Agents In Use
What’s new todayLast update
1005 additions

OpenClaw repairs Home Assistant motion lighting

Malone reports OpenClaw identifying unavailable Hue entities, replacing a broken motion-light automation, adding nap exclusions and updating his dashboard.

Published dan-malone.comAgent: OpenClaw
OUR ADAPTATION

Suggested prompt

Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.

Inspect [failed Home Assistant automation] and its entity availability. Explain the fault and propose a replacement preserving [day/night rules] and [bypass conditions]. Show the complete rule for review before enabling it.

What happened

Malone reports OpenClaw identifying unavailable Hue entities, replacing a broken motion-light automation, adding nap exclusions and updating his dashboard.

Reported by Dan Malone

How the agent did it

  1. Connect OpenClaw to Home Assistant through ha-mcp in an isolated container.
  2. Inspect the entities referenced by the failed motion automation.
  3. Identify unavailable Hue bulbs and replace the uneditable blueprint rule.

Where a person stepped in

Malone requests the repair and adds household nap conditions. He isolates the agent and limits access, while the account does not document a separate confirmation for every automation edit.

Dan Malone connected an OpenClaw assistant to Home Assistant through ha-mcp, running the agent in a separate LXC container beside his home-automation system. His concrete repair task concerned landing motion lights that had stopped working for two days. He asked the assistant to investigate from Telegram.

The agent inspected available light entities and found that the old automation referenced two unavailable Hue bulbs. It could not edit the blueprint-based automation directly, so it replaced that automation with a native one implementing the same purpose. Malone reports a nighttime mode with very low warm lighting and a shorter timeout, plus brighter daytime operation. He then requested weekend nap exclusions, and the assistant updated the dashboard to reflect the new automation.

Malone estimates about five minutes of Telegram interaction for the repair. That is his own comparison, not a measured benchmark. He keeps the agent isolated and limits its access to home automation and email drafting rather than critical accounts. The reusable detail is to begin with entity availability and the existing rule’s references before rewriting behaviour. A replacement automation should still be inspected for timing, bypass conditions and household expectations; successful entity discovery does not alone prove a safe or complete rule.

Prompt
Suggested prompt (written by Agents In Use, not the author's)

Source: Blog · Dan Malone

Frequently asked questions

What caused the motion-light failure?

The old automation referenced two Hue light entities that were unavailable.

Why was a new automation created?

Malone says the agent could not directly edit the existing blueprint-based automation, so it made a native replacement.

How was access limited?

The agent ran in an isolated LXC container, and Malone excluded critical systems and financial credentials.

More details

Remaining steps

  1. Apply requested day/night and nap-bypass behaviour.
  2. Update the dashboard and review the replacement automation.

About this account

OpenClaw inspects device state, diagnoses dead references and chooses a native automation replacement when it cannot modify the existing blueprint.

Firsthand account by Dan Malone; original publication date verified against Primary page JSON-LD datePublished 2026-01-30. Outcomes are the author’s reports, not independently reproduced. Suggested prompt written by Agents In Use; not the author’s original prompt.

Credit: Dan Malone

Source published 30 Jan 2026

Source checked 14 Oct 2026

Tools and topics

Home Assistant · ha-mcp · Proxmox LXC · Telegram · Hue lights

Smart home repair · Entity checks · Motion lighting

Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.