Facilitate an evidenced migration decision and follow-up plan

Repair facilitation-plan.json. Fictional sources, all available day 0 (criterion:metric,comparison,value,source,owner): units:totalUnits,eq,8,P8,product-owner; outage:maximumOutage,lte,2,O2,operations-owner; verification:minimumVerification,gte,.5,Q05,quality-owner; budget:totalCost,lte,4000,B4,sponsor; finish:finishAt,lte,8,D8,sponsor; single-window:batchCount,prefer-eq,1,C1,product-owner. Only single-window is a soft preference. Session availability: product-owner,operations-owner,quality-owner 1–3; all three required. sponsor 2–4 is optional at session and sole decision owner. Outcome review by 9. Each batch: integer units≥1; setup .25 day plus .25 day per unit, then verification≥.5 day before reopening. verifyStart=start+.25+.25×units; finish>verifyStart. Batches must not overlap. Cost=700 per batch+200 per unit. session: start,finish,participants,agenda rows {criterion,source,kind:constraint/preference,bound}. Construct ≥2 distinct alternatives {id,batches:[{id,units,start,verifyStart,finish}],assessment}. Assess totalUnits,totalCost,maximumOutage,totalOutage,minimumVerification,finishAt,batchCount; constraintFailures lists failed hard criterion IDs; executionIssues lists migration-boundary and/or overlap when present. decision: proposal ID,owner,at,acknowledgements (all three represented parties),grounds (all six source refs),tradeoffs. Decide after session/evidence, before chosen work, within sponsor availability. Each unmet preference needs {criterion,owner,actual,preferred,followup:action ID}; never soften hard criteria. actions: {id,kind:verification/outcome-review,batch,owner,at,source}. One quality-owner/Q05 verification per chosen batch during verification; one product-owner/P8 outcome-review, batch:null, after final finish by 9. Link preference to that review. Two or three batches can work. Save, test and repair without Reset.