maintenance-downtime

Run the metrics subcommand with an equipment ID and date range. MTBF equals total operating time divided by number of failures, and MTTR equals total repair time divided by number of repairs, with availability also reported.

Use the pareto subcommand with a date range and a group-by option of equipment, reason, shift, or operator. The output ranks contributors by downtime hours and highlights the vital few driving most losses.

The skill uses a YAML catalog with codes like mechanical_failure, electrical_failure, changeover, material_shortage, no_operator, and quality_hold, grouped into equipment, process, supply, labor, and quality categories. You can customize the catalog file path via DT_REASON_CODES_PATH.

No, it only tracks unplanned downtime events. For planned maintenance scheduling use the maintenance-pm skill, and for real-time machine state monitoring use the mes-connector skill instead.

The estimate multiplies downtime hours by a flat DT_COST_PER_HOUR rate, which ignores product margin, downstream delays, and customer impact. Treat it as a directional indicator rather than a financial figure.

Overlapping events double-count downtime for the same equipment. The skill documentation recommends enforcing a single active event per equipment policy to keep metrics accurate.