Construct a staffing and successor-capability forecast

Repair capability-plan.json. Time is elapsed hours from 0. Members (initial skills; available window; booked-hour capacity): expert (configure,verify; 0–6; 4), specialist (verify; 0–3; 2), successor-a (deploy; 0–12; 8), successor-b (deploy; 0–12; 8). Critical work (id,skill,duration,predecessor): configure,configure,2,none; verify,verify,2,configure; deploy,deploy,2,verify. All must finish by hour 10. By hour 6, at least one of the two successors must cover configure and at least one must cover verify; the same person may cover both if feasible. transfers rows contain skill,mentor,learner,teaching,practice,verification. Each stage has start,finish and lasts at least 1 hour, in that order without overlap. Mentor must initially know the skill and differ from learner. Mentor attends teaching and verification; learner attends all three stages. Practice is independent. Skill-dependent work may start only after initial evidence or the transfer's verification finish. This is conditional on successful verification, not completed training evidence. work rows contain id,member,start,finish. Preserve durations and predecessor finish-to-start handoffs. Every participant's full booked intervals must fit availability/capacity with no collisions. coverage rows contain skill,backup,availableAfter and must reconcile to the chosen eligible successor's initial skill or verification time, no later than hour 6. forecast contains finishHour (latest work finish), trainingPersonHours (both mentor and learner time) and workPersonHours. One expert can stage both transfers; a second mentor can enable parallel transfers and earlier delivery. Work and successor assignments may differ. Row order is immaterial. Save, rerun and repair without Reset. These durations and gates are scenario planning assumptions.