Back to skills
SKILL.md
Plausibility Ranking
ASecuritySOP: 对候选解释按合理性进行多维度加权排序
- 417 stars
- 0 votes
- 0 copies
- 3 views
- Added May 27, 2026
Security analysis
100/100Pro scans all 2 files and shows the line behind each finding
npx -y skills add yogsoth-ai/de-anthropocentric-research-engine --skill plausibility-ranking --agent claude-codeAre you the author of Plausibility Ranking?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/yogsoth-ai-plausibility-ranking)---
name: plausibility-ranking
description: "SOP: 对候选解释按合理性进行多维度加权排序"
version: 1.0.0
category: hypothesis-formation
type: sop
campaign: hypothesis-formulation
input: "候选解释列表 + 相关证据(来自 explanation-generation 输出)"
output: "合理性排序列表 + 各维度评分 + 排序理由"
dependencies:
skills:
- subagent-spawning
---
# Plausibility Ranking
对候选解释从证据一致性、简约性、解释范围三个维度评分,加权排序,产出优先级列表。
## HARD-GATE
<HARD-GATE>
前置条件(全部满足才能开始):
1. 已有 ≥2 个候选解释(来自 explanation-generation)
2. 每个解释有 mechanism 和 predictions 字段
不满足 → 停止,返回错误:需要至少 2 个候选解释才能排序。
</HARD-GATE>
## Pipeline
1. 前置检查:验证候选解释列表完整性
2. 证据一致性评分(0-10):与已知证据的符合程度
3. 简约性评分(0-10):解释所需假设数量(越少越高分)
4. 解释范围评分(0-10):能解释多少相关现象(不仅仅是当前异常)
5. 加权排序:默认权重 0.5/0.3/0.2(证据/简约/范围),可调整
6. 输出排序列表 + 理由
## Output Format
```json
{
"weights": {"evidence": 0.5, "parsimony": 0.3, "scope": 0.2},
"rankings": [
{
"rank": 1,
"explanation_id": "E1",
"statement": "...",
"scores": {
"evidence_consistency": 8,
"parsimony": 7,
"explanatory_scope": 6
},
"weighted_score": 7.4,
"rationale": "Why this ranks here",
"key_weakness": "Main reason it might be wrong"
}
]
}
```
Files in this skill
- SKILL.md
- prompt.md
Attribution
Comments
Loading comments…
Newsletter
Get the weekly list of safe new skills, and the ones we pulled.