# Litmind Evidence

> LitMind Evidence Finder — 科研证据检索与归纳系统

- Skill: `meishiwhy/litmind-evidence` (Agent Skill)
- Install (CLI): `npx skillmds@latest add meishiwhy/litmind-evidence`
- Raw SKILL.md: https://api.skillmd.com/api/skills/meishiwhy/litmind-evidence/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: meishiwhy (https://skillmd.com/u/meishiwhy)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/meishiwhy/litmind-evidence

---


# LitMind Evidence Finder

输入一个科学观点，系统自动在 Knowledge Base 中寻找支持证据、反对证据和中性证据，并生成证据总结。

## 工作流程

1. 用户输入研究观点
2. ClaimRetriever 多路检索（语义 + LIKE）
3. ClaimClassifier 判断关系（support/oppose/neutral）
4. Evaluator 评估证据强度
5. 输出结构化证据汇总

## 调用方式

### Python API

```python
from litmind_knowledge.service import KnowledgeBase
from litmind_evidence import EvidenceFinderService
from litmind_analyzer.providers import AnthropicProvider

kb = KnowledgeBase()
provider = AnthropicProvider()
service = EvidenceFinderService(kb=kb, llm_provider=provider)

result = service.find_evidence("Flatfoot increases MTP ROM")
print("Strength:", result.evidenceStrength)
print("Support:", len(result.support), "papers")
print("Oppose:", len(result.oppose), "papers")
```

## 输出字段

| 字段 | 说明 |
|---|---|
| evidenceStrength | Strongly/Moderately/Weakly Supported, Mixed/Insufficient Evidence |
| confidence | 置信度 0.0-1.0 |
| support | 支持证据列表 |
| oppose | 反对证据列表 |
| neutral | 中性证据列表 |