Memos Memory Guide

Memos Memory Guide:

Technical Overview

What Is This Skill?

The "Memos Memory Guide" is a specialized skill for the Happycapy Skills platform, designed to interface with the MemOS Local memory system. This system allows agents to efficiently search, retrieve, and manage a user's historical conversation data, preferences, and team-shared knowledge within the OpenClaw workspace. The core purpose of this skill is to provide intelligent context recall and facilitate seamless collaboration by differentiating between locally shared and team-shared memory planes.

Agents utilizing this skill gain access to a suite of memory management tools, including search, retrieval, write, sharing, and task summarization functionalities. The skill operates within the OpenClaw environment, leveraging Node.js and SQLite for robust local data management. It is essential for any workflow requiring conversational context, long-term user preference tracking, or collaborative task management.

Why Use Memos Memory Guide?

Modern conversational agents often struggle to maintain long-term context, leading to fragmented user experiences and repetitive queries. The Memos Memory Guide addresses this by:

• Enabling Persistent Context: Agents can recall user history, preferences, and past conversations, offering coherent and personalized responses.
• Supporting Collaboration: With distinct sharing planes, agents can selectively share or withhold data from the broader team, maintaining privacy or promoting teamwork as needed.
• Automating Memory Retrieval: When explicit recall fails, the agent can auto-generate search queries to retrieve relevant data, minimizing manual intervention.
• Integrating with Task and Skill Management: Beyond conversation history, agents can summarize tasks, search for relevant skills, and facilitate knowledge transfer.

This skill is indispensable for agents operating in environments where context continuity, privacy boundaries, and efficient information flow are critical.

How to Use It

Prerequisites

• OpenClaw workspace set up with required binaries: Node.js, npm, and either curl or wget.
• Environment variables: OPENCLAW_STATE_DIR, OPENCLAW_CONFIG_PATH.
• Configuration file: ~/.openclaw/openclaw.json.
• Dependency: better-sqlite3 Node.js package.

Core Tools and Example Usage

1. Searching Memory

   Use memory_search to query past conversation data. The agent should generate a concise query based on user input if auto-recall is insufficient.

   javascript

   // Example: Search for user's favorite color
   const query = "favorite color";
   const results = await memory_search({ query });

2. Getting Specific Memory Items

   Use memory_get to retrieve a specific memory item by its identifier, typically obtained from a search result.

   javascript

   // Example: Retrieve memory by ID
   const memoryID = "abc123";
   const item = await memory_get({ id: memoryID });

3. Writing to Public Memory

   Use memory_write_public to save new information that should be accessible in the shared plane.

   javascript

   // Example: Store user's timezone
   await memory_write_public({ key: "timezone", value: "UTC+2" });

4. Sharing and Unsharing Memory

   Use memory_share and memory_unshare to control visibility of memory items between local and team planes.

   javascript

   // Example: Share a memory item with the team
   await memory_share({ id: memoryID });
   
   // Example: Unshare the same memory item
   await memory_unshare({ id: memoryID });

5. Task and Skill Management

   Summarize tasks, search for skills, and manage installed skills:

   javascript

   // Get a summary of active tasks
   const summary = await task_summary();
   
   // Search for a specific skill
   const skills = await skill_search({ query: "calendar" });
   
   // Install a new skill
   await skill_install({ id: "calendar-tool" });

Advanced Tools

• network_memory_detail for inspecting distributed memory.
• network_skill_pull for updating team skills.
• memory_timeline and memory_viewer for visualizing memory changes.

When to Use It

This skill should be invoked in the following scenarios:

• User Refers to Past Interactions: Whenever a user references previous conversations, preferences, or asks for information already provided in the past.
• Personalization Required: When tailoring responses based on long-term user behavior or preferences.
• Collaboration and Task Sharing: When agents need to share or withhold information with/from teammates, ensuring the right level of data exposure.
• Skill Discovery and Management: When searching for, installing, or updating skills relevant to the current workflow.
• Context Recovery: When an agent's auto-recall does not yield results, generate a focused search query to retrieve the necessary context.

Important Notes

• Two Distinct Sharing Planes: Local agent sharing is confined to the current OpenClaw workspace, while team sharing extends visibility to all teammates. These must not be confused, as it affects privacy and accessibility.
• Data Privacy: Agents must ensure that sensitive user data is not inadvertently shared across planes. Always verify the required visibility level before sharing memory items.
• Fallback Mechanism: If memory auto-recall is inconclusive, the agent should construct a concise search query and use memory_search rather than returning an empty response.
• Skill and Task Management: Use only the relevant tools for searching, installing, or sharing skills and tasks. Avoid unnecessary installations or publishing actions unless required by the workflow.
• Configuration and Dependencies: Ensure all environment variables and dependencies are correctly set up before deploying the skill.

In summary, the Memos Memory Guide skill is a comprehensive toolset for managing conversational context, user preferences, and collaborative workflows in the OpenClaw environment, with strict boundaries for data sharing and privacy. Proper use of its tools will result in more intelligent, context-aware, and collaborative agents.