midday-checkin

To detect task drift, you can run a midday check-in script that scans tasks and recent meeting records to compute delta findings. This identifies overdue items, stale work, and unprocessed meeting records before your afternoon starts.

Pattern detection works by querying an optional DuckDB tasks database and reading meeting markdown records. The script scans these sources to find workload-related signals, stale tasks, and meeting load changes, then outputs findings as categorized JSON.

No, DuckDB access to tasks.db is optional for running a midday check-in. The script can still detect task drift using a required state file for delta detection and optional reading of meeting markdown records.

Yes, the midday check-in supports selective SMS delivery based on specific thresholds. It sends an SMS only when high-severity overdue items or time-sensitive meetings are detected during the pattern detection scan.

Context warming pulls remaining calendar events for the day and pre-associates external participants with recent meeting history and open threads. This prepares relevant context in advance to speed up your afternoon meetings.

A state file is required to compute delta findings between scheduled midday check-in runs. By maintaining state, the script can accurately identify new anomalies, meeting load changes, and task drift since the previous execution.