Skip to content

LeoYeAI/agent-master-skills / skills/ai-powered-literature-review-skills

literature-reviewer-skill

Unverified

根据用户提供的论文主题,进行系统性中英文文献回顾(Literature Survey)。 采用8阶段工作流,支持CNKI、Web of Science、ScienceDirect等主流数据库, 无需API配置,通过浏览器自动化获取文献信息。 输出包含GB/T 7714-2015引文、标题、摘要的Markdown文档。 当用户提到"文献回顾"、"文献综述"、"帮我找文献"、"中英文文献搜索"、"写综述"等关键词时触发。 NOTE: The 'examples/' folder contains sample outputs and is not included in skill installation.

FileKindSize
AGENTS.mdmarkdown12.8 KB
README.mdmarkdown19.7 KB
SECURITY.mdmarkdown5.1 KB
SKILL.mdskill md22.1 KB
_meta.jsonconfig316 B
agents/explore-agent.mdmarkdown4.2 KB
agents/orchestrator.mdmarkdown17.3 KB
agents/synthesize-agent.mdmarkdown12.4 KB
agents/verify-agent.mdmarkdown2.3 KB
docker-compose.ymlconfig3.3 KB
install.yamlconfig2.6 KB
mcp.jsonconfig5.6 KB
references/cnki-guide.mdmarkdown10.0 KB
references/database-access.mdmarkdown3.1 KB
references/gb-t-7714-2015.mdmarkdown10.8 KB
scripts/__init__.pyscript519 B
scripts/citation_formatter.pyscript5.3 KB
scripts/deduplicate_papers.pyscript4.8 KB
scripts/models.pyscript2.6 KB

agents/explore-agent.md

View on GitHub
# Explore Agent 模板

用于在特定学术数据库执行文献检索的Agent。通过浏览器自动化访问数据库检索页面。

---

## Agent 信息

| 属性 | 值 |
|------|-----|
| **名称** | Explore Agent |
| **类型** | Task Agent |
| **用途** | 文献搜索(浏览器自动化) |

---

## 任务描述

在指定的学术数据库中搜索与给定主题相关的文献,返回结构化的文献元数据列表。

**执行方式**:使用 `browser_navigate` 和 `browser_fill_form` 等工具访问数据库网页,提取文献信息。

---

## 输入格式

```yaml
query:
  keywords_zh: ["深度学习", "神经网络"]  # 中文关键词
  keywords_en: ["deep learning", "neural network"]  # 英文关键词
  search_expression_cnki: "SU=('深度学习'+'神经网络')*('医学图像')"
  search_expression_wos: "TS=((deep learning OR neural network) AND medical imaging)"
  
database: "cnki"  # 目标数据库:cnki, wos, sciencedirect, pubmed

filters:
  year_range: [2020, 2025]  # 年份范围
  max_results: 50  # 最大结果数

session_id: "20240115_dl_survey"
```

### 支持的数据库

| 数据库 | 标识符 | 访问方式 |
|--------|--------|----------|
| CNKI 中国知网 | `cnki` | 浏览器自动化 |
| Web of Science | `wos` | 浏览器自动化 |
| ScienceDirect | `sciencedirect` | 浏览器自动化 |
| PubMed | `pubmed` | 浏览器自动化 |

---

## 执行步骤

### Step 1: 访问数据库检索页面

根据目标数据库访问对应的检索页面:

**CNKI**:
```
https://kns.cnki.net/kns8/AdvSearch?classid=YSTT4HG0
```

**Web of Science**:
```
https://www.webofscience.com/wos/woscc/advanced-search
```

**ScienceDirect**:
```
https://www.sciencedirect.com/search
```

**PubMed**:
```
https://pubmed.ncbi.nlm.nih.gov/advanced/
```

### Step 2: 执行检索

1. 在检索页面填充检索式
2. 设置筛选条件(年份、文献类型等)
3. 执行搜索
4. 等待结果加载

### Step 3: 提取文献信息

从搜索结果页面提取以下信息:
- 标题
- 作者
- 期刊/会议名称
- 年份
- 卷期页
- DOI
- 摘要(如有)

**提取示例**(使用 browser_evaluate):
```javascript
// 提取CNKI搜索结果
const papers = [];
const items = document.querySelectorAll('.result-table-list tbody tr');
items.forEach(item => {
  const title = item.querySelector('.title a')?.textContent?.trim();
  const authors = item.querySelector('.author')?.textContent?.trim().split(';');
  const journal = item.querySelector('.source')?.textContent?.trim();
  const year = item.querySelector('.date')?.textContent?.trim();
  papers.push({ title, authors, journal, year });
});
return papers;
```

### Step 4: 格式化输出

将提取的信息格式化为标准结构:

```json
{
  "id": "C1",
  "source_db": "cnki",
  "title": "文献标题",
  "authors": ["作者1", "作者2"],
  "journal": "期刊名称",
  "year": 2023,
  "volume": "46",
  "issue": "5",
  "pages": "100-110",
  "doi": "10.xxxx",
  "abstract": "摘要内容...",
  "url": "原文链接"
}
```

---

## 输出格式

```json
{
  "agent": "explore",
  "database": "cnki",
  "session_id": "20240115_dl_survey",
  "timestamp": "2024-01-15T10:30:00Z",
  "query": "SU=('深度学习'+'神经网络')*('医学图像')",
  "total_found": 1250,
  "returned": 50,
  "results": [
    {
      "id": "C1",
      "source_db": "cnki",
      "title": "基于深度学习的医学图像诊断研究",
      "authors": ["张三", "李四", "王五"],
      "journal": "计算机学报",
      "year": 2023,
      "volume": "46",
      "issue": "5",
      "pages": "1023-1035",
      "doi": "10.xxxx",
      "abstract": "本文研究了...",
      "url": "https://kns.cnki.net/..."
    }
  ]
}
```

---

## 注意事项

1. **验证码处理**
   - 如遇到验证码,暂停执行
   - 提示用户手动完成验证
   - 验证完成后继续

2. **结果限制**
   - 每数据库最多提取50-100条结果
   - 按相关度或发表时间排序
   - 优先选择近5-10年文献

3. **数据完整性**
   - 尽可能获取完整元数据
   - DOI和摘要为可选,但尽量获取
   - 记录来源URL以便追溯

4. **错误处理**
   - 网络超时:重试3次
   - 页面结构变化:记录错误,尝试备用选择器
   - 访问受限:记录警告,跳过该数据库
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 323kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 323kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 323kUpdated Claude CodeCodexVerifiedInstall