Standing queue-health watcher
Runs on a recurring cadence without prompting, scans the calendar and pipeline state, and raises an alert only when a threshold breach appears — an empty channel window, overdue handoff, or same-day collision.
In / outCalendar + pipeline state (periodic scan) → alert on breach; silence when clean
You might say…
“I don't want to check the calendar every week. I want to be told when something is actually about to go wrong.”
What it does
Scheduled twin of the health-check skill. Runs on a recurring cadence (e.g. weekly) without being asked, scans the calendar and pipeline state, and raises an alert only when a threshold breach appears — an empty channel window inside the planning horizon, an overdue handoff, or a same-day collision. Earns agent status because the value is unattended, periodic vigilance rather than a one-shot judgement call.
Trigger: Deploy once to monitor the content calendar on a recurring schedule; fires only on threshold breaches, not routine updates.
Autonomy: Acts autonomously — runs end-to-end without a human in each loop.
Recognise the problem?
The primitives are the commodity part. The fastest next step is a conversation about composing them into something that works for you.
Start a conversation