Automation health: for each source things arrive from (notes, voice memos, email, a connected account's events, a schedule), count what arrived over the last day and week, say what the last arrival started (which rule, which run) and whether it worked, count the runs that worked, failed or are still going, say the last error in plain words, and say whether the source is working, quiet, failing or not connected
Give now and sources: each { source_id, kind (carrier, event-source, schedule), label, connected, arrived: [{ id, at, title? }], rules: [{ rule_id, label, enabled }], runs: [{ at, state, error?, rule_id?, arrival_id?, run_id?, task_id? }] }. Returns each source with last { at, title } (or nothing_yet in plain words), last_started (each rule the last arrival set off, its run and how it ended), arrivals per period, fires, results per run result, last_error { words, at } and state with its words and tone. Periods, how long each kind may stay silent, states and error words come from Trillion's own settings (src/automation-health-looks.json).
Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this tool by what it does and runs it.
Say to your AI: “Use Trillion to automation health: for each source things arrive from (notes, voice memos, email, a connected account's events, a schedule), count what arrived over the last day and week, say what the last arrival started (which rule, which run) and whether it worked, count the runs that worked, failed or are still going, say the last error in plain words, and say whether the source is working, quiet, failing or not connected”
What your AI calls: run_tool {"tool_id":"tool:read-source-health"}