sensitivity-ranking
$
npx mdskill add yogsoth-ai/de-anthropocentric-research-engine/sensitivity-rankingRanks constraints by how much they affect outcomes when changed
- Helps prioritize constraints that most influence system performance
- Uses resource quantification, assumption challenging, and critical chain identification
- Calculates sensitivity using gap magnitude, chain position, and assumption vulnerability
- Returns a ranked list of constraints based on computed sensitivity scores
SKILL.md
.github/skills/sensitivity-rankingView on GitHub ↗
--- name: sensitivity-ranking description: "Rank constraints by sensitivity — which ones most impact the outcome if they shift" version: 1.0.0 category: experiment-execution type: tactic used-by: constraint-analysis orchestrates: - resource-quantification - assumption-challenging - critical-chain-identification --- # Tactic: Sensitivity Ranking ## Orchestration Pattern 1. **Quantify Resources** → spawn `resource-quantification` - For each identified constraint, quantify its current state (demand, supply, gap) - Express gaps in comparable units where possible 2. **Challenge Assumptions** → spawn `assumption-challenging` - For each constraint, identify the assumptions that make it binding - Score assumption vulnerability (confidence × evidence / testability) 3. **Identify Critical Chain Impact** → spawn `critical-chain-identification` - Determine which constraints lie on the critical chain - Constraints on the critical chain have higher sensitivity 4. **Compute Sensitivity Score** → synthesize - Sensitivity = (Gap magnitude) × (Chain position weight) × (Assumption vulnerability) - Chain position weight: on critical chain = 2.0, feeding chain = 1.0, off-chain = 0.5 - Rank all constraints by sensitivity score ## Decision Criteria - **When to use**: After multiple constraints have been identified and need prioritization - **When to skip**: If only 1-2 constraints exist, ranking is trivial - **Threshold**: Constraints with sensitivity score >2× the median are "binding" - **Output**: Ordered list with scores, used by parent strategy to focus effort