Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Structured Consensus Campaign — converge multiple perspectives into shared agreement through iterative structured dialogue using Delphi variants, NGT, RAND/UCLA, Consensus Conference methods.
.claude/skills/yogsoth-ai-structured-consensus/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 4% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 50% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 61% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 12% | 0% |
Converge multiple independent perspectives into shared agreement through iterative structured dialogue. This campaign orchestrates Delphi variants, Nominal Group Technique, RAND/UCLA Appropriateness Method, and Consensus Conference protocols to systematically reduce disagreement while preserving legitimate dissent.
| Signal | Strategy | |--------|----------| | Iterative convergence to single answer / establish guidelines / determine threshold | convergence-distillation | | Map disagreement structure / wicked problems / value conflicts | disagreement-cartography | | Aggregate probability judgments / technology timeline / market forecast | futures-calibration | | Establish acceptability standards / medical guidelines / regulatory standards | appropriateness-bounding | | Distill strongest arguments / policy deliberation / interdisciplinary dispute | argument-crystallization |
| Strategy | Method Family | Purpose | |----------|--------------|---------| | convergence-distillation | Classic Delphi, Modified Delphi, NGT | Iterative convergence to single answer | | disagreement-cartography | Policy Delphi, Argument Delphi, SAST | Map disagreement structure | | futures-calibration | Real-Time Delphi, Prediction Markets | Aggregate probability judgments | | appropriateness-bounding | RAND/UCLA Appropriateness, Consensus Conference | Establish acceptability standards | | argument-crystallization | Argument Delphi, Dialectical Delphi | Distill strongest arguments |
| Tactic | Purpose | SOPs | |--------|---------|------| | iterative-convergence-round | Collect → feedback → revise → check → decide | judgment-collection, feedback-distribution, consensus-measurement, round-decision | | disagreement-mapping | Collect → cluster → extract arguments → visualize | judgment-collection, cluster-analysis, argument-extraction, disagreement-visualization | | threshold-calibration | Adjust thresholds → observe consensus changes | threshold-sweep, consensus-classification, consensus-measurement |
| SOP | Input | Output | |-----|-------|--------| | judgment-collection | question, perspectives] | judgments] | | feedback-distribution | judgments], round_n | feedback_report | | consensus-measurement | judgments] | consensus_score, method_used | | round-decision | consensus_score, round_n, stability | continue/stop | | cluster-analysis | judgments] | clusters], cluster_characterization] | | argument-extraction | cluster, judgments] | arguments] | | disagreement-visualization | clusters], arguments] | disagreement_map | | threshold-sweep | judgments], threshold_range | threshold_curve | | consensus-classification | judgments], threshold | consensus_items], dissensus_items] | | consensus-synthesis | rounds_history, final_judgments | consensus_report |
| Parameter | Constraint | |-----------|-----------| | Perspectives/experts | >=4 independent perspectives | | Iteration rounds | 2-4 rounds (until consensus threshold or stability) | | Consensus threshold | >=70% agreement or IQR <= 1 | | Dissent documentation | All non-consensus items must document reasons |
mcp__semantic-scholar__relevanceSearch — find methodological referencesmcp__wiki-vault__vault_search — retrieve prior consensus results from vault<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Strategy | When to use | | --- | --- | | appropriateness-bounding | Establish acceptability standards through RAND/UCLA Appropriateness Method or Consensus Conference protocols. | | argument-crystallization | Distill the strongest arguments from each perspective through Argument Delphi or Dialectical Delphi methods. | | convergence-distillation | Iterative convergence to a single answer through Classic Delphi, Modified Delphi, or Nominal Group Technique rounds. | | disagreement-cartography | Map the structure of disagreement across perspectives using Policy Delphi, Argument Delphi, or SAST methods. | | futures-calibration | Aggregate probability judgments across perspectives using Real-Time Delphi or prediction market mechanisms. |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | context-checkpoint | Append research process and results to the current Phase's context file. Covers both process and results with genuine substance. Use this skill at plan-designated checkpoint points — typically after each strategy completes or at key decision nodes within a research Phase. | | context-init | Create a new context file for a research Phase. Called once at Phase start to initialize the file that subsequent context-checkpoint calls will append to. Use this skill whenever a new research Phase begins and a fresh context file is needed. | | convergence-multi-stakeholder-simulation | Simulates diverse stakeholder perspectives and their strongest objections/support arguments. Shared across steel-manning and consensus campaigns. | | convergence-saturation-detection | Determines when to stop iterating — coverage threshold met or marginal returns diminishing. Shared across all campaigns. |
Optional, no fixed order; the final leaf is always a sop.
| Campaign | When to use | | --- | --- | | convergence-multi-criteria-scoring | Multi-Criteria Scoring Campaign — evaluate and rank candidates against multiple weighted criteria using AHP, BWM, TOPSIS, VIKOR, ELECTRE, PROMETHEE, MAUT methods. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→fail | 99,793 | 30,426 | -70% | 1 | 1 | 0% | 6,955 | 2,738 | -61% | 0 | 0 | — |
case-02 | fail→fail | 71,795 | 44,198 | -38% | 1 | 1 | 0% | 8,274 | 8,219 | -1% | 0 | 0 | — |
case-03 | fail→pass | 19,258 | 12,826 | -33% | 1 | 1 | 0% | 2,141 | 2,529 | +18% | 0 | 0 | — |
case-04 | fail→pass | 13,501 | 9,986 | -26% | 1 | 1 | 0% | 1,975 | 2,053 | +4% | 0 | 0 | — |
case-05 | pass→pass | 12,675 | 10,931 | -14% | 1 | 1 | 0% | 1,186 | 2,241 | +89% | 0 | 0 | — |
case-06 | fail→pass | 17,863 | 16,364 | -8% | 1 | 1 | 0% | 1,938 | 2,914 | +50% | 0 | 0 | — |
case-07 | fail→pass | 15,596 | 13,618 | -13% | 1 | 1 | 0% | 1,548 | 2,490 | +61% | 0 | 0 | — |
case-08 | pass→pass | 17,272 | 14,600 | -15% | 1 | 1 | 0% | 1,832 | 2,958 | +61% | 0 | 0 | — |
case-09 | pass→pass | 17,657 | 28,760 | +63% | 1 | 1 | 0% | 1,899 | 3,984 | +110% | 0 | 0 | — |
case-10 | pass→pass | 12,018 | 3,272 | -73% | 1 | 1 | 0% | 1,834 | 1,740 | -5% | 0 | 0 | — |
case-11 | pass→fail | 12,171 | 8,021 | -34% | 1 | 1 | 0% | 1,262 | 1,700 | +35% | 0 | 0 | — |
case-12 | pass→pass | 33,605 | 52,379 | +56% | 1 | 1 | 0% | 5,832 | 9,461 | +62% | 0 | 0 | — |
case-13 | fail→pass | 13,854 | 9,535 | -31% | 1 | 1 | 0% | 2,514 | 2,811 | +12% | 0 | 0 | — |
case-14 | fail→pass | 15,393 | 11,079 | -28% | 1 | 1 | 0% | 1,552 | 2,070 | +33% | 0 | 0 | — |
case-15 | pass→pass | 19,004 | 14,625 | -23% | 1 | 1 | 0% | 2,189 | 2,883 | +32% | 0 | 0 | — |
case-16 | pass→pass | 15,957 | 10,114 | -37% | 1 | 1 | 0% | 1,759 | 2,278 | +30% | 0 | 0 | — |
case-17 | fail→pass | 13,138 | 8,072 | -39% | 1 | 1 | 0% | 1,721 | 1,755 | +2% | 0 | 0 | — |
case-18 | fail→pass | 32,436 | 2,891 | -91% | 1 | 1 | 0% | 1,756 | 1,676 | -5% | 0 | 0 | — |
case-19 | fail→fail | 15,525 | 2,261 | -85% | 1 | 1 | 0% | 1,779 | 1,604 | -10% | 0 | 0 | — |
case-20 | fail→fail | 16,551 | 7,982 | -52% | 1 | 1 | 0% | 1,666 | 1,763 | +6% | 0 | 0 | — |
case-21 | pass→pass | 16,146 | 2,706 | -83% | 1 | 1 | 0% | 1,898 | 1,688 | -11% | 0 | 0 | — |
case-22 | fail→pass | 8,218 | 7,566 | -8% | 1 | 1 | 0% | 1,384 | 1,603 | +16% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 21 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +32 percentage points is the difference between those two pass rates over the 21 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.