Tailscale SSH flaw could let attackers gain root access

A cybersecurity researcher has disclosed critical vulnerabilities affecting Tailscale SSH implementations, with details published on the Tailworm research portal. The findings describe privilege escalation across an entire tailnet: by compromising an authorised device, an attacker can exploit Unix socket forwarding paths to reach privileged local services and, in the worst case, obtain root access on target nodes.

The report notes that Tailscale’s SSH model replaces traditional OpenSSH configurations with an identity-based architecture controlled by tailnet access lists, which, if trusted devices are compromised, can create serious risks for restricted resources and even allow extraction of node authentication keys to move laterally within the network.

Tailscale has acknowledged the research and is pursuing fixes. Engineers have outlined ongoing remediations, including redesigning Unix socket authorization logic and restricting background forwarding behaviours. Code changes are already appearing in the project repository, with Pull Requests 21937 and 21942 discussed to establish a dedicated SSH Monitor and to delegate forwarding operations to an isolated monitor process. Advisories and patches will follow once the scope of affected configurations is verified.

In the interim, administrators are advised to audit tailnet access policies and consider hardening steps to neutralise the attack vector.

Immediate hardening steps include restricting SSH permissions to essential users and machines, disabling socket forwarding across endpoints, and enabling an environment flag in the tailscaled daemon: TS_SSH_DISABLE_FORWARDING=1. Additionally, organisations should consider SSH check mode to require interactive re-authentication for sensitive logins.