moltbot-ha_skill

Overview

This skill controls Home Assistant smart home devices using the moltbot-ha CLI. It provides discovery, state inspection, and action commands for lights, switches, covers, scenes, and automations with a configurable safety confirmation system. It is designed for scripted and agent-driven workflows with optional JSON output for integrations.

How this skill works

moltbot-ha talks to Home Assistant via its REST API using a long-lived access token (recommended stored in an environment variable). The CLI can list entities, read states, call services, and set entity attributes. A three-level safety system blocks or requires confirmation for critical actions and logs every write to an actions log for auditability.

When to use it

• Discover devices and check sensor states before automations or reports
• Turn devices on/off, set brightness, color temperature, or thermostat setpoints
• Trigger scenes, automations, scripts, or custom service calls from scripts or agents
• Control covers, locks, and alarms while enforcing safety confirmations
• Integrate Home Assistant control into task runners, bots, or CI workflows

Best practices

• Store the Home Assistant token in an environment variable (HA_TOKEN) rather than config files
• Keep safety level at 3 for production to force confirmation on critical domains
• Use allowed_entities or blocked_entities to limit automated access and reduce risk
• Always present the user with the exact critical action and wait for explicit approval before retrying with --force
• Use --json output for programmatic parsing and logging of results

Example use cases

• Morning routine: turn on bedroom light, open blinds, set thermostat
• Night mode: power off lights, activate goodnight scene, close all blinds
• Agent-driven query: list lights and report available entities to a user
• Safe critical action flow: ask user to confirm before opening a garage cover
• Automation testing: call services or scripts with JSON payloads from CI