Does this skill support existing memory data from legacy systems?

Yes, the skill includes built-in migration strategies and adapters for moving data from SQLite, Markdown, and other legacy memory managers into the unified AgentDB backend.

What is the main benefit of the HNSW index in this skill?

HNSW (Hierarchical Navigable Small World) indexing allows for extremely fast approximate nearest neighbor searches, providing up to a 12,500x performance boost over traditional linear search methods for vector embeddings.

What are the performance targets for query latency?

The architecture is designed to maintain sub-100ms query latency even when handling over 1 million memory entries.

How does it improve multi-agent collaboration?

By unifying memory into a single service, multiple agents can access a shared pool of knowledge and learning patterns in real-time, preventing information silos.

Unified AgentDB Memory & Vector Search

Name: Unified AgentDB Memory & Vector Search
Author: nbossn

bynbossn

0•

数据库管理

Consolidates disparate agent memory systems into a high-performance AgentDB backend with HNSW vector indexing for ultra-fast retrieval.

This skill optimizes Claude's internal memory architecture by unifying legacy systems—including SQLite, Markdown, and Distributed Memory—into a single, high-performance AgentDB implementation. By leveraging HNSW (Hierarchical Navigable Small World) indexing for semantic vector search, it achieves search performance improvements ranging from 150x to 12,500x while maintaining full backward compatibility. It is an essential upgrade for developers building complex agentic workflows that require low-latency access to massive datasets, cross-agent memory sharing, and integrated SONA learning patterns.

主要功能

01HNSW vector indexing for 150x-12,500x faster semantic search

020 GitHub stars

03Automated migration paths for SQLite and Markdown-based data

04Real-time cross-agent memory synchronization and sharing

05Unified AgentDB backend replacing 7+ legacy memory systems

06SONA learning pattern integration for adaptive behavior storage

使用场景

01Scaling agentic workflows that require rapid retrieval from millions of historical memory entries

02Implementing shared memory architectures for multi-agent systems to enable collective learning

03Migrating legacy file-based memory storage to high-performance vector-first databases

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add nbossn/claude-code-base v3-memory-unification

For use in Claude.ai and ChatGPT

主要功能

01HNSW vector indexing for 150x-12,500x faster semantic search

020 GitHub stars

03Automated migration paths for SQLite and Markdown-based data

04Real-time cross-agent memory synchronization and sharing

05Unified AgentDB backend replacing 7+ legacy memory systems

06SONA learning pattern integration for adaptive behavior storage

使用场景

01Scaling agentic workflows that require rapid retrieval from millions of historical memory entries

02Implementing shared memory architectures for multi-agent systems to enable collective learning

03Migrating legacy file-based memory storage to high-performance vector-first databases

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add nbossn/claude-code-base v3-memory-unification

For use in Claude.ai and ChatGPT