milvus

What problem does it solve?

This Skill solves the problem of turning user intent into working code that connects to Milvus and performs vector database tasks reliably through the pymilvus SDK, without guessing connection details or schema requirements.

Core Features & Use Cases

• Connect and manage environments: Support Milvus Lite, self-hosted Standalone/Cluster, and Zilliz Cloud by guiding users to provide URI and authentication parameters before any connection code is written.
• Create and manage collections: Create collections with quick schemas or fully custom schemas, including loading/releasing and schema/index-related setup required before searching.
• Do vector, hybrid, and full-text search: Insert/upsert vectors, run similarity search, combine dense and sparse results with reranking, and perform BM25 full-text search using Milvus functions.
• Admin operations and guardrails: Manage partitions, databases, and RBAC while emphasizing confirmations before destructive actions and preventing common mistakes like fake vectors.

Quick Start

Use the milvus skill to connect to your Milvus instance by first providing your deployment type, URI, authentication, and database name, then create a collection and run a vector similarity search from an embedding model.