hypothesis-comparison-matrix
$
npx mdskill add yogsoth-ai/de-anthropocentric-research-engine/hypothesis-comparison-matrix构建竞争假设的多维度对比矩阵,高亮关键差异,支持知情选择。
SKILL.md
.github/skills/hypothesis-comparison-matrixView on GitHub ↗
---
name: hypothesis-comparison-matrix
description: "SOP: 构建竞争假设的多维度对比矩阵"
version: 1.0.0
category: hypothesis-formation
type: sop
campaign: hypothesis-formulation
input: "竞争假设集 + 区分性预测(来自上游输出)"
output: "结构化对比矩阵(表格 + 综合评估)"
dependencies:
skills:
- subagent-spawning
---
# Hypothesis Comparison Matrix
构建竞争假设的多维度对比矩阵,高亮关键差异,支持知情选择。
## HARD-GATE
<HARD-GATE>
前置条件(全部满足才能开始):
1. 已有 ≥2 个假设(含主假设 + ≥1 个竞争假设)
2. 假设包含 mechanism、predictions 等基本组件
不满足 → 停止,返回错误:需要至少 2 个假设才能构建对比矩阵。
</HARD-GATE>
## Pipeline
1. 前置检查:验证假设集完整性
2. 维度确定:选择对比维度(默认:机制类型/证据支持/可测试性/简约性/解释范围/理论依据)
3. 逐假设逐维度填充:为每个假设在每个维度上赋值
4. 差异高亮:标记假设间差异最大的维度
5. 综合评估:基于矩阵给出综合推荐
6. 输出结构化对比表
## Output Format
```json
{
"dimensions": ["mechanism_type", "evidence_support", "testability", "parsimony", "scope", "theoretical_basis"],
"matrix": [
{
"hypothesis_id": "H1",
"label": "Primary Hypothesis",
"statement": "...",
"mechanism_type": "...",
"evidence_support": "strong | moderate | weak | none",
"testability": "high | medium | low",
"parsimony": "high | medium | low",
"scope": "broad | moderate | narrow",
"theoretical_basis": "Theory name(s)"
}
],
"key_differentiators": ["Dimension where hypotheses differ most"],
"recommendation": "Which hypothesis to prioritize and why",
"caveats": "Important limitations of the comparison"
}
```