Recipe: Calendar Audit

Analyze meeting load over a week or custom time range. Count meetings, estimate total hours, identify the busiest days, flag back-to-back chains, and surface free blocks.

Prerequisite: Read the spark base skill for tool reference and filter syntax.

Access level required: read-only.

Steps

Step 1: Pull events for the period

For the current week:

events { "week": true }

For a custom range:

events { "start": "2026-04-13", "end": "2026-04-17" }

Record each event's date, start time, end time, and title.

Step 2: Check free time

availability { "week": true }

Or for the same custom range:

availability { "start": "2026-04-13", "end": "2026-04-17" }

This shows free slots within working hours (08:00-20:00), skipping weekends and events marked "free."

Step 3: Compute metrics

From the events list, calculate:

• Total meetings: count of events
• Total meeting hours: sum of all event durations
• Per-day breakdown: meetings and hours per day
• Busiest day: the day with the most meeting hours
• Lightest day: the day with the fewest meetings (best for deep work)
• Back-to-back chains: sequences of 2+ meetings with no gap between them
• Longest free block: the largest contiguous free window from the availability output

Step 4: Check across calendars (optional)

If the user has multiple calendars or accounts:

events { "week": true, "in": "user@work.com" }
events { "week": true, "in": "user@personal.com" }

Compare meeting loads across contexts.

Step 5: Present the audit

Report:

• Summary: N meetings totaling X hours this week (Y% of working hours)
• Per day: table of meetings and hours per day, flagging the busiest and lightest
• Back-to-back: list any chains of consecutive meetings (flag chains of 3+)
• Free blocks: the longest free windows available for deep work
• Observations: any patterns worth noting (e.g., mornings packed but afternoons free, one day completely booked)

Tips

• Run this on Monday to plan the week, or Friday to review how the week went.
• Back-to-back chains of 3+ meetings are worth flagging - they leave no buffer for breaks or follow-ups.
• The ratio of meeting hours to total working hours (assume 8-10 hours/day) gives a quick "meeting load" percentage.
• For a forward-looking audit, use events's start and end with next week's dates.
• Events marked "free" in the calendar don't count against availability - they won't inflate the meeting load.
• Combine with recipe-schedule-meeting to find the best slots for new meetings in a busy week.
• If the user has recurring meetings, patterns will be consistent week-to-week. Focus commentary on what's unusual.