Cyberdione Labs researcher Erica Windisch published lab research called Tailworm showing that Tailscale SSH port forwarding could let an already-compromised device escalate to root on other servers in a tailnet. Tailscale SSH authorizes by device rather than user credentials, and with port forwarding enabled the researcher connected the root-level Tailscale process to privileged system sockets, then used a race condition to capture a device-enrollment key and enroll a rogue server, a step blocked by Tailnet Lock. She tested Tailscale 1.104.2 and headscale 0.29.5. No CVE or fix existed at publication on October 25, no in-the-wild attacks are reported, and Tailscale is working on fixes.
A leaked post disclosed two critical PHP remote code execution flaws, dubbed PHP2Shell, prompting PHP to schedule an out-of-cycle security release at 14:00 UTC on October 21 for supported branches 8.2 through 8.5. PHP rated both flaws 9.2 and has not described the mechanism, affected versions, or feature requirements, which will come with the release; the leaked identifiers CVE-2026-109618 and CVE-2026-109619 are not confirmed by PHP. watchTowr verified the leak is real and that the flaws allow code execution. No exploitation was reported as of October 21, and end-of-life branches 8.1 and older will receive no fix.
JFrog disclosed CVE-2026-105192, a critical flaw rated 9.8 in LMCache, open-source software that speeds up LLM servers such as vLLM. In multiprocess mode the cache runs as a standalone server reached over ZeroMQ with no authentication, and one message type is deserialized with pickle before its type is checked, letting a crafted message run code. The server is exposed only when an operator binds it to a routable address rather than the default localhost, and the code runs with LMCache's privileges, which are root in the project's official container images. No fix exists yet, and JFrog advises keeping the port on localhost or a trusted cluster network.
ConnectWise warned of a new vulnerability in ScreenConnect, its widely used remote-access platform, affecting both cloud and on-premises deployments, and issued temporary mitigations while it prepares a patch. The flaw involves file-transfer behavior in remote-access sessions and has not yet received a CVE identifier. As an interim measure, administrators are told to disable file-transfer permissions in the ScreenConnect console. ScreenConnect is a favorite tool of managed-service providers and IT teams, which also makes it a repeated target: three earlier ScreenConnect flaws are in the exploited-vulnerabilities catalog, two abused in ransomware, and nearly 6,000 instances are exposed online. There is a concurrent campaign spreading malware through rogue ScreenConnect clients.
CERT/CC disclosed two unpatched flaws in Kaltura's HTML5 video player library that let a remote, unauthenticated attacker read files from a server and run code, with only network access to the endpoint required. Both stem from the same unsafe deserialization in the mwEmbedLoader.php endpoint, which takes a user-controlled ServiceUrl parameter and passes fetched data to PHP's unserialize without validating it. Supplying a file path lets an attacker read any file the web server can access, including credentials and API keys, while the deserialization also enables code execution. CERT/CC could not reach the vendor, and because the endpoint is exposed on Kaltura's shared multi-tenant infrastructure, the flaws can affect many tenants at once.
An unpatched flaw in Calix residential routers used by several US broadband providers lets a remote, unauthenticated attacker create port-forwarding rules that expose devices on the local network to the internet. Tracked as CVE-2026-75501, the missing-authentication issue affects the Calix GS7 XGS model on a specific firmware version, and Calix supplies gear to large providers including Cox and Brightspeed. The researcher who found it reported it to the vendor in June, got no response, and disclosed through CERT/CC after further attempts failed. Because it lets an attacker punch holes through the router's network address translation, internal devices that were never meant to be reachable can be exposed. No fix is available.
A researcher disclosed an unpatched flaw in internet-connected Shark robot vacuums that lets an attacker take control of other owners' units across the same cloud region. The certificate a vacuum uses to authenticate to its maker's Amazon cloud broker was never restricted to that one device, so a certificate pulled from a vacuum's flash memory can send root commands to any Shark vacuum the broker serves. That means watching the camera, driving the robot, reading the stored map of a home, and taking the Wi-Fi password in plaintext. No memory corruption or password guessing is needed. The researcher says the maker has had the report since March with no fix.
Security firm Mindgard disclosed an unpatched flaw in Cursor, the widely used AI code editor, that runs code automatically when a developer opens a malicious repository on Windows. When loading a project, Cursor looks for a Git binary in several places, including the workspace itself, so a file named git.exe planted in the repository root gets executed with no click, prompt, or warning, and it re-runs on a cadence while the project stays open. In a proof of concept, a renamed Windows Calculator launched on its own. Mindgard reported it in December and went public after seven months with no fix, so there is no patch and no CVE, only workarounds.
A researcher who calls themselves Chaotic Eclipse - and who has weaponized every prior Windows flaw they have leaked this year - dropped working proof-of-concept code for two unpatched zero-days on May 12. YellowKey lets anyone with physical access to a Windows 11 or Server 2022/2025 machine plug in a USB stick, hold CTRL during a reboot into the Windows Recovery Environment, and get a shell with full access to the BitLocker-protected drive. GreenPlasma is a privilege escalation against the CTFMON service that hands an unprivileged user a path to SYSTEM. Independent researchers including Will Dormann and Kevin Beaumont have confirmed that YellowKey works as advertised.
Researcher Hyunwoo Kim disclosed Dirty Frag yesterday after an unrelated third party broke the embargo five days early. The flaw chains two Linux kernel page-cache write bugs (xfrm-ESP and RxRPC) to give any local user root access on every major distribution - Ubuntu, RHEL, CentOS Stream, AlmaLinux, openSUSE Tumbleweed, Fedora. Like Dirty Pipe and last week's Copy Fail, it's a deterministic logic bug with no race condition required and no kernel panic on failure. PoC is public on GitHub. The ESP variant patch was merged into the netdev tree on May 7 but distribution kernels remain unpatched. No CVE assigned yet because the embargo broke early.