---
name: "qdrant-payload-filtered-hnsw"
description: "Perform vector similarity search with strict metadata payload filtering using Qdrant."
metadata:
  author: "Qdrant"
  version: "1.0.0"
  runtime_compatibility:
    - "codex"
    - "claude-code"
    - "cursor"
    - "windsurf"
triggers:
  - "setup qdrant collection"
  - "payload filter qdrant vector search"
  - "tune qdrant hnsw"
permissions:
  fs_read: true
  fs_write: "workspace"
  shell_exec: false
  net_access: false
---

# Qdrant Vector Database Payload-Filtered HNSW

Configures Qdrant collections: setting up HNSW graph vector storage, custom payload schemas, payload-based index filtering without recall degradation, snapshot backups, and Rust-powered gRPC client queries.

## When to Invoke
Invoke this skill whenever you need to:
- setup qdrant collection
- payload filter qdrant vector search
- tune qdrant hnsw

## Execution SOP & Rules
1. Initialize client: client = QdrantClient(url='http://localhost:6333').
2. Create collection: client.create_collection('articles', vectors_config=VectorParams(size=768, distance=Distance.COSINE)).
3. Query with filter: client.search(collection_name='articles', query_vector=vec, query_filter=Filter(...)).
4. Enable scalar quantization in collection optimizers config for high-density deployments.

## Verification Checklist
- Run local linting, syntax validation, and type-checks before reporting completion.
- Ensure all created or modified files adhere to workspace boundaries and avoid uncommitted secrets.
- Verify tests or build commands exit cleanly with return code 0.