菜单
Remember and recall information using MemOS long-term memory. Use when you want to explicitly save important facts, user preferences, or project context, or when you need to search past memories.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/ouchanip/memos-openclaw-local --skill memos-memory命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | memos-memory |
| description | Remember and recall information using MemOS long-term memory. Use when you want to explicitly save important facts, user preferences, or project context, or when you need to search past memories. |
Agent skill for actively remembering and recalling information. Works alongside the lifecycle plugin (auto-capture) for hybrid memory.
Use the bundled script scripts/memos.sh for all operations:
# Remember important information
./scripts/memos.sh remember "User prefers TypeScript over JavaScript"
# Search past memories
./scripts/memos.sh recall "programming language preferences"
# Save as conversation pair (richer context)
./scripts/memos.sh remember-pair "What framework do you use?" "User prefers Next.js + Supabase"
# Check MemOS status
./scripts/memos.sh status
Actively save important information when you discover:
Search past memories when:
| Component | Role |
|---|---|
| MemOS API (localhost:8000) | Memory management server |
| Neo4j (localhost:7687) | Graph DB (relationships) |
| Qdrant (localhost:6333) | Vector DB (semantic search) |
| Ollama (localhost:11434) | Embedding + LLM |
| Tag | Meaning |
|---|---|
| openclaw, auto | Auto-captured by lifecycle plugin |
| openclaw, manual | Explicitly saved by this skill |