ssci-subagent-skills 1.0.0

package/skills/INTEGRATION_PLAN.md

@@ -0,0 +1,405 @@
+# 文献智能体与论文下载技能集成方案
+
+**创建日期**: 2025-12-28
+**状态**: 待集成
+
+---
+
+## 🔍 当前状况分析
+
+### literature-expert 智能体
+
+**文件位置**:
+- `agents/literature-expert.md`
+- `agents/literature-expert-v2.md`
+
+**核心技能配置**:
+```yaml
+core_skills:
+  - processing-citations
+  - writing
+  - validity-reliability
+```
+
+**问题**:
+1. ❌ 这些技能在 `archive/skills/` 中，主要是指导性内容
+2. ❌ **缺少实际的论文检索和下载能力**
+3. ❌ archive中的 `literature-search-skill.md` 明确说明"**不提供任何文献下载功能**"
+
+### 新建的论文下载技能
+
+| 技能 | 平台 | 语言 | 功能 | 状态 |
+|------|------|------|------|------|
+| `pubscholar-auto-search` | PubScholar | 中文 | 自动搜索+下载 | ✅ 已创建，**未关联** |
+| `arxiv-paper-search` | arXiv | 英文 | API搜索+下载 | ✅ 已创建，**未关联** |
+
+**问题**:
+- ❌ 两个新技能与 literature-expert **没有任何关联**
+- ❌ literature-expert 不知道这两个技能的存在
+- ❌ 用户需要手动选择技能，无法自动触发
+
+---
+
+## 🎯 集成目标
+
+让 literature-expert 智能体能够：
+1. ✅ **识别用户需求** - 自动判断需要中文还是英文论文
+2. ✅ **自动调用技能** - 触发 pubscholar-auto-search 或 arxiv-paper-search
+3. ✅ **整合结果** - 统一返回中英文文献结果
+4. ✅ **提供指导** - 基于检索结果提供质量评估和趋势分析
+
+---
+
+## 🔧 集成方案
+
+### 方案1: 更新 literature-expert 的 core_skills（推荐）
+
+更新 `agents/literature-expert.md` 和 `agents/literature-expert-v2.md`:
+
+```yaml
+---
+name: literature-expert
+description: 文献管理专家，整合中英文论文检索、下载、引用管理和研究趋势分析。
+core_skills:
+  - pubscholar-auto-search    # 中文论文检索下载
+  - arxiv-paper-search        # 英文论文检索下载
+  - processing-citations       # 引用格式化
+  - validity-reliability       # 质量评估
+---
+```
+
+**优点**:
+- ✅ 简单直接
+- ✅ 符合现有架构
+- ✅ AI可以自动调用这两个技能
+
+**实现步骤**:
+1. 编辑 literature-expert.md
+2. 添加技能关联说明
+3. 更新触发条件
+
+### 方案2: 创建统一接口技能
+
+创建新技能 `literature-integration`:
+
+**文件**: `skills/literature-integration/SKILL.md`
+
+```markdown
+---
+name: literature-integration
+description: 中英文文献检索统一接口，自动判断并调用合适的检索技能。
+core_skills:
+  - pubscholar-auto-search
+  - arxiv-paper-search
+---
+
+# 文献检索统一接口
+
+## 自动路由逻辑
+
+当用户提出文献检索需求时：
+1. **语言检测**: 中文→pubscholar, 英文→arxiv
+2. **平台检测**: 提到"PubScholar"→pubscholar, "arXiv"→arxiv
+3. **默认行为**: 同时检索中英文文献
+
+## 使用示例
+
+```python
+from skills.pubscholar_auto_search.scripts.pubscholar_searcher import SynchronousPubScholarSearcher
+from skills.arxiv_paper_search.scripts.arxiv_searcher import ArxivPaperSearcher
+
+def search_papers(query, language='both'):
+    results = []
+
+    if language in ['chinese', 'both']:
+        cn_searcher = SynchronousPubScholarSearcher()
+        results.extend(cn_searcher.search(query))
+
+    if language in ['english', 'both']:
+        en_searcher = ArxivPaperSearcher()
+        results.extend(en_searcher.search(query))
+
+    return results
+```
+```
+
+**优点**:
+- ✅ 提供统一接口
+- ✅ 自动语言检测
+- ✅ 结果合并去重
+
+### 方案3: literature-expert 作为协调器
+
+让 literature-expert 作为**主控智能体**，协调其他技能：
+
+```markdown
+## 协调能力
+
+### 自动技能调用
+当用户需要文献检索时：
+1. 识别检索需求（语言、平台、领域）
+2. 自动调用对应技能：
+   - 中文论文 → pubscholar-auto-search
+   - 英文论文 → arxiv-paper-search
+   - 引用格式化 → processing-citations
+   - 质量评估 → validity-reliability
+3. 整合多个技能的结果
+4. 提供综合分析和建议
+
+### 工作流程示例
+```
+用户: "帮我找关于数字鸿沟的论文"
+  ↓
+literature-expert 识别需求:
+  - 关键词: 数字鸿沟
+  - 语言: 中文
+  ↓
+调用 pubscholar-auto-search
+  ↓
+接收检索结果
+  ↓
+提供质量评估和趋势分析
+  ↓
+返回综合报告
+```
+```
+
+**优点**:
+- ✅ literature-expert 作为中心协调器
+- ✅ 整合所有文献相关能力
+- ✅ 提供端到端服务
+
+---
+
+## 📋 推荐实施方案
+
+### 混合方案（方案1 + 方案3）
+
+#### 步骤1: 更新 literature-expert 配置
+
+**编辑文件**: `agents/literature-expert-v2.md`
+
+```yaml
+---
+name: literature-expert
+description: 文献管理专家，整合中英文论文检索、下载、引用管理和研究趋势分析。
+model: claude-3-5-sonnet-20241022
+core_skills:
+  - pubscholar-auto-search    # 中文论文检索下载（新增）
+  - arxiv-paper-search        # 英文论文检索下载（新增）
+  - processing-citations       # 引用格式化
+  - validity-reliability       # 质量评估
+  - writing                   # 学术写作
+---
+```
+
+#### 步骤2: 添加技能调用说明
+
+在 literature-expert 中添加：
+
+```markdown
+## 📚 可用技能
+
+### 自动检索技能
+- **pubscholar-auto-search**: 中文论文检索（PubScholar平台）
+  - 触发: "搜索中文论文"、"在PubScholar搜索"
+  - 功能: 浏览器自动化、智能关键词扩展
+
+- **arxiv-paper-search**: 英文论文检索（arXiv平台）
+  - 触发: "搜索英文论文"、"在arXiv搜索"
+  - 功能: API集成、批量下载摘要/PDF
+
+### 辅助技能
+- **processing-citations**: 引用格式化（GB/T 7714）
+- **validity-reliability**: 文献质量评估
+- **writing**: 学术写作指导
+
+## 🔄 自动工作流程
+
+### 场景1: 中文文献检索
+```
+用户: "搜索关于人工智能的中文论文"
+  ↓
+literature-expert 识别:
+  - 语言: 中文
+  - 关键词: 人工智能
+  ↓
+调用: pubscholar-auto-search
+  ↓
+返回: 中文论文列表 + 质量评估
+```
+
+### 场景2: 英文文献检索
+```
+用户: "在arXiv搜索transformer相关论文"
+  ↓
+literature-expert 识别:
+  - 平台: arXiv
+  - 关键词: transformer
+  ↓
+调用: arxiv-paper-search
+  ↓
+返回: 英文论文列表 + 引用格式
+```
+
+### 场景3: 中英文综合检索
+```
+用户: "帮我找关于社会网络分析的文献"
+  ↓
+literature-expert 识别:
+  - 语言: 未指定 → 同时检索中英文
+  ↓
+并发调用:
+  - pubscholar-auto-search → 中文结果
+  - arxiv-paper-search → 英文结果
+  ↓
+整合并去重
+  ↓
+返回: 中英文文献列表 + 综合分析
+```
+```
+
+#### 步骤3: 创建示例配置文件
+
+创建 `agents/literature-expert-integration.md`:
+
+```markdown
+# 文献智能体集成配置
+
+## 技能关联
+
+| 智能体 | 关联技能 | 触发条件 |
+|--------|---------|---------|
+| literature-expert | pubscholar-auto-search | 中文论文检索 |
+| literature-expert | arxiv-paper-search | 英文论文检索 |
+| literature-expert | processing-citations | 引用格式化 |
+| literature-expert | validity-reliability | 质量评估 |
+
+## 调用示例
+
+### Claude Code 调用方式
+
+```python
+# 文献智能体内部调用
+from skills.pubscholar_auto_search.scripts.pubscholar_searcher import SynchronousPubScholarSearcher
+from skills.arxiv_paper_search.scripts.arxiv_searcher import ArxivPaperSearcher
+
+# 根据用户需求自动选择
+if is_chinese_query(user_query):
+    searcher = SynchronousPubScholarSearcher()
+    results = searcher.search(extract_keywords(user_query))
+else:
+    searcher = ArxivPaperSearcher()
+    results = searcher.search(extract_keywords(user_query))
+```
+
+## 预期效果
+
+- ✅ 用户只需说出需求，智能体自动选择合适的技能
+- ✅ 中英文文献自动合并
+- ✅ 提供端到端的文献服务
+```
+
+---
+
+## 🎯 实施优先级
+
+### 高优先级（立即实施）
+
+1. **更新 literature-expert-v2.md**
+   - 添加 pubscholar-auto-search 到 core_skills
+   - 添加 arxiv-paper-search 到 core_skills
+   - 添加技能调用说明
+
+2. **测试集成**
+   - 验证技能能否被自动调用
+   - 测试中英文文献检索流程
+
+### 中优先级（后续实施）
+
+3. **创建 literature-integration 技能**
+   - 统一接口
+   - 自动语言检测
+   - 结果合并去重
+
+4. **优化工作流程**
+   - 添加结果质量评估
+   - 提供研究趋势分析
+
+### 低优先级（可选）
+
+5. **添加更多平台**
+   - PubMed（生物医学）
+   - IEEE Xplore（工程）
+   - Google Scholar（通用）
+
+---
+
+## ✅ 预期效果
+
+### 集成前
+
+```python
+# 用户需要手动选择技能
+from skills.pubscholar_auto_search.scripts.pubscholar_searcher import SynchronousPubScholarSearcher
+
+searcher = SynchronousPubScholarSearcher()
+results = searcher.search("人工智能")
+```
+
+**问题**:
+- ❌ 需要知道具体技能名称
+- ❌ 需要手动导入代码
+- ❌ 需要判断使用哪个技能
+
+### 集成后
+
+```python
+# 用户只需表达需求
+user_query = "帮我找关于人工智能的论文"
+
+# literature-expert 自动:
+# 1. 识别语言（中文）
+# 2. 调用 pubscholar-auto-search
+# 3. 返回结果 + 质量评估
+# 4. 提供引用格式化建议
+```
+
+**优势**:
+- ✅ 自然语言交互
+- ✅ 自动技能选择
+- ✅ 端到端服务
+- ✅ 综合分析报告
+
+---
+
+## 📝 总结
+
+### 当前状态
+
+| 项目 | 状态 |
+|------|------|
+| literature-expert 智能体 | ✅ 存在 |
+| pubscholar-auto-search 技能 | ✅ 已创建，**未关联** |
+| arxiv-paper-search 技能 | ✅ 已创建，**未关联** |
+| 三个组件的集成 | ❌ **未完成** |
+
+### 需要做什么
+
+1. ✅ 更新 literature-expert 的 core_skills 配置
+2. ✅ 添加技能调用说明到 literature-expert
+3. ✅ 创建集成示例和文档
+4. ⏳ 测试集成效果
+5. ⏳ 根据测试结果优化
+
+### 预期收益
+
+- 🎯 **用户体验**: 从"手动选择技能"到"自动智能匹配"
+- 🎯 **功能整合**: 文献检索、下载、管理一站式
+- 🎯 **智能化**: 自动语言检测、质量评估、趋势分析
+
+---
+
+**文档日期**: 2025-12-28
+**作者**: Claude Code
+**状态**: 待实施

package/skills/QWEN.md

@@ -0,0 +1,230 @@
+# SSCI Subagent Skills - Comprehensive Documentation
+
+## Project Overview
+
+This directory contains a comprehensive collection of specialized skills for Chinese social science research, implemented as part of the Stigmergy multi-agent system. Each skill is designed to support specific analytical tasks in social science research, following the agentskills.io standard and incorporating progressive disclosure principles.
+
+The skills cover a wide range of methodologies including:
+- Quantitative analysis (statistics, network analysis)
+- Qualitative analysis (grounded theory coding)
+- Theoretical frameworks (Actor-Network Theory, Bourdieu's Field Theory)
+- Mixed-method approaches
+
+## Architecture & Structure
+
+The directory contains multiple skill implementations, each organized in its own subdirectory with the following structure:
+```
+skill-name/
+├── SKILL.md          # Skill documentation with metadata
+├── pyproject.toml    # Dependencies and project config
+├── scripts/          # Implementation scripts
+├── references/       # Additional documentation
+├── tests/            # Test files (where applicable)
+```
+
+Each skill follows a standardized format with:
+- YAML frontmatter metadata
+- Clear usage instructions
+- Progressive disclosure of functionality
+- Standardized JSON output format
+- Integration with the Claude skills system
+
+## Key Skills Categories
+
+### Quantitative Analysis Skills
+- `mathematical-statistics`: Comprehensive statistical analysis toolkit
+- `validity-reliability`: Psychometric validation tools
+- `network-computation`: Social network analysis tools
+- `performing-centrality-analysis`: Network centrality calculations
+
+### Qualitative Analysis Skills
+- `performing-open-coding`: Grounded theory open coding
+- `performing-axial-coding`: Grounded theory axial coding
+- `performing-selective-coding`: Grounded theory selective coding
+- `writing-grounded-theory-memos`: Memo writing for grounded theory
+
+### Theoretical Framework Skills
+- `ant`: Actor-Network Theory analysis
+- `ant-participant-identification`: ANT participant identification
+- `ant-translation-process`: ANT translation process analysis
+- `ant-network-analysis`: ANT network analysis
+- `field-analysis`: Bourdieu's field theory analysis
+- `field-boundary-identification`: Field boundary identification
+- `field-capital-analysis`: Field capital analysis
+
+### Specialized Research Skills
+- `digital-marx`: Digital Marx analysis
+- `digital-durkheim`: Digital Durkheim analysis
+- `digital-weber`: Digital Weber analysis
+- `conflict-resolution`: Conflict resolution in research
+- `dissent-resolution`: Dissent resolution in research
+- `research-design`: Research design framework
+
+## Implementation Standards
+
+### Technical Implementation
+- Python-based implementations with standardized command-line interfaces
+- Argparse for parameter handling
+- Standardized three-layer JSON output format:
+  1. Summary: Key metrics and overview
+  2. Details: Complete analysis results
+  3. Metadata: Processing information and context
+- Error handling and validation
+- Support for Chinese language processing (e.g., jieba for Chinese text)
+
+### Theoretical Rigor
+- Skills grounded in established social science theories
+- Methodological principles clearly articulated
+- Validity and reliability considerations addressed
+- Cultural context integration (especially for Chinese research)
+
+### Progressive Disclosure
+- Core functions presented first
+- Detailed instructions provided as needed
+- Advanced features available for complex analyses
+- Examples provided for common use cases
+
+## Dependencies & Requirements
+
+Most skills require Python 3.8+ and common data science libraries:
+- pandas: Data manipulation
+- numpy: Numerical computations
+- scipy: Statistical functions
+- networkx: Network analysis
+- scikit-learn: Machine learning algorithms
+- jieba: Chinese text processing
+- matplotlib/seaborn: Visualization
+
+Dependencies are specified in individual pyproject.toml files.
+
+## Usage Patterns
+
+Skills are designed to be used in the Claude environment with the following patterns:
+1. Automatic skill selection based on user request
+2. Standardized parameter interface
+3. Rich output with actionable insights
+4. Integration with broader research workflows
+
+## Quality Assurance
+
+- Skills follow established methodological standards
+- Implementation includes validation and error checking
+- Output formats are standardized across skills
+- Cultural appropriateness for Chinese research context
+- Integration with quality checklists and troubleshooting guides
+
+## Integration with Claude Skills Standard
+
+All skills:
+- Follow the agentskills.io standard format
+- Include YAML frontmatter with metadata
+- Provide clear usage instructions
+- Define parameters and expected outputs
+- Include examples of use
+- Specify compatibility and requirements
+
+## Development Conventions
+
+### Coding Standards
+- Python 3.8+ compatibility
+- Type hints for function parameters and return values
+- Comprehensive error handling
+- Standardized command-line interface using argparse
+- UTF-8 encoding for Chinese language support
+
+### Documentation Standards
+- Complete SKILL.md documentation with metadata
+- Clear "When to Use" sections
+- Progressive disclosure of functionality
+- Examples and use cases
+- Quality checklists
+
+## Building and Running
+
+Skills are executed as command-line tools within the Claude environment:
+
+```bash
+# Example execution of a statistical analysis skill
+python scripts/descriptive_statistics.py --input data.csv --output results.json
+
+# Example execution of a network analysis skill
+python scripts/calculate_centrality.py --input network.json --output centrality.json
+
+# Example execution of an open coding skill
+python scripts/auto_loader.py --input interview.txt --output concepts.json
+```
+
+## Future Development
+
+### Planned Enhancements
+- Additional statistical methods
+- Advanced network analysis techniques
+- Integration with more data formats
+- Enhanced visualization capabilities
+- More theoretical frameworks
+- Cross-skill integration workflows
+
+### Maintenance Requirements
+- Regular review of theoretical foundations
+- Updates to implementation based on feedback
+- Expansion of cultural context considerations
+- Integration with new analytical tools
+
+### Organizational Structure and Fusion Plan
+
+Based on analysis of the skills directory, the skills are organized around coherent theoretical frameworks rather than being redundant. The following organizational structure is recommended to better integrate related skills:
+
+#### 1. Actor-Network Theory (ANT) Subagent
+- `ant` (main orchestrator skill)
+- `ant-network-analysis` (network analysis module)
+- `ant-participant-identification` (participant identification module)
+- `ant-translation-process` (translation process module)
+
+These skills form a cohesive ANT analysis framework where the main skill orchestrates specialized modules.
+
+#### 2. Field Analysis Subagent
+- `field-analysis` (main orchestrator skill)
+- `field-boundary-identification` (boundary identification module)
+- `field-capital-analysis` (capital analysis module)
+- `field-habitus-analysis` (habitus analysis module)
+
+These skills form a comprehensive field analysis framework based on Bourdieu's theoretical approach.
+
+#### 3. Grounded Theory Subagent
+- `performing-open-coding` (open coding stage)
+- `performing-axial-coding` (axial coding stage)
+- `performing-selective-coding` (selective coding stage)
+- `checking-theory-saturation` (saturation assessment tool)
+- `writing-grounded-theory-memo` (documentation tool)
+
+These skills represent the sequential stages of grounded theory methodology with supporting tools.
+
+#### 4. Digital Marx Subagent
+- `digital-marx` (main orchestrator skill)
+- `historical-materialist-analysis` (historical materialism module)
+- `class-structure-analysis` (class analysis module)
+- `practical-marxist-application` (application module)
+- `alienation-analysis` (alienation analysis module)
+- `dialectical-quantitative-synthesis` (methodology module)
+
+These skills form a comprehensive Marxist analysis framework integrating various theoretical components.
+
+#### 5. Social Network Analysis (SNA) Subagent
+- `network-computation` (main orchestrator skill)
+- `performing-network-computation` (technical implementation)
+- `processing-network-data` (data processing module)
+- `performing-centrality-analysis` (centrality analysis module)
+
+These skills form a complete network analysis workflow from data processing to analysis.
+
+#### Fusion Plan
+
+A detailed fusion plan has been developed to organize these skills into cohesive subagent systems. The plan maintains specialized functionality while creating clear parent-child relationships:
+
+1. **ANT Subagent**: Integrates all ANT-related skills with the main `ant` skill as orchestrator
+2. **Field Analysis Subagent**: Combines field theory skills under the main `field-analysis` skill
+3. **Grounded Theory Subagent**: Creates a sequential workflow for grounded theory stages
+4. **Digital Marx Subagent**: Unifies Marxist analysis components under `digital-marx`
+5. **SNA Subagent**: Consolidates network analysis skills into a comprehensive framework
+
+For complete details of the fusion plan, see the `fusion_plan.md` file in this directory.