Files
Szpitale-graph/README.zh.md
T
jilinchenandClaude Opus 4.5 013cc823b6 docs: add real workflow examples from AI Agent Demo research
- Show actual outline.yaml, JSON output, and report
- Keep examples concise and elegant

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2026-01-07 02:03:28 +08:00

2.4 KiB
Raw Blame History

Deep Research Skill for Claude Code

English | 中文

灵感来源:RhinoInsight: Improving Deep Research through Control Mechanisms for Model Behavior and Context

Claude Code 的结构化调研工作流技能,支持两阶段调研:outline生成(可扩展)和深度调查。人在回路设计确保每个阶段的精确控制。

使用场景

  • 学术研究:论文综述、benchmark评测、文献分析
  • 技术研究:技术对比、框架评估、工具选型
  • 市场研究:竞品分析、行业趋势、产品比较
  • 尽职调查:公司研究、投资分析、风险评估

安装

# 中文版
cp -r skills/research-zh ~/.claude/skills/research

# 必需:安装agent
cp agents/web-search-agent.md ~/.claude/agents/

命令

注意:使用 run /research 而非直接 /research,因为斜杠命令与内置命令冲突。

命令 描述
run /research 生成包含items和fields的调研outline
run /research/add-items 向现有outline添加更多items
run /research/add-fields 向现有outline添加更多fields
run /research/deep 使用并行agents对每个item进行深度调研
run /research/report 从JSON结果生成markdown报告

工作流 & 示例

示例:调研 "AI Agent Demo 2025"

阶段1:生成Outline

run /research AI Agent Demo 2025

输出: ai-agent-demo/outline.yaml

topic: "AI Agent Demo & 测评 (2025.9-2025.12)"
items:
  - name: "ChatGPT Agent"
    category: "浏览器Agent"
    brief: "OpenAI统一Agent2025年7月发布"
  - name: "Claude Computer Use"
    category: "桌面Agent"
    brief: "Anthropic桌面操控Agent"
  # ... 另外15个items

阶段2:深度调研

run /research/deep

输出: ai-agent-demo/results/ChatGPT_Agent.json

{
  "basic_info": {
    "name": "ChatGPT Agent",
    "company": "OpenAI",
    "release_date": "2025-07-17",
    "pricing": "Pro $200/月, Plus $20/月"
  },
  "tech_specs": {
    "underlying_model": "GPT-5系列",
    "agent_type": "统一型自主Agent"
  }
}

阶段3:生成报告

run /research/report

输出: ai-agent-demo/report.md - 带目录和所有item的Markdown报告

参考文献

  • RhinoInsight: Improving Deep Research through Control Mechanisms for Model Behavior and Context

许可证

MIT