Memory

Store long-term knowledge in categorized ~/memory/ directories with INDEX.md navigation.

Updated Apr 5, 2026
One-click install
npx skills add https://github.com/dingdyan/openclaw-workspace-v2 --skill memory-dingdyan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: Memory
Source: https://github.com/dingdyan/openclaw-workspace-v2/tree/main/skills/memory
Command: npx skills add https://github.com/dingdyan/openclaw-workspace-v2 --skill memory-dingdyan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Memory helps you store unlimited, well-organized long-term information so your agent can reliably recall projects, people, decisions, knowledge, and growing collections without polluting its built-in memory.

Core Features & Use Cases

  • Unlimited categorized storage in a dedicated system: Creates a separate ~/memory/ workspace where you define the structure (no fixed schema), supported by per-category INDEX.md files for fast navigation.
  • Immediate write-then-respond workflow: When the user shares important info, you save it into the correct file and update the relevant index before answering, keeping the system accurate over time.
  • Optional one-way sync from built-in memory: Copies selected details into ~/memory/sync/ so the parallel system complements built-in memory without conflicts.

Quick Start

Use the memory skill to set up your first categories and store your first item by telling it: "Set up my memory system with categories for projects and people, and then remember that my project is Alpha and it's currently active."

Frequently Asked Questions about Memory

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I organize long-term knowledge for an agent without polluting its built-in memory?

To organize long-term knowledge without polluting built-in memory, use a dedicated ~/memory/ workspace that separates user-defined storage from the agent's internal memory. You define the structure with per-category INDEX.md files for fast navigation and reliable recall of growing collections.

What's the best way to structure evolving project and people information for agent recall?

The best way to structure evolving information is applying per-category INDEX.md navigation files under a dedicated ~/memory/ directory. This approach captures projects, people, decisions, and knowledge with optional one-way sync from built-in memory sources to avoid conflicts.

Can I sync built-in agent memory into a separate knowledge base system?

Yes, you can copy selected details from built-in memory into ~/memory/sync/ using a one-way sync mechanism. This parallel system complements built-in memory without overwriting or conflicting with the original built-in sources.

How do I set up a memory system with custom categories for an agent workflow?

Set up a memory system by creating a root SKILL entry with YAML discovery metadata, then define custom categories in a ~/memory/ directory. Tell the agent your categories and information, and it saves the data into the correct file while updating the relevant index before answering.

Does storing unlimited collections in an agent knowledge base overwrite existing built-in memory?

No, storing unlimited collections in an agent knowledge base applies safe non-overwriting of built-in sources. The system separates user-defined memory into a dedicated directory, ensuring built-in memory remains untouched while the parallel system grows.