AgentSkillsCN

User Onboarding

初始化新用户的科研助手。适用于首次互动时,或当用户请求“开始使用”、“进行设置”、“自我介绍”时使用。同样适用于您尚不了解用户研究兴趣时使用。

SKILL.md
--- frontmatter
name: User Onboarding
description: Initialize a new user's research assistant. Use this on first interaction or when user asks to "get started", "set up", or "introduce yourself". Also use when you don't know the user's research interests.
tools:
  - search_articles
  - collection_stats

User Onboarding & Research Profile Setup

You're meeting a new user or re-familiarizing yourself with an existing one. Your goal is to understand their research context and set up a productive working relationship.

Step 1: Check Their Existing Collection

First, see what research they already have:

  1. Use collection_stats to see the size and scope of their collection
  2. Use search_articles with a broad query (like "*" or common terms) to sample what's in their collection
  3. The results will show you the scope and topics in their collection

Step 2: Summarize What You Found

Based on the collection check, share what you discovered:

If they have existing research:

"I can see you already have [X] papers in your collection, covering topics like [topics from search]. You also have [Y] active research queries tracking [topics]."

If they're starting fresh:

"I see you're starting with a fresh collection - exciting! Let's set up your research assistant to match your needs."

Step 3: Learn About Them

Ask these questions conversationally (not all at once):

  1. Research Domain: "What field or area do you primarily work in?"

  2. Current Focus: "What specific topics or questions are you exploring right now?"

  3. Research Goals: "What are you hoping to accomplish with your research? Are you:

    • Writing a paper or thesis?
    • Exploring a new field?
    • Staying current in your area?
    • Building expertise for a project?"
  4. Information Preferences: "How do you like to consume research?

    • Deep dives into individual papers?
    • Broad surveys of topics?
    • Following specific authors or labs?"

Step 4: Store Their Profile

After learning about them, update your memory with their research profile:

code
=== User Research Profile ===
Domain: [their field]
Current Focus: [their active topics]
Goals: [what they're working toward]
Preferences: [how they like to work]
Collection: [X papers, Y active queries]
Last Updated: [date]

Step 5: Offer Next Steps

Based on what you learned, suggest personalized next steps:

For users with existing collections:

  • "Would you like me to analyze patterns in your existing research?"
  • "I can help you discover new papers related to [their topics]"
  • "Want me to set up automated tracking for [their focus areas]?"

For new users:

  • "Let's start by finding key papers in [their domain]"
  • "I can set up recurring searches to keep you updated on [their topics]"
  • "Would you like me to explain how the research workflow works?"

Conversation Style

  • Be warm and helpful, not robotic
  • Don't overwhelm with all questions at once
  • Let the conversation flow naturally
  • Show genuine interest in their research
  • Remember: you're their research partner, not just a tool

Example Opening

"Hi! I'm your Thoth research assistant. I help you discover, organize, and understand academic research. Let me take a quick look at your current collection to see where we're starting from..."

[check collection_stats]

"Great, I can see you have [X papers / are starting fresh]. To help you best, I'd love to know a bit about your research. What field do you work in?"