jilinchenandClaude Opus 4.5
779c447670
- report/SKILL.md: add AskUserQuestion to allowed-tools - validate_json.py: remove unused imports (List, Any) - validate_json.py: sort output lists for deterministic order - validate_json.py: limit recursion to category level only - validate_json.py: add list-of-dict handling - EN validate_json.py: remove Chinese from CATEGORY_MAPPING - EN deep/SKILL.md: change output language from Chinese to English - deep/SKILL.md: add slug handling for filenames - report/SKILL.md: fix uncertain_fields -> uncertain naming - README: add PyYAML dependency note 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2.7 KiB
Deep Research Skill for Claude Code
Inspired by RhinoInsight: Improving Deep Research through Control Mechanisms for Model Behavior and Context
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
- 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
# English version
cp -r skills/research-en ~/.claude/skills/research
# Chinese version
cp -r skills/research-zh ~/.claude/skills/research
# Required: Install agent
cp agents/web-search-agent.md ~/.claude/agents/
# Required: Install Python dependency
pip install pyyaml
Commands
Note
: Use
run /researchinstead of/researchdirectly, as slash commands conflict with built-in commands.
| 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 |
Workflow & Example
Example: Researching "AI Agent Demo 2025"
Phase 1: Generate Outline
run /research AI Agent Demo 2025
💡 What happens: Tell it your topic → It creates a research list for you
You get: A list of 17 AI Agents to research (ChatGPT Agent, Claude Computer Use, Cursor, etc.) + what info to collect for each
Phase 2: Deep Research
run /research/deep
💡 What happens: AI automatically searches the web for each item, one by one
You get: Detailed info for each Agent (company, release date, pricing, tech specs, reviews...)
Phase 3: Generate Report
run /research/report
💡 What happens: All data → One organized report
You get: report.md - A complete markdown report with table of contents, ready to read or share
Need Help?
If you have questions, ask Claude Code to explain this project:
Help me understand this project: https://github.com/Weizhena/deep-research-skills
References
- RhinoInsight: Improving Deep Research through Control Mechanisms for Model Behavior and Context
License
MIT