From 5dcafa4ef0aaf97d591cd6ef5d2470b305ab1c17 Mon Sep 17 00:00:00 2001 From: jilinchen Date: Wed, 7 Jan 2026 02:00:15 +0800 Subject: [PATCH] docs: split README into separate EN/ZH files MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - README.md for English (default) - README.zh.md for Chinese - Add language switch links at top 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 --- README.md | 113 +++++++++------------------------------------------ README.zh.md | 79 +++++++++++++++++++++++++++++++++++ 2 files changed, 99 insertions(+), 93 deletions(-) create mode 100644 README.zh.md diff --git a/README.md b/README.md index 7d89316..2fa8b4b 100644 --- a/README.md +++ b/README.md @@ -1,44 +1,43 @@ # Deep Research Skill for Claude Code -[English](#english) | [中文](#中文) - ---- - -## English +[English](README.md) | [中文](README.zh.md) > Inspired by [RhinoInsight: Improving Deep Research through Control Mechanisms for Model Behavior and Context](https://arxiv.org/abs/2511.18743) A structured research workflow skill for Claude Code, supporting two-phase research: outline generation (extensible) and deep investigation. Human-in-the-loop design ensures precise control at every stage. -### Use Cases +## Use Cases - **Academic Research**: Paper surveys, benchmark reviews, literature analysis - **Technical Research**: Technology comparison, framework evaluation, tool selection - **Market Research**: Competitor analysis, industry trends, product comparison - **Due Diligence**: Company research, investment analysis, risk assessment - -### Installation - -Choose your language version: +## Installation ```bash # English version cp -r skills/research-en ~/.claude/skills/research -# Chinese version -cp -r skills/research-zh ~/.claude/commands/research - # Required: Install agent cp agents/web-search-agent.md ~/.claude/agents/ ``` -### Commands + +## Commands > **Note**: Use `run /research` instead of `/research` directly, as slash commands conflict with built-in commands. -### Workflow +| Command | Description | +|---------|-------------| +| `run /research` | Generate research outline with items and fields | +| `run /research/add-items` | Add more items to existing outline | +| `run /research/add-fields` | Add more fields to existing outline | +| `run /research/deep` | Deep research each item with parallel agents | +| `run /research/report` | Generate markdown report from JSON results | -#### Phase 1: Generate Outline +## Workflow + +### Phase 1: Generate Outline ``` run /research ``` @@ -47,7 +46,7 @@ run /research - User confirms and adjusts - Outputs: `outline.yaml` (items + config) + `fields.yaml` (field definitions) -#### Phase 2: Deep Research +### Phase 2: Deep Research ``` run /research/deep ``` @@ -56,13 +55,13 @@ run /research/deep - Supports checkpoint resume - Outputs: `results/*.json` -#### Optional: Expand Outline +### Optional: Expand Outline ``` run /research/add-items # Add research targets via user input or web search run /research/add-fields # Add field definitions ``` -#### Phase 3: Generate Report +### Phase 3: Generate Report ``` run /research/report ``` @@ -71,82 +70,10 @@ run /research/report - Skips uncertain values automatically - Outputs: `report.md` ---- - -## 中文 - -> 灵感来源：[RhinoInsight: Improving Deep Research through Control Mechanisms for Model Behavior and Context](https://arxiv.org/abs/2511.18743) - -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 - -选择语言版本： - -```bash -# 英文版 -cp -r skills/research-en ~/.claude/skills/research - -# 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。 - -### 工作流 - -#### 阶段1：生成Outline -``` -run /research -``` -- 模型知识生成初始items和字段框架 -- 网络搜索补充最新items -- 用户确认并调整 -- 输出：`outline.yaml`（items + 配置）+ `fields.yaml`（字段定义） - -#### 阶段2：深度调研 -``` -run /research/deep -``` -- 并行agents调研每个item（batch_size可配置） -- 每个agent读取fields.yaml并输出结构化JSON -- 支持断点续传 -- 输出：`results/*.json` - -#### 可选：扩展Outline -``` -run /research/add-items # 通过用户输入或网络搜索添加调研对象 -run /research/add-fields # 添加字段定义 -``` - -#### 阶段3：生成报告 -``` -run /research/report -``` -- 生成Python脚本将JSON转换为markdown -- 用户选择目录中显示的摘要字段 -- 自动跳过不确定值 -- 输出：`report.md` - ---- - -## References / 参考文献 +## References - RhinoInsight: Improving Deep Research through Control Mechanisms for Model Behavior and Context -## License / 许可证 +## License MIT diff --git a/README.zh.md b/README.zh.md new file mode 100644 index 0000000..3cf810f --- /dev/null +++ b/README.zh.md @@ -0,0 +1,79 @@ +# Deep Research Skill for Claude Code + +[English](README.md) | [中文](README.zh.md) + +> 灵感来源：[RhinoInsight: Improving Deep Research through Control Mechanisms for Model Behavior and Context](https://arxiv.org/abs/2511.18743) + +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 + +```bash +# 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 + +### 阶段1：生成Outline +``` +run /research +``` +- 模型知识生成初始items和字段框架 +- 网络搜索补充最新items +- 用户确认并调整 +- 输出：`outline.yaml`（items + 配置）+ `fields.yaml`（字段定义） + +### 阶段2：深度调研 +``` +run /research/deep +``` +- 并行agents调研每个item（batch_size可配置） +- 每个agent读取fields.yaml并输出结构化JSON +- 支持断点续传 +- 输出：`results/*.json` + +### 可选：扩展Outline +``` +run /research/add-items # 通过用户输入或网络搜索添加调研对象 +run /research/add-fields # 添加字段定义 +``` + +### 阶段3：生成报告 +``` +run /research/report +``` +- 生成Python脚本将JSON转换为markdown +- 用户选择目录中显示的摘要字段 +- 自动跳过不确定值 +- 输出：`report.md` + +## 参考文献 + +- RhinoInsight: Improving Deep Research through Control Mechanisms for Model Behavior and Context + +## 许可证 + +MIT