Agent skill

connect-users

Find community members with relevant expertise using vector search

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/testing/connect-users-mlai-aus-inc-roo

SKILL.md

Connect Users Skill

This skill enables Claude to find and recommend MLAI community members based on their expertise, interests, and what they're working on.

Capabilities

  • Search for users with specific expertise using vector similarity
  • Match users based on topics, skills, and interests
  • Provide warm introductions and connection suggestions

Parameters

  • query: The expertise or topic the user is looking for (required)
  • exclude_user_id: Slack user ID to exclude from results (optional - usually the requester)
  • limit: Maximum number of users to suggest (default: 5)

Workflow

Step 1: Extract Topics

Parse the user's query to identify the specific expertise areas they're looking for.

Common patterns to recognize:

  • "who knows about X" → extract X
  • "anyone working on Y" → extract Y
  • "expert in Z" → extract Z
  • "looking for someone in [field]" → extract field
  • "recommend someone for [topic]" → extract topic

Step 2: Vector Search

Use the vector_search function to find users with matching expertise.

The search uses cosine similarity on embeddings stored in the user_expertise table.

sql
SELECT u.id, u.name, u.slack_id, e.topic, e.relationship,
       1 - (e.embedding <=> $query_embedding) as similarity
FROM users u
JOIN user_expertise e ON u.id = e.user_id
WHERE 1 - (e.embedding <=> $query_embedding) > 0.7
ORDER BY similarity DESC
LIMIT 5;

Step 3: Format Response

Generate a warm, friendly response suggesting the matched users.

Include for each user:

  • Their name (with Slack mention if available)
  • Their expertise area that matched
  • The relationship type (expert, working on, interested)

Response Style

  • Start with an acknowledgment of what they're looking for
  • Use casual Australian language occasionally (mate, no worries, legend, etc.)
  • Be encouraging about making connections
  • If no users found, offer alternative suggestions or ask for clarification

Example Responses

Successful Match

G'day! Looking for folks in AI research, eh? Here are some legends who might help:

• **@sam** - Expert in machine learning and neural networks
• **@jane** - Currently working on computer vision projects
• **@bob** - Interested in deep learning applications

Feel free to reach out to them! 🦘

No Matches Found

Hmm, I couldn't find anyone specifically matching "quantum computing" in our community yet.

A few things we could try:
• Broaden the search - maybe "physics" or "advanced computing"?
• Post in #introductions asking if anyone's into this space
• Check out our upcoming events - might meet someone there!

Want me to try a different search? 🤔

Error Handling

If the database search fails:

  1. Apologize briefly
  2. Suggest alternative ways to find help (Slack channels, events)
  3. Offer to try again

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results