Which AI agents and hardware are required to use this skill?

Any agent that supports SKILL.md and MCP, such as Claude Code, Codex CLI or Cursor. You also need the netatmo-energy MCP server, a Netatmo account with a Smart Thermostat or Smart Radiator Valves, and your own free Netatmo developer app.

How does the skill prevent accidental or unwanted temperature changes?

Write mode is off unless you log in with --write. Then the MCP server previews every change and returns a single-use confirmation token, valid for 5 minutes. The skill tells the agent to show you the preview and to apply the change only after your explicit yes. Temperatures outside 7 to 28 °C are refused by default.

What does it add compared with just asking an AI about my Netatmo data?

It tells the agent which server tool answers which question, uses the server's own statistics and rankings instead of raw history, follows a fixed order to investigate a cold room, and keeps answers honest: no energy figures, data gaps reported, causes given as hypotheses.

Is it free, and where is the source?

Yes. The skill and the netatmo-energy MCP server are free and open source (Apache-2.0): https://github.com/christophe77/netatmo-energy-mcp. Updates are published there first.

Can this skill help me track my actual electricity or gas consumption in kWh?

No. Netatmo does not measure gas or electricity, so the skill never turns boiler heat-demand time into kWh, cost or CO2.

Netatmo Energy Heating Assistant

Example session with this skill installed

The bedroom feels cold tonight. Can you find out why?

Read your context and instructions
Compiled the netatmo energy heating

Example with sample data, not a real home.

Bedroom: 17.1 °C for a 19 °C target

The room has been about 2 °C below its setpoint since 18:30, while the schedule has been in

Comfort (19 °C) since 18:00.

Observed data

Check	Result
Current temperature / setpoint	17.1 °C / 19.0 °C, schedule mode
Heating demand	100 % since 18:00
Boiler	Requested to heat
Open window detected	No
Bedroom valve	Reachable, battery low, radio signal weak
Anomalies (last 24 h)	1 finding, medium confidence: sustained deviation below setpoint, 18:30 to now

Derived metrics

Time below target over the last 24 h: 3 h 40 min (from netatmo_get_heating_summary).
The other rooms reached their setpoints within 45 minutes of the same 18:00 change.

Hypotheses (uncertain)

The valve may not be opening fully: its battery is low and the signal is weak.
The radiator may be too small for this room on a cold evening.

The setpoint and schedule do not explain the gap: the room is asking for heat and the boiler is on.

Next steps

Replace the valve batteries, then check again tomorrow evening.
If you want it warmer tonight, I can set the bedroom to 20 °C until 23:00. I will show you the change first and apply it only if you say yes.

Heat-demand time is when the thermostat asked for heat, not gas used.

Connects securely to your tools. The creator never sees your data.

About this skill

Unofficial skill, not affiliated with or endorsed by Netatmo.

This skill teaches your agent how to use the open-source netatmo-energy MCP server to answer questions about a Netatmo Energy heating system (Smart Thermostat, Smart Radiator Valves) and, if you enable it, to change the heating safely.

Requires the MCP server: netatmo-energy-mcp (free, Apache-2.0), a Netatmo account with Energy devices, and your own free Netatmo developer app. The skill runs no code itself.

What it adds

The cheapest call for each question: current status, one room, boiler demand, batteries and signal, period summary, room ranking, anomalies, schedules.
A cold or overheated room, step by step: room status, then whether the setpoint or schedule explains it, then 24 h history, anomalies, device status and boiler demand. Possible causes are given as hypotheses, never as a diagnosis.
Daily and weekly reports and room comparisons in a fixed table format, using the server's own statistics and rankings.
Safe heating changes: read first, pick the smallest change (a temporary setpoint before a schedule edit), show the preview, and send the confirmation token only after you explicitly say yes. Never batched, never reused.
Honest numbers: boiler heat-demand time is not gas or energy use, and the skill never turns it into kWh, cost or CO2. Data gaps are reported, not filled in.
Error handling for login, rate limits and unsupported hardware.

Setup

Install the MCP server in your agent (for Claude Code: claude mcp add --transport stdio --scope user netatmo-energy -- npx -y netatmo-energy-mcp), or install the Claude Code plugin from the repository, which bundles this skill.
Run npx -y netatmo-energy-mcp login once (add --write to allow changes).
Add this skill and ask about your heating.

The server is read-only by default, stores your Netatmo tokens only on your computer and sends requests only to api.netatmo.com.

How to use

Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.

Reviews

No reviews yet

Be one of the first to try it. Every listed skill passes our trust checks below.

Security scanned

Passed our 8-point scan before listing

1 install

Downloaded by developers to date

Free forever

No account required to browse

Netatmo Energy Heating Assistant