V3 Memory Unification

Unify six memory systems into AgentDB with HNSW indexing.

2|Updated May 8, 2026
One-click install
npx skills add https://github.com/xotong/claude-marketplace --skill v3-memory-unification-xotong
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: V3 Memory Unification
Source: https://github.com/xotong/claude-marketplace/tree/main/plugins/ruflo/skills/v3-memory-unification
Command: npx skills add https://github.com/xotong/claude-marketplace --skill v3-memory-unification-xotong

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Unify six memory systems into a single AgentDB backend, enabling fast, cross-agent memory sharing and scalable search with HNSW indexing.

Core Features & Use Cases

  • Unified memory backend that consolidates multiple storage systems (SQLiteBackend, MarkdownBackend, etc.)
  • High-performance semantic search using HNSW indexing
  • Smooth data migration from legacy backends with backward compatibility

Quick Start

Migrate and unify your existing memories into AgentDB and enable semantic search.

Frequently Asked Questions about V3 Memory Unification

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I unify multiple memory backends into a single database for semantic search?

To unify multiple memory backends, you consolidate disparate storage systems like SQLite and Markdown into a single AgentDB. This process uses HNSW indexing to deliver high-performance semantic search across all consolidated memories.

What is the best way to migrate legacy memory data to a new vector search backend?

The best way to migrate legacy memory data is using a DataMigrator with embedding generation capabilities. It ensures smooth data migration from legacy backends to AgentDB while maintaining backward compatibility and enabling HNSW indexing.

How does HNSW indexing improve agent memory search performance?

HNSW indexing improves agent memory search performance by applying hierarchical navigable small world graphs to the unified AgentDB backend. This mechanism delivers 150x to 12,500x search improvements compared to querying disparate legacy memory systems.

Do I need an AgentDBAdapter and HNSWIndexer to enable cross-agent memory sharing?

Yes, you need an AgentDBAdapter and an HNSWIndexer to enable cross-agent memory sharing. These components, along with a DataMigrator for embedding generation, are required to unify disparate backends and deliver scalable semantic search.

Can I consolidate SQLite and Markdown memory systems while maintaining backward compatibility?

Yes, you can consolidate SQLite and Markdown memory systems into AgentDB while maintaining backward compatibility. The data migration workflow ensures smooth transition from legacy backends without losing existing memory data during the unification process.