AgentSkillsCN

research

借助 AI 生成并分析 SaaS 商业创意。适用于头脑风暴创业构想、生成商业点子,或评估 SaaS 机遇。AI 会自动生成结构化的创意文件,并同时进行多维度分析。

SKILL.md
--- frontmatter
name: research
description: Conduct and document research with consistent file organization. Use when researching topics, gathering information, or when the user asks to investigate a subject. Saves findings to research/yyyy-mm-dd-title.md format.

Research

Workflow

  1. Create research file: research/YYYY-MM-DD-topic-slug.md
  2. Conduct research: Use web search for current information
  3. Document findings: Keep content short and scannable
  4. Multiple topics: Launch sub-agents in parallel

File Naming

code
research/
├── 2026-01-31-vue-composables.md
├── 2026-01-31-tailwind-v4.md
└── 2026-01-30-typescript-patterns.md

Format: research/YYYY-MM-DD-slug.md

  • Use today's date
  • Lowercase slug with hyphens
  • One file per topic

Document Structure

markdown
# [Topic Title]

## Summary

[2-3 sentence overview]

## Key Findings

- Finding 1
- Finding 2
- Finding 3

## Sources

- [Source 1](url)
- [Source 2](url)

## Next Steps (optional)

- Action item 1
- Action item 2

Research with Web Search

  1. Use WebSearch tool to find current information
  2. Use WebFetch for specific documentation pages
  3. Extract key points, avoid copying large blocks
  4. Always cite sources with URLs

Multiple Topics

For 2+ topics, launch parallel sub-agents:

code
Task tool calls (in parallel):
- Agent 1: "Research [topic A], save to research/YYYY-MM-DD-topic-a.md"
- Agent 2: "Research [topic B], save to research/YYYY-MM-DD-topic-b.md"

Each agent:

  • Researches one topic
  • Creates one file
  • Returns summary of findings

Quality Checklist

  • File saved in research/ directory
  • Filename follows YYYY-MM-DD-slug.md format
  • Content is scannable (bullets, headers)
  • Sources cited with URLs
  • No walls of text