A recent security disclosure known as the "Tailworm" report, published by researcher Erica Windisch of Cyberdione Labs, has highlighted a series of unpatched vulnerabilities and design concerns affecting Tailscale and its open-source equivalent, Headscale.

Testing versions such as* Tailscale 1.104.2* and Headscale 0.29.5, the report outlines 11 specific findings, with the most critical focusing on Tailscale SSH and port forwarding.

Key Takeaways of the Vulnerabilities

1. How the Attack Works (Lateral Movement & Privilege Escalation):

Tailscale SSH differs from traditional SSH by relying on device identity rather than manually managed passwords or user-supplied keys.

If an attacker gains an initial foothold on a single approved device within the mesh (via phishing, a vulnerable web app, etc.), they can exploit local and remote Unix socket/port forwarding behaviors built into Tailscale's SSH server (tailscaled, which operates with root privileges).

In a lab setting, this allowed a low-privilege session to abuse race conditions (TOCTOU) and forwarding channels to escalate straight to root privileges on other permitted servers in less than a second—without stealing passwords, keys, or requiring user interaction.

2. The "Wormable" Concern vs. Reality: The researcher demonstrated that the process could chain together rapidly across a network if broad access rules are in place.

However, security analysts note important context: Tailscale does not cause the initial device compromise, and the exploit relies entirely on existing network access control (ACL) permissions and configurations allowing SSH and forwarding.

3. Current Status & Vendor Response: At the time of disclosure, no official CVE numbers or security patches have been released by Tailscale.

Tailscale's engineering team has acknowledged receipt of the report and stated they are actively reviewing the areas highlighted, while urging users to rely on proper access rules.

Recommended Immediate Mitigations Because no official patch has been deployed yet, administrators managing Tailscale or Headscale networks should adopt defense-in-depth measures immediately:

Tighten Access Control Lists (ACLs): Replace broad, identity-wide, or tag-wide SSH permissions with strict, single-pair source-to-destination rules (one specific device allowed to talk to one specific target). In testing, granular per-pair rules effectively halted the lateral spread.

Disable Port Forwarding: Where possible, explicitly disable SSH port and socket forwarding in your Tailscale rules or configuration. (Note: According to the report, Headscale 0.29.5 currently lacks the option to disable port forwarding entirely, meaning every SSH grant there should be treated as a potential root path).

Turn off Unused SSH: If a host does not require Tailscale SSH functionality, disable it locally using:

tailscale set --ssh=false

Enable Extra Security Features: Turn on node key expirations, configure Tailnet Lock (if available on your plan) to prevent unauthorized signing of rogue nodes, and consider enabling SSH check mode to force user re-authentication.