Threat briefs

OpenSSH is affected by multiple vulnerabilities that allow a remote, unauthenticated attacker to bypass security controls, cause a denial-of-service (DoS) condition, or perform privilege escalation on the affected system.

A heap-based buffer overflow in GIMP's DirectDraw Surface (DDS) loader allows for arbitrary code execution via a maliciously crafted image file.

Multiple vulnerabilities in Gitea enable unauthenticated or authenticated attackers to perform remote code execution, privilege escalation, security bypass, data manipulation, and cross-site scripting.

An unidentified, likely Chinese-speaking threat actor is targeting South Korean financial entities using the open-source ARTEX penetration testing framework managed by Large Language Models to conduct reconnaissance and data exfiltration.

Multiple vulnerabilities in WordPress versions prior to 7.1.3 allow for remote denial of service, data compromise, and remote code execution via XSS and SQL injection.

Multiple vulnerabilities, including RCE, SSRF, and XSS, affect SonicWall SMA 1000 series appliances, posing significant risks if chained to bypass authentication.

This detection brief identifies potential persistence mechanisms where non-system users modify Windows scheduled tasks to execute malicious code.

AdFind.exe is a legitimate Active Directory query tool frequently repurposed by threat actors for post-exploitation reconnaissance in ransomware and criminal campaigns.

Adversaries may attempt to enable the Windows Subsystem for Linux (WSL) using the legitimate DISM utility to facilitate the execution of Linux-based tools and bypass Windows-specific security controls.

Adversaries may abuse compromised or legitimate SolarWinds binaries to disable security services by modifying Windows Registry 'Start' values, a technique used for defense evasion.

Adversaries manipulate the Windows PortProxy registry keys to establish unauthorized port forwarding, enabling lateral movement and the bypassing of network segmentation.

Adversaries may achieve persistence on macOS by modifying calendar (.ics) files to trigger the execution of malicious programs at recurring intervals.

Adversaries can exploit the inclusion of sshd in the macOS Full Disk Access list by using localhost-targeted SCP commands to bypass privacy restrictions and access sensitive files.

Adversaries may attempt to evade detection and destroy forensic evidence by deleting critical Linux system logs, requiring monitoring of file deletion events for sensitive log paths.

Attackers can leverage misconfigured pods with hostPID enabled to gain visibility into host-level processes, facilitating container escape and privilege escalation to root on the underlying host node.

Creation or modification of Kubernetes pods with the HostNetwork attribute enabled allows containers to escape their network namespace, potentially enabling host-level traffic interception and privilege escalation.

Adversaries can exploit the hostIPC namespace in Kubernetes pods to access host-level shared memory and IPC mechanisms, enabling container escape and privilege escalation.

Attackers with initial cluster access are abusing the Kubernetes TokenRequest API to mint arbitrary service account tokens, facilitating privilege escalation, cloud resource pivoting, and persistence without leaving filesystem artifacts.

Adversaries are utilizing stolen GCP user or service account credentials to perform unauthorized AI tasks across geographically distant locations, indicating LLMjacking and credential abuse.

High volumes of GenerateContent requests from a single IP against Vertex AI models indicate potential model extraction, automated scraping, or LLMjacking attacks.

An unauthenticated remote attacker can trigger a use-after-free or double-free condition in llama.cpp version b11393 or earlier by sending a malformed POST /completion request to achieve memory corruption.

An unauthenticated arbitrary file upload vulnerability (CVE-2026-81780) in the Hash Form WordPress plugin allows attackers to achieve remote code execution through the 'admin-ajax.php' endpoint.

GitAhead versions 2.5.0 through 2.7.1 suffer from an insecure update mechanism that fails to verify update integrity, allowing attackers to perform a man-in-the-middle attack and execute arbitrary code.

The patool library prior to version 4.0.6 contains an OS command injection vulnerability on Windows that allows arbitrary command execution via crafted archive filenames.

Authenticated users can execute arbitrary OS commands on Windows nodes in Rundeck versions prior to 6.2.0 by injecting shell metacharacters into job options.

The OMGF WordPress plugin is vulnerable to stored Cross-Site Scripting via the 's' parameter in comments-atom feeds, allowing unauthenticated script injection in environments where web servers permit MIME-sniffing.

An unauthenticated arbitrary file read vulnerability (CVE-2026-21589) in multiple Atlassian Data Center products allows attackers to access sensitive configuration files via a path traversal flaw in the webresource router.

This detection logic identifies outbound network connections from Linux hosts originating from anomalous locations or via suspicious binaries to ASNs associated with bulletproof or high-abuse hosting providers.

LiteLLM contains a critical authentication flaw where failure to validate JWT email claims allows attackers to impersonate arbitrary users and escalate to administrative privileges.

A flaw in the Candlepin authorization filter allows low-privileged authenticated users to bypass security checks and access unauthorized resources by exploiting flawed parameter verification.