Enhances agent memory with semantic similarity and hybrid BM25 search for precise context retrieval.
This skill integrates advanced retrieval capabilities into the Claude Code environment by leveraging vector embeddings and hybrid search algorithms via the agent-memory system. It enables semantic queries that transcend simple keyword matching, making it indispensable for finding conceptually related discussions, handling synonyms, and navigating large codebases or documentation. By combining traditional BM25 keyword search with modern vector similarity through Reciprocal Rank Fusion (RRF), it ensures that developers receive the most relevant context for their tasks, even when the exact terminology is unknown or difficult to recall.
Key Features
01Configurable weights to balance keyword precision vs. semantic breadth
02Real-time index status monitoring and vector statistics reporting
03Targeted filtering for specific memory segments and TOC nodes
043 GitHub stars
05Hybrid BM25 and vector search using Reciprocal Rank Fusion (RRF)
06Semantic vector similarity search for deep conceptual queries
Use Cases
01Optimizing RAG context by fusing keyword matches with semantic embeddings
02Retrieving documentation fragments when searching with synonyms or technical jargon
03Locating design patterns or architectural discussions based on conceptual similarity
