What is the difference between vector search and BM25?

BM25 is a traditional keyword-based search that looks for exact word matches, while vector search uses mathematical embeddings to find content that is conceptually similar even if the specific words differ.

How do I check if my vector index is up to date?

You can use the 'teleport vector-stats' command to view the index status, the number of vectors stored, and the timestamp of the last indexing operation.

When should I use hybrid search instead of pure vector search?

Use hybrid search when your query contains specific technical terms or function names that need exact matching, but you also want to capture broader conceptual context.

Does this skill require a background service to run?

Yes, the vector-search skill depends on the memory-daemon being active to handle the gRPC requests and manage the local vector index.

Vector Memory Search

Name: Vector Memory Search
Author: SpillwaveSolutions

bySpillwaveSolutions

•

Productivity & Workflow

Enables semantic similarity and hybrid keyword-vector searching within the agent-memory system to find conceptually related discussions.

The Vector Search skill provides advanced retrieval capabilities for the agent-memory system, allowing Claude to perform semantic similarity searches and hybrid BM25+vector fusion. It is specifically designed for scenarios where standard keyword searches return poor results or when the user needs to find conceptually related topics, synonyms, or technical patterns across past interactions. By utilizing Reciprocal Rank Fusion (RRF), it allows developers to balance the precision of keyword matching with the breadth of semantic embeddings, ensuring high-quality context retrieval for complex agent workflows.

Key Features

01Real-time vector index statistics and status monitoring

02Semantic vector search for conceptual similarity matching

03Hybrid search mode combining BM25 keyword matching and vector embeddings

04Reciprocal Rank Fusion (RRF) for optimized search result ranking

054 GitHub stars

06Configurable search weights to prioritize either keyword or semantic relevance

Use Cases

01Resolving queries involving synonyms or typos that standard search tools would miss

02Finding related code snippets or documentation based on conceptual meaning rather than syntax

03Locating past architectural discussions or patterns when exact keywords are forgotten

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add spillwavesolutions/agent-memory vector-search

For use in Claude.ai and ChatGPT

Download Skill