Oracle PeopleSoft Servers Hacked in ShinyHunters Data Theft Attacks

Article Link: https://www.bleepingcomputer.com/news/security/oracle-peoplesoft-servers-hacked-in-shinyhunters-data-theft-attacks/

• Oracle disclosed a critical zero-day vulnerability in their PeopleSoft PeopleTools software that allows unauthenticated remote code execution (RCE) on affected systems. The flaw impacts PeopleTools versions 8.61 and 8.62.
• The vulnerability can be exploited remotely without valid credentials, making internet-facing PeopleSoft environments particularly vulnerable. Threat actors reportedly leveraged the flaw as part of a broader attack chain to gain initial access and establish persistence within compromised environments.
• Security researchers and incident responders confirmed active exploitation of the vulnerability in real-world attacks, with a significant concentration of victims in the education sector.
• Attackers linked to the ShinyHunters allegedly used the vulnerability to compromise hundreds of PeopleSoft instances across more than 100 organizations, enabling large-scale data theft, extortion, and potential regulatory and reputational consequences.
• Additional information: https://www.bleepingcomputer.com/news/security/oracle-mitigates-peoplesoft-zero-day-exploited-in-data-theft-attacks/

Agentjacking Attack Tricks AI Coding Agents Into Running Malicious Code

Article Link: https://thehackernews.com/2026/06/agentjacking-attack-tricks-ai-coding.html

• Researchers at Tenet Security disclosed a new attack class called “Agentjacking” that can trick AI coding agents such as Claude Code and Cursor. The attack exploits how these tools ingest external error data such as logs from Sentry, allowing adversaries to inject malicious inputs that can trick the AI into running attacker-controlled commands on a developer’s machine.
• The attack exploits Sentry’s publicly accessible Data Source Name (DSN) to inject malicious error events. These crafted events are retrieved through the Model Context Protocol (MCP) and misinterpreted by AI coding agents as legitimate debugging guidance, leading them to execute embedded commands with the developer’s permissions.
• Successful exploitation can result in arbitrary code execution on developer endpoints and the exposure of sensitive data without the need for phishing, malware delivery, or direct infrastructure compromise.
• Researchers identified thousands of organizations with potentially exposed DSNs and observed high success rates in controlled testing across popular AI coding assistants, suggesting the attack could scale as AI-driven development tools become more widely adopted.
• Additional Information: https://tenetsecurity.ai/blog/agentjacking-coding-agents-with-fake-sentry-errors/

Critical HVAC and UPS Vulnerabilities Could Let Hackers Disrupt Data Centers

Article Link: https://www.securityweek.com/critical-hvac-and-ups-vulnerabilities-could-let-hackers-disrupt-data-centers/

• Security researchers at Claroty have disclosed multiple high-severity vulnerabilities affecting widely deployed data center infrastructure systems, including Vertiv uninterruptible power supply (UPS) network management cards and Trane HVAC controllers, which could allow attackers to remotely compromise critical facility management systems.
• The vulnerabilities include authentication bypass, remote code execution (RCE), and information disclosure flaws in web-based management interfaces. When combined, these issues could allow unauthenticated attackers to gain full remote control of UPS and HVAC systems.
• Compromise of UPS systems could enable attackers to disrupt backup power, trigger unsafe shutdowns, or interfere with power conditioning. Similarly, exploitation of HVAC systems could disable cooling, both scenarios pose a direct threat to server availability and hardware integrity.
• With most modern digital services rely on data center uptime, a successful attack could cause outages across cloud platforms, enterprise applications, and other critical processes.
• This highlights the importance of Business Continuity and Disaster Recover planning, especially for critical processes and technologies an organization relies on.

CISA Directive Orders Agencies to Prioritize Vulnerability Patching in a New Way

Article Link: https://cyberscoop.com/cisa-vulnerability-remediation-directive-bod-26-04/

• The Cybersecurity and Infrastructure Security Agency (CISA) issued a new Binding Operational Directive (BOD) requiring U.S. federal agencies to prioritize vulnerability remediation based on exploitability and risk. These criteria include whether vulnerabilities are internet-exposed, actively exploited, easily automated, or capable of full system compromise.
• The directive introduces a risk-based patching approach focused on “patch smarter, not harder,” with remediation timelines as short as three days for the highest-risk vulnerabilities. It aims to address the growing speed of exploitation and the increasing role of artificial intelligence in vulnerability discovery and weaponization.
• The directive signals a shift toward intelligence-driven vulnerability management that prioritizes real-world exploit risk over patch volume. Although mandatory only for federal agencies, it is likely to influence private-sector practices by reshaping patch cycles, security tools, and risk prioritization strategies.
• The directive underscores a broader trend: attackers are moving faster, and the most critical vulnerabilities must now be remediated in days, not weeks. Organizations that adopt similar prioritization can reduce exposure windows and strengthen overall cyber resilience.
• Additional information: https://www.cisa.gov/news-events/directives/bod-26-04-prioritizing-security-updates-based-risk#Note9

Critical LangGraph Vulnerability Chain Enables Full Server Takeover

Article Link: https://cyberpress.org/critical-langgraph-vulnerability/

• Researchers disclosed a critical vulnerability chain in the LangGraph framework that can lead to full remote code execution (RCE) on affected self-hosted deployments, impacting systems used to build and run AI agents at scale.
• The attack chains multiple vulnerabilities, including a SQL injection flaw in the SQLite checkpointer and unsafe msgpack deserialization, allowing attackers to manipulate stored agent state and inject malicious payloads that execute during checkpoint loading.
• Successful exploitation can give attackers full control of the server by running system commands, allowing them to take over AI agent environments and access sensitive data such as API keys, credentials, conversation histories, and connected system
• Because AI agent frameworks often operate with broad privileges and long-lived credentials, a compromise can expose enterprise data, enable lateral movement, and undermine multiple downstream systems, turning the AI agent layer into a high-value attack surface.
• The vulnerabilities have been patched in updated releases, and organizations running self-hosted LangGraph deployments are advised to upgrade immediately, restrict access to agent interfaces, and apply least-privilege controls to connected data sources and credentials.

Pharma Giant Novo Nordisk Discloses Breach of Clinical Trials Data

Article Link: https://www.bleepingcomputer.com/news/security/pharmaceutical-giant-novo-nordisk-discloses-security-breach/

• Danish pharmaceutical company Novo Nordisk disclosed a data breach involving unauthorized access to internal systems and the exfiltration of clinical trial data, including pseudonymized patient records and healthcare professional contact information.
• Attackers gained access to internal IT systems and copied non-public clinical trial datasets. While the patient data was pseudonymized (using IDs and non-direct identifiers rather than names), it still included sensitive medical, demographic, and lifestyle information.
• Exposed information includes health and biomarker data, trial participation details, and lifestyle factors such as smoking, alcohol use, and BMI. Although not directly linked to patient names, this type of data can still carry privacy risks if combined with other datasets or used in targeted phishing and social engineering campaigns.
• Data from healthcare professionals was also exposed in the breach, including contact details that increase the risk of targeted phishing and impersonation attacks across clinical networks.
• The incident highlights the importance of data minimization and built-in anonymization in clinical research systems. Even pseudonymized data can pose residual risk, so organizations should limit stored data, separate identifiers from clinical information, and enforce strong access controls and encryption to reduce exposure in the event of a compromise.

ServiceNow Flaw Exploited to Gain Unauthorized Access to Customer Instances

Article Link: https://thehackernews.com/2026/06/servicenow-flaw-exploited-to-gain.html

• ServiceNow disclosed a security incident involving a vulnerability that allowed unauthenticated users, under certain conditions, to gain unauthorized and elevated access to customer instances.
• The issue stemmed from endpoint configuration weaknesses affecting certain platform releases and specific instance configurations. Attackers were able to bypass intended authentication controls and query instance tables before a security update was deployed.
• ServiceNow confirmed that a subset of customer instances were successfully queried, indicating real-world access to internal data such as IT workflows, system configurations, and potentially sensitive organizational information.
• Because ServiceNow is deeply embedded in enterprise IT operations, even limited access can enable attackers to map internal systems, identify privileged users, and gather intelligence for follow-on attacks such as phishing or lateral movement.
• ServiceNow deployed a security update, notified affected customers, and is investigating anomalous activity while reinforcing endpoint access controls to prevent unauthenticated access.

Critical Microsoft Outlook and Word Flaws Enable Malicious Code Execution

Article Link: https://cyberpress.org/microsoft-outlook-and-word-flaws/

• Microsoft disclosed three critical vulnerabilities affecting Microsoft Outlook and Microsoft Word that could enable arbitrary code execution without privileges or user interaction.
• The flaws stem from memory corruption issues, including type confusion, use-after-free, and heap-based buffer overflow, within Office rendering components. Because Outlook classic uses Word’s rendering engine for emails, crafted content can be processed automatically via the Preview Pane.
• A malicious email viewed in Outlook’s Preview Pane can trigger code execution, meaning attackers do not need users to open attachments or click links, significantly lowering the barrier for exploitation in phishing scenarios.
• Although no active exploitation has been observed, the vulnerabilities are well-suited for future phishing and ransomware campaigns targeting enterprise environments due to the widespread use of Microsoft Outlook and Microsoft Word. Microsoft has released patches in the June 9, 2026 update cycle, and organizations should prioritize immediate patching, monitor Office activity via EDR, and consider disabling the Outlook Preview Pane in high-risk environments as a temporary mitigation.
• Additional information: https://nvd.nist.gov/vuln/detail/CVE-2026-45456, https://nvd.nist.gov/vuln/detail/CVE-2026-45458, https://nvd.nist.gov/vuln/detail/CVE-2026-47635