Claude Skill
kevin-hs-sohn/hipocampus
Hipocampus is a drop-in memory harness for AI agents with 3-tier memory, compaction tree, and hybrid search. One-command setup, works with Claude Code and OpenClaw.
Overview
Repository
Install this Skill
git clone https://github.com/kevin-hs-sohn/hipocampus.gitRegistry
Summary
Hipocampus is a drop-in memory harness for AI agents, featuring a 3-tier memory architecture, compaction tree, and hybrid search. It can be set up with a single command and works seamlessly with Claude Code and OpenClaw.
AI代理的即插即用记忆框架——三级记忆架构、压缩树与混合搜索。一键部署,兼容Claude Code与OpenClaw。
Key features
- 3-tier memory architecture for structured recall
- Compaction tree for efficient memory management
- Hybrid search combining vector and keyword methods
- One-command setup for rapid integration
- Compatible with Claude Code and OpenClaw
Use cases
- Enhancing AI agent long-term memory in coding assistants
- Building context-aware conversational agents
- Improving knowledge retrieval in RAG pipelines
- Managing agent context across multiple sessions