Cisco warned of active attacks exploiting CVE-2026-76519, a flaw rated 9.8 in Secure Email Gateway, found during its own internal security scanning. A sender with no login can email a crafted attachment through the gateway to write files to the appliance operating system, after which an attacker could run commands as root. Both the File Reputation service enabled on a mail policy and a Content Scanner Tools version earlier than 25.3.0.5525 must be present. Cisco's advisory lists this file-write flaw and a September SQL injection flaw as actively exploited. Fixes shipped since September 14 in AsyncOS 15.5.5-014 and 16.5.0-780, plus Content Scanner Tools 25.3.0.5525.
Palo Alto Networks patched CVE-2026-0322, a remote code execution flaw rated 8.6 in the PAN-OS management interface that lets an authenticated attacker with network access run code with root privileges on the firewall. It affects PA-Series and VM-Series firewalls and Panorama, while Cloud NGFW and Prisma Access are not affected, and no special configuration is required to exploit it. Palo Alto reports very limited exploitation, and CISA added the flaw to its Known Exploited Vulnerabilities catalog with a federal patch deadline of October 17. Fixed builds are available across the 10.2, 11.1, 11.2, 12.1, and 12.2 branches.
BeyondTrust patched CVE-2026-102764, a critical authentication bypass in Remote Support and Privileged Remote Access that lets an unprivileged remote attacker perform file operations on unpatched appliances, after which Mandiant observed attackers enabling SUID permissions to escalate to root. Google Threat Intelligence Group and Mandiant spotted the initial attacks, which date to late August. BeyondTrust patched SaaS customers on October 8, while self-hosted appliances must be updated manually unless automatic updates are enabled, per advisory BT26-05. Shadowserver tracks more than 1,400 exposed instances, and any internet-reachable appliance left vulnerable during the exploitation window should be treated as potentially compromised.
Previdian warned of exploitation of CVE-2026-105740, a flaw rated 9.9 in Langflow, the open-source platform for building AI applications and agentic workflows now owned by IBM. Any authenticated user can add an MCP server using the stdio transport, and the user-supplied command is passed to bash with no validation, allowlisting, or sandboxing, running when the server list is fetched. Previdian recorded 1,706 exploit attempts from 28 IPs since September 19, with payloads trying to extract the Langflow secret key and superuser password and read the host passwd file. Versions prior to 1.9.0 are affected, and operators should update to the latest release.
F5 released NGINX 1.30.6 and 1.31.7 on October 14 to fix five flaws it rates medium. Three memory bugs, CVE-2026-15176, CVE-2026-93479, and CVE-2026-97205, carry a grouped 9.2 score and could crash a worker process or possibly allow code execution, each requiring a specific configuration. CVE-2026-97629, rated 8.3, is a TLS session ticket key flaw that left a key slot uninitialized after rotation, allowing a forged session that could bypass client-certificate verification on servers using a shared SSL session cache without a defined ticket key. A fifth flaw passes extra HTTP/3 bytes to the backend. None are known to be exploited.
Drupal patched CVE-2026-108097, a flaw in the Field UI module it rates 19 out of 25 on its own critical scale. Field UI built admin tables that ran each row's region callback through call_user_func_array without checking the callable, a path present since Drupal 8.9.0 in 2020, and the fix stops calling the callback. Several exploit paths run through contributed modules, and at least one needs no login, raising an unauthenticated remote code execution risk. Fixes are in 11.4.10, 11.3.19, and 10.6.19, with no fix for end-of-life branches. No exploit was public as of mid-October, but the team warned exploits could appear within hours.
Patchstack detected active exploitation of CVE-2026-85097, a flaw rated 10.0 in the Bricksforge WordPress plugin, beginning October 7. The plugin does not validate the url field inside its temporaryFileUploads parameter, letting an unauthenticated attacker upload and execute arbitrary PHP code on vulnerable sites. Exploitation came from 63 unique source IPs, with six responsible for roughly 46 percent of requests. Versions up to and including 3.1.8.9 are affected; the vendor released a fix in 3.1.8.10, and the current release is 4.0.1. Wordfence published an advisory. Site operators should update immediately and check for unexpected PHP files and administrative changes.
GitLab disclosed CVE-2026-97229, a critical integer overflow in the CI/CD regular expression compiler that parses and evaluates regular expressions in pipeline configuration files. Under certain conditions, an unauthenticated user can execute arbitrary code on the GitLab server through a specially crafted regular expression. It affects versions from 19.2 before 19.2.8, 19.3 before 19.3.4, and 19.4 before 19.4.2, with fixes in 19.2.8, 19.3.4, and 19.4.2. Self-managed installations must upgrade immediately, while GitLab.com already runs the patched version and Dedicated customers need no action. A researcher reported it through GitLab's HackerOne program, and no exploitation is reported.
Threat intelligence firms Defused Cyber and Previdian detected active exploitation of CVE-2026-102255, the pre-authentication server-side request forgery flaw rated 10.0 in SonicWall SMA 1000 appliances, starting October 8, two days after SonicWall released patches on October 6. Previdian's telemetry recorded 135 exploitation attempts from two United States IP addresses as of October 12. The flaw lets an unauthenticated attacker make the appliance issue requests on their behalf to reach internal functionality. Fixed builds are 12.4.3-03670 and 12.5.0-03082. As of reporting, SonicWall had not updated its advisory to reflect the active exploitation now underway.
Bitdefender detailed Midnight Mimosa, a campaign in which low-cost Android phones built on MediaTek platforms ship with a preinstalled system component, com.android.system.lite, that runs with system privileges. It grants itself accessibility, notification, and SMS permissions, installs and removes apps silently, performs ad fraud through an invisible overlay, and turns devices into residential proxy relays and botnet nodes. It disables the Play Store while dropping payloads to evade Play Protect. Bitdefender also found 13 related Play Store apps, and most infections are in Mexico, France, and Italy. Because the component sits in the system partition, users cannot remove it without firmware-level intervention.