openmemory
Persistent long-term agent memory for storing and querying past work, patterns, and learnings.
安装与使用
请根据来源仓库和你的 AI 工具说明安装此技能。先阅读技能文件和依赖要求，再添加到工作流。
查看来源 ↗使用指南技能说明
OpenMemory (Persistent Agent Memory)
You have access to OpenMemory for persistent long-term memory across sessions.
Memory Tools
| Tool | Purpose |
|---|---|
openmemory_query |
Semantic search across memories by similarity |
openmemory_store |
Store new memories with sector classification |
openmemory_list |
List recent memories for a user/agent |
openmemory_get |
Retrieve specific memory by ID |
openmemory_reinforce |
Boost salience of important memories |
Memory Sectors
Memories are classified into sectors:
| Sector | Use Case | Example |
|---|---|---|
| episodic | Events, task history | "Implemented auth flow for project X" |
| semantic | Facts, learned patterns | "Always add Context7 lookup before Rust implementation" |
| procedural | How-to knowledge | "Steps to deploy with ArgoCD" |
Usage Patterns
Before starting a task:
openmemory_query({ query: "similar implementations", sector: "episodic" })
After completing a task:
openmemory_store({
content: "Implemented OAuth2 with PKCE for React app using Effect",
sector: "episodic",
tags: ["auth", "react", "effect"]
})
For important learnings:
openmemory_reinforce({ memory_id: "mem_xyz", boost: 1.5 })
Best Practices
- Query before implementing - Check for similar past work
- Store after completing - Save successful patterns and solutions
- Reinforce important memories - Boost salience of critical learnings
- Tag memories well - Include relevant technologies and patterns
使用示例
查看来源仓库中的使用示例与配置说明。
查看来源 ↗来源历史
- 收录时间
- 2026-01-20
- 索引更新时间
- 2026-03-04
- Branch
- main
收录与同步日期属于本站记录。源码修改历史请查看来源仓库。
GitHub ↗chat Comments (0)
Sign in to join the discussion and leave a comment.
No comments yet. Be the first to share your thoughts!