A high-performance, file-based semantic memory system enabling AI agents to persist and recall learnings and decisions with sub-20ms search speeds.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memory-system-v2-1-0-0
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install memory-system-v2-1-0-0 using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Memory System v2.0?
Memory System v2.0 is a lightweight, pure Bash and jq-based infrastructure designed to provide AI agents with a persistent memory layer. Unlike standard context windows that clear between sessions, this skill allows agents to store learnings, decisions, insights, and events in a structured JSON index for near-instant retrieval. It is a vital component for developers building sophisticated Openclaw Skills that require historical context and long-term intelligence.
By prioritizing speed and simplicity, the system avoids the overhead of traditional databases while maintaining rigorous data organization. It utilizes a file-based architecture that scales effectively for agent workflows, ensuring that critical information like architectural decisions or user preferences is never lost. This enables a more seamless and intelligent interaction model where the agent grows smarter with every session.
Memory System v2.0 Use Cases
- Persistent tracking of technical skills and patterns acquired during development sessions.
- Recording architectural decisions and strategy shifts to maintain a historical log of why choices were made.
- Capturing breakthrough insights and realizations to improve future problem-solving.
- Logging project milestones and event completions for automated status reporting.
- Storing user feedback and interaction details to personalize future agent responses.
How Memory System v2.0 Works
- The agent or user invokes the capture command to record a memory with specific metadata including type, importance (1-10), and context.
- The system appends the raw memory entry to a daily Markdown file for human-readable persistence.
- Simultaneously, the skill updates a centralized JSON index to facilitate high-speed semantic searches.
- When context is needed, the agent runs a search query against the index, filtering by keywords or minimum importance levels.
- An automated consolidation process can be triggered to summarize daily logs into weekly overview files for high-level analysis.
Memory System v2.0 Setup
To integrate this into your environment, follow these steps:
# Install the jq dependency (required for JSON processing)
brew install jq
# Ensure the memory-cli.sh script is executable
chmod +x ./memory/memory-cli.sh
# Initialize the system by capturing your first memory
./memory/memory-cli.sh capture --type learning --importance 10 --content "Initialized Memory System v2.0" --tags "setup,init"
Memory System v2.0 Data Schema & Taxonomy
The system organizes data across several directories to balance speed and readability:
| Directory | File | Purpose |
|---|---|---|
memory/ |
memory-cli.sh |
The primary command-line interface for all operations. |
memory/index/ |
memory-index.json |
High-speed JSON index containing metadata and file pointers. |
memory/daily/ |
YYYY-MM-DD.md |
Human-readable Markdown logs of every captured memory. |
memory/consolidated/ |
YYYY-WW.md |
Weekly summaries generated by the consolidation engine. |
Memory System v2.0 Advanced Features
- Sub-20ms search performance optimized for real-time agent responses.
- Multi-word search functionality with importance-based weighting and filtering.
- Automatic weekly consolidation that uses AI to summarize daily events into coherent reports.
- Detailed analytics and statistics to track memory distribution across different types and importance levels.
- Native compatibility with Openclaw Skills workflows via AGENTS.md auto-capture protocols.
SKILL.md
Loading
Related Openclaw Skills
mem0 Memory Integration
The mem0 skill provides an intelligent, adaptive memory layer that learns and recalls user preferences, patterns, and context across conversations.
gog: Google Workspace CLI
A versatile command-line tool for managing Gmail, Calendar, Drive, Sheets, and Docs via simple terminal commands.
ElevenLabs Agent Manager
A comprehensive automation skill for creating, syncing, and deploying ElevenLabs conversational voice agents and webhook integrations.
Clawd Cursor
An AI-powered desktop agent that enables OpenClaw users to control any Windows or macOS application through natural language commands.
NEON-SOUL
NEON-SOUL provides an automated framework for synthesizing an AI agent's identity and core principles directly from its interaction history and memory files.
QMD Memory Skill
A local hybrid search engine that eliminates API costs by running embeddings and semantic queries directly on your machine.