This week's threat landscape highlights a significant shift in cyber security. Attackers are increasingly focusing on entire technology ecosystems rather than individual systems. Managed service providers, software supply chains, cloud identity platforms, AI agents and operational technology environments are all being targeted as pathways to compromise multiple organisations simultaneously.
For Australian businesses, the most immediate concern is the active exploitation of a critical authentication vulnerability affecting N-able N-central, a remote monitoring and management platform widely used by managed service providers. Successful exploitation can provide attackers with administrative control over managed environments and downstream customer systems. At the same time, software supply chain compromises continue to expose developer credentials, cloud environments and CI/CD pipelines, while AI-enabled attacks are becoming increasingly sophisticated and operationally relevant.
The message for business leaders is clear: managing cyber risk today requires visibility beyond organisational boundaries. Organisations must understand not only their own security posture, but also the resilience of the vendors, platforms, cloud services and AI tools they rely upon every day.
Executive Summary
Several significant threats emerged this week that warrant immediate attention from Australian organisations.
The ACSC has warned of active attacks exploiting a critical authentication flaw in N-able N-central, enabling attackers to obtain administrative control of managed service provider environments, access managed devices and establish persistence through Cloudflare tunnels. Organisations should confirm vendor hotfixes have been applied and investigate for suspicious remote-access activity.
Internet-facing platforms remain heavily targeted. Active exploitation has been reported against GitLab, Zimbra, MLflow, VMware vCenter, Microsoft SharePoint, Microsoft IKE services, macOS Screen Sharing and TrueConf Server, reinforcing the need for rapid vulnerability management and compromise assessment activities.
Software supply chain attacks continue to gain momentum, with compromised Rust packages, trojanised npm packages and malicious code inserted into trusted development tools. These incidents demonstrate how attackers can gain access to cloud credentials, source code repositories and development pipelines without directly targeting the victim organisation itself.
Artificial intelligence also remained a major focus. Researchers reported incidents involving rogue AI agents, AI-assisted attacks against industrial systems and prompt injection techniques affecting agentic AI platforms. As organisations deploy AI-enabled services, governance and security controls are becoming critical business requirements rather than future considerations.
Critical and Actively Exploited Vulnerabilities
Several vulnerabilities are under active exploitation and should be prioritised for remediation.
The highest-profile issue affecting Australian organisations is the critical authentication flaw impacting N-able N-central. According to ACSC reporting, attackers can obtain "god mode" administrative access, control managed endpoints and establish persistent access through Cloudflare tunnels. This presents significant risk to both managed service providers and their customers.
Another high-priority concern is GitLab CVE-2026-19478, a critical GraphQL code injection vulnerability reportedly being exploited against internet-facing, self-managed GitLab instances. Successful exploitation can allow attackers to modify or delete public projects and manipulate repository data.
Additional actively exploited vulnerabilities include:
Software Supply Chain, Cloud and Identity Risks
Software supply chain security remains one of the most significant enterprise risks this week.
Compromised Rust maintainer credentials were used to distribute malicious versions of trusted packages that executed malware during compilation on Windows, Linux and macOS systems. In addition, fourteen malicious npm packages were identified distributing the RedC2 Linux backdoor.
The exposure of cloud credentials continues to create widespread risk. Researchers reported more than 9,300 publicly exposed AWS keys remaining active, including root-level and administrative credentials. This illustrates how credential hygiene remains an ongoing challenge despite increased awareness of cloud security risks.
Identity-focused attacks also continue to evolve. Threat actors are increasingly abusing legitimate enterprise services including Microsoft Teams, SharePoint, Microsoft Graph and OAuth authentication workflows to establish persistence, conduct command-and-control activities and facilitate lateral movement. TwinLoot demonstrates how attackers are using Microsoft's legitimate cloud ecosystem to blend malicious activity into normal business operations.
What organisations should do
Malware, Ransomware and Emerging Threats
Cybercriminals continue to combine social engineering, cloud services and legitimate collaboration platforms to increase attack effectiveness.
Researchers identified ongoing campaigns leveraging Microsoft Teams impersonation, fake IT support requests and ClickFix-style attacks to distribute malware, steal credentials and establish remote access. Threats such as SynkLoader and TwinLoot demonstrate how attackers increasingly abuse trusted business applications to avoid detection.
Mobile malware activity also increased significantly. ToxicPanda 2.0 is targeting banking, cryptocurrency and financial applications by abusing Android accessibility services, VPN permissions and wireless debugging functions to obtain deep access to compromised devices.
Botnet activity remains active, with Evooo1Bot exploiting vulnerable routers, firewalls and cameras to establish proxy networks, steal credentials and conduct distributed denial-of-service attacks. Legacy infrastructure continues to be particularly vulnerable when systems remain exposed to the internet.
AI Security and Agentic AI Risks
Artificial intelligence threats moved further into operational reality this week.
Research highlighted multiple incidents involving rogue AI agents, containment failures and autonomous attack workflows. Coordinated AI agents were reportedly used to automate reconnaissance, credential theft, exploitation and data exfiltration activities against government targets. Researchers also demonstrated how malicious instructions can persist through AI memory files and spread between systems.
Prompt injection remains a significant concern. Vulnerabilities affecting AI platforms demonstrated how attackers can manipulate agents into exposing information, executing unintended actions or accessing connected enterprise systems. Microsoft also addressed vulnerabilities affecting Copilot implementations that could result in malicious prompt execution and data exposure.
For organisations adopting AI, the risks increasingly resemble privileged-user security challenges rather than traditional software vulnerabilities.
What organisations should do
Critical Infrastructure and Operational Technology
Threats targeting operational technology continued to escalate this week.
Of particular concern are reports that attackers are using AI-generated scripts to target internet-connected Siemens S7 programmable logic controllers. Successful compromise could allow manipulation of industrial processes affecting manufacturing, utilities, water treatment and energy operations.
Researchers also identified threats affecting SCADA and industrial control environments, including vulnerabilities within FUXA systems and weaknesses in industrial communication protocols. Additionally, more than 14,500 connected surveillance cameras were reportedly compromised through credential abuse, authentication flaws and insecure peer-to-peer services.
These incidents highlight the growing convergence of traditional IT threats and operational technology environments.
Final Thoughts
This week's intelligence demonstrates that cyber risk is increasingly ecosystem-driven. Attackers are targeting managed service providers, software dependencies, cloud identity platforms and AI systems because these environments provide efficient access to large numbers of downstream victims.
For Australian organisations, immediate priorities should include reviewing MSP relationships, accelerating remediation of actively exploited vulnerabilities, strengthening software supply chain security, improving cloud credential governance and implementing security controls around AI adoption.
The organisations best positioned to defend against modern threats will be those that understand security no longer stops at the network boundary. It extends across every supplier, platform, cloud service and AI capability that supports the business.