op1-product-idea-screening

name	op1-product-idea-screening
description	Score and prioritize startup opportunities with an explicit weighted framework. Use when selecting which Stage-1 opportunities should advance into MVP design (Stage 3).

Idea Screening (Stage 2)

1. Read Stage-1 input:
   • research/stage1_idea_discovery/opportunity_records.json
2. Run scoring with explicit config:
   • python3 scripts/score_stage2.py --input research/stage1_idea_discovery/opportunity_records.json --weights framework/config/scoring_weights.default.json --trust-config framework/config/source_trust_rank.json --csv-out research/stage2_idea_screening/scoring.csv --json-out research/stage2_idea_screening/decision_log.json
3. Review outputs:
   • research/stage2_idea_screening/scoring.csv
   • research/stage2_idea_screening/decision_log.json
4. Verify per-opportunity decision packet includes:
   • hard_gate_pass
   • evidence_gate (stage1_pass, loss_signal_ok, budget_or_intent_ok)
   • confidence (level, multiplier, reason)
   • per-dimension score, reason, risk_if_wrong
   • final decision and decision_reason
5. Verify ranking summary exists and is coherent:
   • selection_summary.top_ranked_opportunity_id
   • selection_summary.runner_up_opportunity_id
   • selection_summary.score_gap_vs_runner_up
   • selection_summary.why_top_beats_runner_up
6. If top-vs-runner gap is narrow, add manual review notes before advancing.

Quality bar

• No generic comments; tie scores and decisions to evidence.
• Top-ranked idea must explicitly beat runner-up with auditable reasoning.
• Rejections/holds must include concrete rationale and risk notes.