What is the Smart Memory System?

It is a retrieval-augmented intelligence layer that adds semantic search and long-term memory capabilities to Claude environments like OpenClaw.

How does it reduce token consumption?

Instead of sending the entire chat history, it uses semantic retrieval to only inject the most relevant context, reducing token usage by up to 80%.

Which models are required for this skill?

It requires an embedding model like BAAI/bge-m3 and a re-ranker model like bge-reranker-v2-m3 for high-precision semantic retrieval.

Does it support automated context cleanup?

Yes, it includes a sophisticated context compression feature that uses soft-trim and hard-clear ratios to prune irrelevant data while protecting recent history.

Smart Memory System

Name: Smart Memory System
Author: dvcrn

bydvcrn

•

生産性とワークフロー

Enhances AI context management using Retrieval-Augmented Generation (RAG) to provide semantic long-term memory and reduce token consumption.

The Smart Memory System is a sophisticated retrieval-augmented intelligence layer designed to overcome context window limitations by providing persistent semantic memory. It utilizes advanced embedding models like BAAI/bge-m3 to cluster related memories, score them by importance, and automatically inject the most relevant historical data into your current session. By replacing keyword-based lookups with semantic search and intelligent context compression, it reduces token usage by up to 80% while significantly improving response accuracy and personalization for complex, long-term projects.

主な機能

013 GitHub stars

02Real-time context enhancement and history injection

03Significant 80% reduction in token consumption via RAG optimization

04Advanced context compression with dual-layer pruning mechanisms

05Automatic memory clustering and importance-based scoring system

06Semantic search powered by BAAI/bge-m3 1024-dimension embeddings

ユースケース

01Building a searchable personal knowledge base for AI-assisted research

02Maintaining persistent technical context across long-running development projects

03Sharing domain-specific project history and decision logs within development teams

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dvcrn/openclaw-skills-marketplace smart-memory-system

For use in Claude.ai and ChatGPT

主な機能

013 GitHub stars

02Real-time context enhancement and history injection

03Significant 80% reduction in token consumption via RAG optimization

04Advanced context compression with dual-layer pruning mechanisms

05Automatic memory clustering and importance-based scoring system

06Semantic search powered by BAAI/bge-m3 1024-dimension embeddings

ユースケース

01Building a searchable personal knowledge base for AI-assisted research

02Maintaining persistent technical context across long-running development projects

03Sharing domain-specific project history and decision logs within development teams

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dvcrn/openclaw-skills-marketplace smart-memory-system

For use in Claude.ai and ChatGPT