Qdrant VectorStore Skill

Skill Information

Name: qdrant-vectorstore Description: Manage Qdrant Cloud vector operations for document embedding storage and semantic search Version: 1.0.0 Author: Claude Code

Purpose

Handle document embedding storage and semantic search for RAG chatbot using Qdrant Cloud.

Inputs

• •operation (required): "create_collection" | "upsert" | "search"
• •collection_name (required): string - Name of the Qdrant collection
• •documents (optional): string[] - Array of documents to embed (for upsert operation)
• •query (optional): string - Search query text (for search operation)
• •top_k (optional): number - Number of results to return (default: 5)

Outputs

• •For create_collection: { collection_id: string }
• •For upsert: { inserted_count: number }
• •For search: Array of { text: string, score: number, metadata: object }

Implementation Requirements

• •Platform: Qdrant Cloud Free Tier endpoint
• •Embedding Model: text-embedding-3-small (OpenAI)
• •Vector Dimension: 1536
• •Distance Metric: Cosine

Environment Variables

• •QDRANT_URL: Qdrant Cloud endpoint URL
• •QDRANT_API_KEY: Qdrant Cloud API key
• •OPENAI_API_KEY: OpenAI API key for embeddings

Error Handling

• •Retry on rate limits with exponential backoff
• •Validate collection exists before operations
• •Log all operations for debugging
• •Handle network timeouts gracefully

Usage Examples

Create Collection

{
  "operation": "create_collection",
  "collection_name": "my_documents"
}

Upsert Documents

{
  "operation": "upsert",
  "collection_name": "my_documents",
  "documents": [
    "Document 1 text content",
    "Document 2 text content"
  ]
}

Search Documents

{
  "operation": "search",
  "collection_name": "my_documents",
  "query": "What is the meaning of life?",
  "top_k": 3
}

Dependencies

• •@qdrant/js-client-rest
• •openai
• •dotenv

Implementation Notes

• •Documents are automatically chunked if too long for embedding model
• •Each document gets a unique ID and timestamp metadata
• •Search results are ranked by similarity score (0-1)
• •All operations are logged with timestamps and success/failure status