The **Tailscale** security team is in a race against time to patch nearly a dozen security vulnerabilities that pose significant risks, including the potential for credential theft, server crashes, and root privilege escalation. **Tailscale**, known for its **WireGuard**-based software-defined mesh VPN, allows users to securely connect various computing resources—from individual machines to cloud instances—into a single encrypted 'tailnet' via direct peer-to-peer (P2P) connections. ### The 'Tailworm' Discovery The vulnerabilities, collectively branded as **Tailworm**, were discovered by **Cyberdione Labs** during a security audit of **Tailscale** version 1.104.2 and **headscale** v0.29.5, the self-hosted control server. **Erica Windisch**, founder of **Cyberdione Labs**, confirmed the flaws were reproducible in lab environments on unmodified software and has publicly released the details to expedite remediation. Windisch stated, "It is the opinion of the researchers that rapid awareness of these flaws will expedite remediation across impacted devices. These vulnerabilities were not introduced by the researchers, but were discovered, and discovered with low effort." ### Critical Vulnerabilities Detailed The most critical of the identified issues resides in **Tailscale**'s embedded SSH server. This flaw could allow authorized attackers to gain root code execution through Unix-socket forwarding on a targeted tailnet. Other significant vulnerabilities span various components: * **Tailscale SSH**: Unix-socket forwarding issues could lead to root execution, capture of enrollment keys, and arbitrary root-file deletion. * **Tailscale App Connectors (conn25, experimental)**: Lack of validation allows traffic translation to attacker-chosen destinations, bypassing network ACLs and enabling access to cloud metadata for credential theft, internal LANs, and Kubernetes cluster services. * **headscale Control Server**: Unauthenticated resource exhaustion vulnerabilities could lead to control-server crash loops and block new device onboarding across an organization. * **Tailscale DERP Client**: Unbounded read issues when interacting with malicious relays could cause client-agent memory exhaustion. * **tailscale serve**: Flaws in TCP-forwarding to Unix sockets, combined with race conditions in per-mount authorization, could grant remote administrative-API access and bypass app authorization for some requests. ### Tailscale's Response and Mitigation Advice **Tailscale** acknowledged the report from **Cyberdione Labs**, confirming the "serious" nature of the flaws and their potential for escalated privilege actions by threat actors. "Most of the Tailscale SSH issues we've identified concern privilege escalation. In some circumstances, someone already authorized to connect over Tailscale SSH may be able to perform actions beyond the permissions of their configured local user, including potentially gaining elevated privileges," **Tailscale** stated. Despite the severity, **Tailscale** is not recommending a blanket disabling of the Tailscale SSH feature. **Mike Shaver**, VP of Engineering at **Tailscale**, advised users to instead focus on immediate mitigation steps: * **Review SSH access policies**: Restrict access to only essential users and devices. * **Disable SSH forwarding if not needed**: Set `TS_SSH_DISABLE_FORWARDING=1` in the `tailscaled` service environment and restart the service. This specifically addresses several of the more critical findings. * **Temporarily tighten access to sensitive systems**: Adjust **Tailscale** SSH access policies to grant connectivity only to specific, necessary users and devices during the investigation period. **Tailscale** aims to release a patched client version by the end of Tuesday, October 27, urging users to prepare for and apply the update promptly upon availability.  ## Once attackers have valid credentials, only 37% of their actions are blocked Overall prevention scores can hide what happens after initial access. Once attackers are using valid credentials, prevention drops sharply. The Blue Report 2026 measures defenses technique by technique across 338 million simulations run in customer production environments. [Get the report](https://hubs.li/Q04sB3fb0)