# Memos Memory Guide

> Practical guide for using an external/shared memory system such as MemOS alongside Hermes memory and session search.

- Skill: `undermybelt/memos-memory-guide` (Agent Skill)
- Install (CLI): `npx skillmds@latest add undermybelt/memos-memory-guide`
- Raw SKILL.md: https://api.skillmd.com/api/skills/undermybelt/memos-memory-guide/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: undermybelt (https://skillmd.com/u/undermybelt)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/undermybelt/memos-memory-guide

---


# MemOS Memory Guide

Use this when the user references an external memory stack such as MemOS or wants cross-agent/shared recall beyond Hermes built-in memory.

Principles:
- Hermes built-in memory is best for compact durable facts about user preferences, environment, and conventions.
- Hermes session_search is best for recalling prior work across past sessions.
- External systems like MemOS are useful for shared, local-first knowledge bases, SOPs, inboxes, and richer memory inspection outside the chat session.

When auto-recall is insufficient:
- Generate a short focused search query.
- Use Hermes session_search first for past-session recall.
- If the workflow depends on a separate local knowledge base, inspect that system directly and summarize relevant findings.

Recommended workflow:
1. Use Hermes memory for durable compact facts.
2. Use session_search when the user references earlier conversations or prior work.
3. Use external memory systems for larger knowledge bases, SOP storage, inbox capture, or multi-agent shared context.
4. Keep private and shared scopes distinct.

Operational guidance:
- Prefer short concrete search terms.
- Verify important hits before acting on them.
- Share only what is necessary across agents or systems.
- Treat local-first plain-text knowledge bases as inspectable source-of-truth artifacts.