.5 and Plans Retirement of Legacy Models

Threat Intelligence OpenAI has announced upgrades to GPT-5.5 while outlining plans to retire several legacy ChatGPT models. The company says the newer models offer improved reasoning, coding assistance, and security-related capabilities compared to previous generations. The transition reflects the rapid pace of AI development and the growing demand for more capable systems. Organizations using AI-powered tools may need to evaluate compatibility and workflow impacts as older models are phased out. The announcement highlights how quickly the AI landscape continues to evolve. Read More