Unpatched Tailscale SSH flaws can give authorized users root
Original title: Tailscale warns of SSH flaws allowing root privilege escalation
Cyberdione Labs detailed 11 Tailworm issues in Tailscale 1.104.2 and headscale v0.29.5; patches were expected October 27.
Cyberdione Labs publicly reported 11 vulnerabilities, collectively called Tailworm, after auditing Tailscale 1.104.2 and self-hosted headscale v0.29.5. The most serious is in Tailscale SSH: a user already allowed to connect can race Unix-socket forwarding and run code as root, capture enrollment keys, or delete root files. Other issues include unauthenticated headscale crashes, app-connector traffic that skips ACLs, and memory exhaustion via a hostile DERP relay. Tailscale called the flaws serious, did not advise disabling Tailscale SSH, and aimed to ship fixes by the end of October 27.
Unconfirmed: Tailscale said it was still verifying the full scope and which configurations are affected.
- Findings were reproduced on unmodified tailscale v1.104.2 and headscale v0.29.5.
- Headscale grants SSH forwarding automatically; SaaS Tailscale does so per rule.
- TS_SSH_DISABLE_FORWARDING=1 turns off forwarding without disabling Tailscale SSH.
- No CVE identifiers are given in the report.
Impact: Root code execution and enrollment-key theft for authorized SSH users
| Product | Affected |
|---|---|
| Tailscale | 1.104.2 |
| headscale | v0.29.5 |
- 1.Set TS_SSH_DISABLE_FORWARDING=1 for tailscaled and restart the service.
- 2.Limit Tailscale SSH to users and devices that need it.
- 3.Install the patched Tailscale client when the vendor releases it.
- Vendors
- Tailscale
- Products
- headscaleTailscale SSH
- Organizations
- Cyberdione Labs
From the original article
3 min read at bleepingcomputer.com
The Tailscale security team is working to patch almost a dozen security vulnerabilities that could be exploited to steal credentials, crash servers, gain root privileges, and more. [...]
By Sergiu Gatlan · bleepingcomputer.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.