Home Assistant Repair Automation With BearClaw and Joanna

I rebuilt one of the more useful pieces of my Home Assistant stack this week: a repair loop that lets Home Assistant signal a problem, lets BearClaw decide how to route it, and then dispatches Joanna to handle the fix. The end result is much better than a plain notification because I can see what was dispatched, what completed, and where errors are starting to pile up.

The problem

Home Assistant is excellent at detecting issues, but detection alone is not enough. I wanted a path that could take an automation-triggered issue, hand it to my BearClaw infrastructure, and then let Joanna work the remediation instead of leaving me with another message to chase down later.

That becomes even more important once you have several infrastructure automations in play. Notifications are easy to miss. A repair workflow with telemetry is much easier to trust.

The fix

I kept the Home Assistant side focused on transport and telemetry. Home Assistant now sends the signal, BearClaw handles the dispatch path, and Joanna reports back through the same bridge so I can track both dispatch volume and outcomes from the infrastructure dashboard.

The bridge package exposes a REST-backed telemetry sensor and uses attributes for dispatch statistics. That keeps the Home Assistant activity log quieter while still giving the dashboard enough data to show 24 hour, 7 day, and 30 day activity.

sensor:
  - platform: rest
    name: BearClaw Status Telemetry
    resource: !secret bearclaw_status_url
    value_template: "{{ 'ok' if value_json.ok | default(false) else 'error' }}"
    json_attributes:
      - dispatchStats

If you want to see the Home Assistant side, the bridge logic lives in bearclaw.yaml and the dashboard view is defined in the Joanna infrastructure view.

Bonus win

The other thing I wanted was visibility. I added an infrastructure dashboard view that shows how often Home Assistant automations dispatch Joanna, what happened in the last 24 hours, and what the latest completion state looked like. That gives me a quick operational read instead of a pile of separate traces and logbook entries.

Result

• Home Assistant can trigger repair-worthy work without me manually translating the issue.
• BearClaw acts as the infrastructure brain that accepts the event and dispatches Joanna.
• Joanna activity is visible in one place instead of scattered across notifications and traces.
• The dashboard now gives me a simple view of dispatch count, error count, and the latest repair activity.

This is the kind of workflow I want more of in my lab: Home Assistant detects, BearClaw dispatches, Joanna repairs, and I get a clean feedback loop showing whether the system is actually helping.

Draft generated with help from Codex