CubeSandbox is an instant, concurrent, secure, and lightweight sandbox for AI agents. Prior to 0.5.1, any authenticated tenant permitted to call POST /sandboxes can supply metadata["host-mount"] with an arbitrary existing host path because the control plane performs no authorization check or allowed-prefix validation for the path. injectHostDirMounts forwards the path to prepareHostDirVolume, where Cubelet uses mount --rbind through the function to perform a recursive bind mount as root and exposes the host subtree to the tenant's MicroVM through virtio-fs, read-write by default. This allows host-root file access and modification, host code execution, access to the container runtime, and compromise of co-resident tenants without a kernel or virtual-machine-monitor flaw. This issue is fixed in version 0.5.1.
RedAmon before 7.8.0 contains a missing authentication vulnerability in the kali-sandbox terminal server that allows unauthenticated attackers to obtain a root shell by connecting to its WebSocket listener on port 8016. Attackers who can reach the port from co-resident containers or the host loopback can send keystrokes to an interactive /bin/bash session, executing commands as root inside the kali-sandbox container.
Cudy TR3000 before 2.5.42 contains a remote code execution vulnerability that allows authenticated administrators to run root commands by importing a malicious OpenVPN client configuration. Attackers can include tls-verify or ipchange directives, which the incomplete denylist does not filter, to execute commands as root when the router connects to an attacker-controlled VPN endpoint.
Cudy TR3000 before 2.5.42 contains a remote code execution vulnerability that allows authenticated administrators to execute commands as root by uploading crafted OpenVPN private or tls-auth keys. Attackers can close the inline key block with </key>, inject directives such as tls-verify that bypass the .ovpn denylist, and have OpenVPN execute the hook as root.
Cudy TR3000 before 2.5.42 contains a command injection vulnerability that allows authenticated administrators to execute commands as root by injecting newlines into the IPsec keyexchange field of the vpn.set_conf RPC method. Attackers can append a second leftupdown directive to ipsec.conf, which StrongSwan's updown plugin runs as root when an attacker-controlled peer establishes the CHILD_SA.
Cudy TR3000 before 2.5.42 contains a command injection vulnerability that allows authenticated administrators to execute commands as root by injecting newlines into reboot_time. Attackers can submit a reboot_time value followed by a newline and crontab line, which /etc/init.d/autoreboot writes to the root crontab for crond to run.
Cudy TR3000 before 2.5.42 contains a dnsmasq configuration injection vulnerability that allows authenticated administrators to inject directives by embedding a newline in the net.set_dhcp leasetime field. Attackers can append address= entries via /etc/init.d/dnsmasq to alter DNS answers for LAN clients, or insert invalid lines that stop dnsmasq.
Cudy TR3000 before 2.5.42 contains an option injection vulnerability that allows authenticated administrators to inject arbitrary pppd options by embedding quotes and newlines in L2TP credentials. Attackers can submit crafted usernames or passwords to the net.set_wan RPC method, which l2tp.sh writes into a pppd options file, adding options like logfile to the root pppd process.
A Cross-Site Scripting (XSS) vulnerability exists in Rymera Web Co's Product Feed PRO for WooCommerce plugin, allowing improper neutralization of script-related HTML tags. This flaw enables code injection via crafted input. The vulnerability affects versions up to and including 13.5.8. The CVSS score is 5.3, indicating a medium severity level with network attack vector and no user interaction required.
Cudy TR3000 before 2.5.42 contains a stack buffer overflow vulnerability in the CWMP client that allows a malicious ACS to corrupt stack memory via oversized credentials. Attackers controlling the ACS can send Username and Password values up to 256 bytes in TR-069 Download or Upload requests, overflowing 128-byte stack buffers and crashing the CWMP process.
Cudy TR3000 before 2.5.42 contains an insufficient verification of data authenticity vulnerability in the mt_wifi Smart Mesh Technology (SMT) onboarding, which protects messages only with an AES-CBC key derived from public values and a CRC32 check. Adjacent unauthenticated attackers can forge SMT messages in Probe Request vendor IEs carrying an attacker-chosen SSID and pre-shared key, forcing unpaired devices to join attacker-controlled access points.
Cudy TR3000 before 2.5.42 contains a remote code execution vulnerability that allows unauthenticated adjacent attackers to run root commands by advertising a rogue automesh controller over mDNS. Attackers on the upstream Ethernet segment can advertise _cudycmagent._tcp records pointing to a malicious MQTT broker, then send commands encrypted with a fixed AES key that router/batchcmd.lua executes as root.
Cudy TR3000 before 2.5.42 contains an authentication bypass vulnerability in the Mosquitto plugin auth_plugin_jwt.so that accepts unsigned JWTs using alg none. Attackers holding any Cudy-trusted client certificate can authenticate as admincudydevice on port 8883 and publish batchcmd requests to execute commands as root.
Cudy TR3000 before 2.5.42 contains an unauthenticated stack buffer overflow vulnerability in the stun_client TR-111 UDP Connection Request parser that parses requests before verifying username or HMAC. Remote attackers can send a single crafted UDP datagram lacking a space in the request line, causing strncpy() to overflow a 10-byte stack buffer and crash stun_client for denial of service.
Cudy TR3000 before 2.5.42 contains a stack buffer overflow vulnerability in an obuspa getter for the TR-098 WLANConfiguration KeyPassphrase parameter, which copies the stored value with strcpy() into a 4,096-byte stack buffer without a length check. A malicious or compromised ACS can set a 5,000-byte passphrase via SetParameterValues, and a later read of the parameter overflows the buffer and crashes obuspa, causing denial of service.
Xiaomi Mi Browser Android Application (com.mi.globalbrowser) before 14.66.0-gn contains a path traversal vulnerability in the exported ThirdDocViewerActivity component, where CopyExternalFile.copyFromUri() copies content provider streams using an unsanitized filename from the _display_name column without basename reduction or canonical-path checks. A zero-permission co-resident malicious application can supply a crafted content provider with a _display_name containing .. sequences to make the browser write attacker-controlled bytes to an arbitrary path within its external storage directory, escaping the intended cache_docs staging location without requiring user interaction.
KUSANAGI provided by GMO Prime Strategy Co., Ltd. contains an HTTP header injection vulnerability. No specific affected versions or technical details are provided. There is no information about patches or exploits in the wild.
A security vulnerability (CVE-2026-116591) affecting the Grafana plugin for Performance Co-Pilot in Red Hat Enterprise Linux 9.8 Extended Update Support has been addressed. The issue involves a denial of service (DoS) caused by crafted zstd compressed input in the klauspost/compress library. Red Hat has issued an important security update to fix this vulnerability in the grafana-pcp package for multiple architectures. The update is available and should be applied to affected systems.