Skip to content
Back to skills

Hypothesis Formulation

ASecurity

Campaign: 将 insight 和 gap 转化为结构化的可测试假设

  • 417 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added May 27, 2026
researchgoapi

Works with

  • api
  • mcp

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned May 27, 2026

npx -y skills add yogsoth-ai/de-anthropocentric-research-engine --skill hypothesis-formulation --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Hypothesis Formulation?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Hypothesis Formulation
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/yogsoth-ai-hypothesis-formulation/badge)](https://www.skillsdirectory.com/skills/yogsoth-ai-hypothesis-formulation)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: hypothesis-formulation
description: "Campaign: 将 insight 和 gap 转化为结构化的可测试假设"
version: 1.0.0
category: hypothesis-formation
type: campaign
input: "排序后的 gaps + insights + tensions(来自 gap-prioritization 或上游 repo)"
output: "结构化假设集 + falsifiability criteria + boundary conditions"
strategies:
  - deductive-hypothesis-generation
  - inductive-hypothesis-generation
  - abductive-hypothesis-generation
  - competing-hypothesis-construction
  - hypothesis-operationalization
tactics:
  - theory-mechanism-extraction
  - anomaly-driven-abduction
  - falsifiability-audit
  - competing-hypothesis-matrix
dependencies:
  skills:
    - context-management
    - subagent-spawning
    - web-browsing
    - literature-engine
  mcp:
    - brave-search
    - apify
    - alphaxiv
    - semantic-scholar
---

# Hypothesis Formulation

将 insight 和 gap 转化为结构化的可测试假设 — 回答"如何将 insight 转化为可测试假设?"

## HARD-GATE

<HARD-GATE>
前置条件(全部满足才能开始):
1. 至少 1 个明确的研究 gap 或 insight 已被识别
2. gap/insight 有足够上下文(领域、现有理论、观察到的现象)
3. 研究意图已明确(知道要解释什么或预测什么)

不满足 → 停止,建议先完成 gap-prioritization 或上游工作。
</HARD-GATE>

## Campaign Goal

将模糊的"这里有个 gap"或"这个现象很有趣"转化为精确的、可证伪的、有边界条件的假设。产出不是"解决方案",而是"可测试的命题"。

## Strategy Selection

| Strategy | 适用场景 | 推理方向 |
|----------|---------|---------|
| deductive-hypothesis-generation | 领域有成熟理论,gap 是理论预测与现实的偏差 | 理论 → 预测 |
| inductive-hypothesis-generation | 领域缺乏理论,但有大量经验观察或数据模式 | 数据 → 规律 |
| abductive-hypothesis-generation | 观察到无法被现有理论解释的异常现象 | 异常 → 最佳解释 |
| competing-hypothesis-construction | 需要保持开放性,避免 confirmation bias | 多解释并行 |
| hypothesis-operationalization | 已有方向性假设,需要 formalize 为可测试形式 | 模糊 → 精确 |

CC 根据输入特征(有无理论支撑、有无异常数据、假设成熟度)自主选择。可串行组合。

## Budget Gate

| Tier | 假设产出 | 理论/机制 | 可证伪性 | 竞争假设 |
|------|---------|----------|---------|---------|
| S | ≥2 个结构化假设 | ≥2 个相关理论 | 每个假设 1 个 falsification scenario | 可选 |
| M | ≥3 个结构化假设 | ≥3 个理论 + ≥5 个机制 | 每个假设 ≥1 scenario + boundary conditions | ≥2 个竞争假设 |
| L | ≥5 个结构化假设 | ≥5 个理论 + ≥8 个机制 | 完整 falsifiability audit | ≥3 个竞争假设 + 区分性预测 |

## Hypothesis Structure (标准产出格式)

每个假设必须包含:
- **Statement**: If X then Y(条件-结果形式)
- **Variables**: 自变量、因变量、控制变量、调节变量
- **Mechanism**: 为什么 X 会导致 Y(因果链)
- **Boundary conditions**: 假设成立的前提条件
- **Falsification**: 什么观察结果会推翻此假设
- **Measurability**: 如何测量(操作定义)

## Context Management

- Campaign 开始时调用 context-init
- 每个 strategy 完成后调用 context-checkpoint(硬性约束)
- 所有产出累积在单个 campaign-scoped context file

## Minimum Yield

每次 campaign 执行必须产出:
1. ≥2 个完整结构化假设(含全部 6 个组件)
2. 每个假设的可证伪性论证
3. 假设间的关系说明(互补/竞争/独立)
4. 推荐的后续方向(哪个假设最值得进一步发展为研究问题)

Files in this skill

  • SKILL.md3.7 KB
  • prompt.md1.2 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.