How Muse, Tailscale and MCP let a cloud AI agent securely reach a home network and drive a radio
Dave_Maynor · x · 2026-11-12
Security researcher David Maynor details how he let his cloud AI agent (Muse) securely access his home network and control radio hardware.
- The pain: cloud agents with their own computer/browser are safer, but machines behind NAT (file server, dev box, lab gear) were unreachable — his old workaround was "becoming the USB cable," pasting commands and outputs back and forth
- The fix: Muse ships with built-in Tailscale; combining Tailscale + MCP gives the agent narrow, revocable access to home-network hardware
- Key principle: letting a service you use but don't own reach into your home network only makes sense if access is narrow, user-approved, and revocable
A practical + security-minded writeup of extending cloud agents to local hardware.
More from coding & agent
- Dev builds sketch-guessing app with new Gemini Decisions API, guesses terrible drawings — fofrAI · 2026-11-12
- Together Compiler: agent-driven continual compilation for AI inference — togethercompute · 2026-11-12
- Book Review: Building AI Agents for Finance — a practical multi-agent patterns reference — bibryam · 2026-11-12
- Google solicits Kubernetes users to help shape Agent Substrate for Managed Agents — rakyll · 2026-11-12
- Kalshi partners with AnyAlpha to let AI agents trade prediction markets 24/7 — ycombinator · 2026-11-12
- Building apps in the agent era: elegant abstractions still beat raw code generation — cnakazawa · 2026-11-12