Skip to content

milkoor/openclaw-professional-research-report

Repository files navigation

OpenClaw Professional Research Report Skill

English | 中文


English

Overview

Professional Research Report Skill is an OpenClaw skill for generating high-quality research reports following MBB/Big 4/Gartner standards.

Features

  • MBB/Big 4 Standard: Professional report structure
  • Multiple Frameworks: SWOT, PESTEL, Porter Five Forces
  • Data Annotation: Mandatory source citations
  • SMART Recommendations: Actionable advice templates
  • 10-Dimension Quality Evaluation: Self-assessment system
  • AI对抗训练: Multi-round quality improvement through adversarial review

Installation

1. Install via GitHub (Recommended)

# Clone to OpenClaw skills directory
cd ~/.openclaw/skills
git clone https://github.com/milkoor/openclaw-professional-research-report.git

# Restart OpenClaw
openclaw gateway restart

2. Install via OpenClaw CLI

openclaw skills install professional-research-report

Usage

Generate a Report

# Using the included script
./scripts/generate.sh -t "Your Topic" -o "output.md"

# Options:
# -t: Report topic (required)
# -o: Output file (required)
# -w: Target word count (default: 6000)
# -f: Analysis frameworks (default: SWOT,PESTEL)
# -s: Apply SMART principles
# -v: Verbose output

Quality Training

# Improve report quality through AI adversarial training
./scripts/train.sh -r "your_report.md" -o "training_output/" -t 4.5

# Options:
# -r: Report file (required)
# -o: Output directory (required)
# -t: Target score (default: 4.0)
# -m: Max iterations (default: 5)
# -q: Quick mode
# -v: Verbose output

10-Dimension Quality Evaluation

Dimension Weight Description
Structure Completeness 10% Standard structure coverage
Executive Summary 15% Understandable in 5 minutes
Data Annotation Rate 20% Source citation completeness
Multi-Source Verification 10% Key data verification
Framework Application 10% Analysis framework usage
Action Title Rate 10% Conclusion-style titles
SMART Recommendation Rate 10% SMART-compliant advice
Logical Chain 5% Insight → Analysis → Recommendation
Predictive 5% Future trend forecasting
Benchmark Analysis 5% Industry best practices

File Structure

professional-research-report/
├── SKILL.md                  # Skill documentation
├── scripts/
│   ├── generate.sh          # Report generation script
│   └── to_pdf.sh            # PDF conversion script
├── GITHUB_UPDATE_SUMMARY.md # Update history
└── 验证报告.md               # Verification report

Version History

Version Date Changes
1.0.0 2026-03-19 Initial release with MBB standards

License

MIT License


中文

概述

专业调研报告技能是 OpenClaw 的技能,用于生成符合 MBB/Big 4/Gartner 标准的高质量调研报告。

功能特性

  • MBB/Big 4 标准:专业报告结构
  • 多种分析框架:SWOT、PESTEL、波特五力
  • 数据标注:强制来源引用
  • SMART 建议:可操作建议模板
  • 10 维度质量评估:自我评估系统
  • AI 对抗训练:通过多轮对抗评审提升质量

安装

1. 通过 GitHub 安装(推荐)

# 克隆到 OpenClaw skills 目录
cd ~/.openclaw/skills
git clone https://github.com/milkoor/openclaw-professional-research-report.git

# 重启 OpenClaw
openclaw gateway restart

2. 通过 OpenClaw CLI 安装

openclaw skills install professional-research-report

使用方法

生成报告

# 使用附带的脚本
./scripts/generate.sh -t "您的主题" -o "输出.md"

# 选项:
# -t: 报告主题(必需)
# -o: 输出文件(必需)
# -w: 目标字数(默认:6000)
# -f: 分析框架(默认:SWOT,PESTEL)
# -s: 应用 SMART 原则
# -v: 详细输出

质量训练

# 通过 AI 对抗训练提升报告质量
./scripts/train.sh -r "您的报告.md" -o "训练输出/" -t 4.5

# 选项:
# -r: 报告文件(必需)
# -o: 输出目录(必需)
# -t: 目标分数(默认:4.0)
# -m: 最大迭代次数(默认:5)
# -q: 快速模式
# -v: 详细输出

10 维度质量评估

维度 权重 说明
结构完整性 10% 标准结构覆盖率
执行摘要 15% 5 分钟可理解
数据标注率 20% 数据来源标注完整性
多源验证率 10% 关键数据多源验证
框架应用 10% 分析框架使用
行动标题率 10% 结论式标题比例
SMART 建议率 10% 建议符合 SMART 原则
逻辑链条 5% 发现→分析→建议连贯性
预测性 5% 未来趋势预测
对标分析 5% 行业最佳实践对比

文件结构

professional-research-report/
├── SKILL.md                  # 技能文档
├── scripts/
│   ├── generate.sh          # 报告生成脚本
│   └── to_pdf.sh           # PDF 转换脚本
├── GITHUB_UPDATE_SUMMARY.md # 更新历史
└── 验证报告.md               # 验证报告

版本历史

版本 日期 变更
1.0.0 2026-03-19 初始版本,MBB 标准

许可证

MIT License


作者: 杨博
最后更新: 2026-03-25

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages