Most people set up an AI assistant, let it run, and assume it's working. Then one day they realize it stopped doing things days ago and they have no idea when.
AI assistants break in quiet ways. An API key expires. A config file gets overwritten. A cron job drifts. The assistant keeps "running" but isn't doing anything useful. You only notice when something important doesn't happen.
Five minutes every morning catches almost all of this before it causes problems. Here's the exact checklist.
Open your assistant's log file and look at the timestamps:
tail -30 ~/your-assistant/logs/agent.log
You should see entries from the past few hours. If the most recent entry is from yesterday afternoon, something stopped it.
Your assistant talks to an AI model to do its work. If that connection is broken, the assistant is basically brain-dead — it'll go through the motions but can't think or respond.
Quick test for Claude (Anthropic):
curl -s -o /dev/null -w "%{http_code}" https://api.anthropic.com/v1/messages \
-H "x-api-key: $ANTHROPIC_API_KEY" \
-H "anthropic-version: 2023-06-01" \
-H "content-type: application/json" \
-d '{"model":"claude-3-5-haiku-20241022","max_tokens":5,"messages":[{"role":"user","content":"hi"}]}'
200.401 = key invalid. 429 = rate limited. Timeout = provider outage.The remaining checks cover scheduled task verification, tool access validation, memory file health — plus the one-time setup that turns this 5-minute manual check into a fully automated morning alert.
30-day money-back guarantee. Cancel anytime.