name: academic-writing-workflow description: Main workflow skill that orchestrates the complete academic writing process

Academic Writing Workflow

Overview

Main workflow skill that orchestrates the complete academic writing process: parse requirements, search references, parse papers, and generate academic writing output.

When to Use

• User confirms requirement file and wants to proceed with writing
• This is the main entry point for the academic writing process

Workflow Steps

1. Parse Requirements
   └── requirement-parser skill

2. Generate Outline
   └── brainstorming skill (built-in)
   └── Create structured outline based on topic

3. Search References
   └── reference-manager skill
   └── Multi-phase retrieval + PDF parsing

4. Generate Paper
   └── paper-writer skill
   └── Built-in academic writing capabilities

5. Output Result
   └── Return output file path

Input Format

{
  "requirement_file_path": "./path/to/requirement.md",
  "skip_outline_confirmation": false,
  "skip_reference_confirmation": true
}

Output Format

**学术写作流程完成**

**输出文件:** {output_path}

**写作统计:**
- 标题: {title}
- 目标字数: ~{length}
- 风格: {style}
- 参考文献: {n} 篇
- 生成字数: ~{actual_length}

**建议下一步:**
- [审阅] 查看生成的论文
- [修改] 提出修改意见
- [导出] 导出为其他格式

Integration Points

Requires Sub-Skills

• brainstorming - For outline generation (built-in)
• requirement-parser - For parsing requirements
• reference-manager - For reference management
• paper-writer - For paper generation (includes writing capabilities)

Output Format for Next Steps

This skill passes structured data to downstream skills:

{
  "requirement": Requirement,
  "outline": Outline,
  "references": List[Reference],
  "output_path": str
}

Error Handling

• Requirement parse error: Return error details
• Reference search failed: Report and suggest alternatives
• Paper generation failed: Return partial results + error
• File write error: Report and suggest retry

Example

Input: {"requirement_file_path": "./requirements.md"}

Output:
**学术写作流程完成**

**输出文件:** ./output/deep-learning-paper.md

**写作统计:**
- 标题: Deep Learning for Image Classification
- 目标字数: ~2000
- 风格: academic
- 参考文献: 15 篇
- 生成字数: ~2150

**建议下一步:**
- [审阅] 查看生成的论文
- [修改] 提出修改意见
- [导出] 导出为其他格式