I still don't want to give Claude SSH access, so I built a doctor for my homelab
Description and use
The author built HomeButler to give AI agents a safe, read‑only interface for homelab health checks, reducing risk and simplifying interpretation.
Problem
Giving an AI agent SSH access to a home server exposes it to full shell privileges, creating security risks and making it hard to interpret results.
Outcome
The doctor command provides a calm, focused answer to whether the server is healthy, reducing the blast radius of AI agents and improving operational safety.
Workflow
- Install HomeButler via Homebrew or curl
- Run
homebutler initto set up configuration - Execute
homebutler doctorto get a concise health report - Optionally use
--jsonfor machine‑readable output - Let an AI agent consume the JSON and interpret it
Evidence
HomeButler’s doctor command offers a concise, read‑only, structured summary of critical health metrics
With HomeButler, the agent gets a much smaller surface: homebutler doctor --json
- Record date: 2026-09-08T13:45:50.628Z